Original article
Schistosomiasis and intestinal helminthic infections
in Delo Awraja, Bale administrative region south
Ethiopia
Terefe Wodimagegnehu1, Hailu Birrie2 and Hailu Yeneneh3
Abstract: A cross-sectional survey of schistosomiasis, intestinal helminthic infections and the snail
intermediate hosts of schistosomiasis was conducted in Delo Awraja, Bale Administrative Region,
South Ethiopia. Of 15 accessible communities studied, the prevalence of intestinal schistosomiasis
exceeded 5% in five of them, reaching 48% in Meda. No S. haematobium infection was found except
two imported cases diagnosed at Melkaaman at a temporary shelter for refugees returning from
Somalia. Uninfected snail hosts of S. mansoni were collected from few water bodies. Other intestinal
helminth parasites were also highly prevalent and widespread, the dominant ones being Ascaris
lumbricoides, hookworm species and Trichuris trichiura, infection rates reaching 84%, 67% and 64%
respectively in some communities located at altitudes of 1500-1600 m.a.s.l. The apprehension that
schistosomiasis may further spread with future development plans and the need for initiation of
surveillance and/or control programmes for schistosomiasis in particular and intestinal helminthic
infections in general are discussed. [Ethiop. J. Health Dev. 1997;11(3):183-188]
Introduction
The distribution and prevalence of schistosomiasis and intestinal helminthic infection for parts of
Ethiopia are quite well understood except for the southern regions where persistent security problems
and/or lack of motorable roads have hindered thorough investigation (1). Of all, the Bale
Administrative Region continues to be the least studied. Information on human schistosomiasis is
limited to a survey of few towns conducted by the Institute of Pathobiology, Addis Ababa University,
in the 1980s (2). Now that the security and communication problems have much improved few
developmental activities are already underway and much more are in the planning phases. Norwegian
Church Aid (NCA) in Ethiopia is implementing a multi-disciplinary development programme
known as Dello Development project (DDP) in the area. The programme encompass primary health
care (PHC), water and agricultural development with the major objective of ensuring community selfhelp and nutritional status. Furthermore, the government was conducting a preliminary survey of the
irrigation potential of the Wolmel-Dumel river basin located in the Awraja (Fig. 1). The objective of
this study was to generate pre-development data especially on schistosomiasis since experience has
shown that this disease is known to aggravate with water development and population settlement
schemes (3, 5). Such baseline data are also necessary in planning disease prevention and/or control
programs.
Methods
184 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Study area: Dello is one of the five Awrajas (sub-regions) in Bale Administrative Region, southern
Ethiopia (Fig. 1). It is divided into five woredas (districts) with a total population of about 100.000
people (6). The Awraja is bounded by the Genale and Mena River Basins to the west and east,
______________________________________
1
From the Norwegian Church Aid, Addis Ababa; 2Institute of Pathobiology, Addis Ababa University,
P.O. Box 1176 Addis Ababa, Ethiopia; 3Ethiopian Health and Nutrition Research Institute, Addis
Ababa, Ethiopia
respectively. From the Batu mountain chain to the north, the land falls undulating towards the plains in
the southern part of the Awraja.
The Awraja has great natural potentials such as the Arena Forest Reserve, large livestock and game
resources, big rivers, irrigable basins and unharnessed minerals prospects. It is said that the GenaleDawa basin alone has a gross potential of 600,000 ha of irrigable land (7). Inspite of these potentials,
however, the area remains isolated and under-developed. It is characterized by poor communication,
inadequate health services and low level of living condition. There is only one health center located in
Mena town, the capital city of the Awraja (Fig. 1). The rest of the Woredas (district) are served by
only 9 clinics that are poorly staffed and equipped. Safe water supply and sanitary facilities are either
poor or absent.
The inhabitants are traditionally nomads or semi-nomads dependent on livestock. About half of the
land is still used for grazing and only about 0.6% has been cultivated, most of which is limited, by and
large, to Mena, Berbere and Oborso districts (8). Recurrent drought, specially in the lower agroclimatic zone, has repeatedly affected the inhabitants forcing many into temporary shelters.
The Norwegian Church Aid for Ethiopia together with the Ministry of Health and the Ministry of
Agriculture has initiated an integrated rural development program. This is essentially a PHC program
which includes safe water supply through drilling deep wells and spring development and a package of
agricultural inputs to the rural communities.
There now is a new road connecting the town of Mena with Goba to the north and with Neghele
Borana to the west (Fig. 1). A commercial Bank is already in operation while a modest hydroelectric
power construction was underway as major signs of government investment in the area.
Parasitological Examination: Stool and urine samples were obtained from school children and
residents in accessible communities where
Observation of blood microfilariae 185
────────────────────────────────────────────────────────────
Figure 1: Map of Delo Awraja and sites of parasitological and malacological survey
NCA is undertaking and/or planning to undertake DDP. About ten percent of the population in each
school or community was selected by systematic random sampling. The stool specimens, about 1 gm
of faeces from each individual, were preserved in screw-capped vials pre-filled with 10% formalin
which were then processed and examined by the Ritchie’s concentration technique (9).
Examination of urine for schistosomiasis haematobium was undertaken since most of the people are
nomads and might have travelled to lowlands such as Delo Odo or Somalia where the disease is
suspected or known to occur.
Urine samples were examined on the spot by filtration method (10) for individuals living at lower
altitudes. The examination was done only for one-fifth of those who provided stool except in
Melkaaman and Genale where more people were included because of the lower elevation and history
of peoples’ travel to Somalia. Reagent strips were also used to detect haematuria and proteinuria due
to schistosomiasis haematobium (10) in the urine collected.
186 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 1: Results of stool examination by altitude, study subjects and community
Community
Alt. (m)
1100
1200
1200
1300
1350
1350
Study
Subj.
R
R
S
R
R
R
No.
Exams
99
148
140
159
138
199
Percent positive
Sm
Al
0
2
2
0
11
15
8
6
1.5
0.7
6.5
11
Melkama
Gomgoma
Genale
Burkitu
Erba
Mena
Tt
3
3
14
21
13
14
Ho
0
16
27
56
27
29
Ts
0
3
14
4
3
7
St
0
1.4
4
5
0
3
Others
0
0
11
5
0.7
1.0
Meda
Angetu
Soda Wolmel
Uko Negesso
Oborso
Kerjule
Ela Bidre
Bidre
Berisa
Total
1450
1500
1500
1600
1600
1600
1650
1700
1700
-
R
S
S
R
S
R
R
S
R
-
62
103
27
184
77
213
102
63
73
1787
48
2
7
0.5
0
3
2
0
3
5
0
31
15
21
64
3
1
25
3
13
2
36
67
54
62
16
5
8
8
30
3
1
11
11
5
3
4
8
11
16
0
2
0
11
17
1.4
0
0
0
3
0
7
4
2
2
0
1
4
4
3
2
67
26
49
84
0
0
40
0
18
Sm = Schistosoma mansoni
R=Residents
Al = Ascaris lumbricoides
S=Students
Tt = Trichuris trichiura
Ho = Hook worm species
St = Strongyloides stercolaris
Ts = Taenia saginata
Other = Faciola species, Enterobius vermicularis and Hymenolepis nana
Malacological Survey: Water bodies, including rivers, streams and temporary water pockets, were
surveyed for snails, using a standard scoop. Furthermore, the altitude and water temperature and pH
were also recorded.
The snails encountered were identified into species using appropriate keys; checked for
schistosomal infection by dissection technique; and recorded with the name of the water body
surveyed.
Result
Stool Examination: Altogether 1787 individuals provided stool samples in 15 communities.
Schistosoma mansoni infected individuals were found in 80% (12/15) of the communities studied. In
five communities the prevalence exceeded 5%, reaching 48%, 11%, 7% and 6.5% in Meda, Genale,
Burkitu, Sodo Wolmel and Mena respectively (Table 1).
Other intestinal helminth parasites detected include: Ascaris lumbricoides, Tricuris trichiura,
hookworm species, Taenia saginata, Strongyloides stercolaris, Hymenolepis nana, Enterobius
vermicularis and Fasciola hepatica (Table 1). The most prevalent was Ascaris lumbricoides, reaching
as high as 84% at Oborso followed by hookworm (67%) at Sodo Wolmel and Trichuris trichiura
(64%) at Oborso. The prevalence of Taenia Saginata ranged from 1% at Angetu to 14% at Genale.
The wormiest communities, particularly for Ascaris, Trichuris, hookworm and Strongyloides, are
located at altitude of 1500-1600 meters above sea level. Ascaris lumbricoides, Trichuris Trichiura
and hookworms appear to be coexsisting in communities where they are prevalent.
Urine Examination: Five hundred and thirty seven urine specimens were examined for Schistosoma
haematobium infection by filtration and reagent strip methods. Only two individuals at Melkaaman
(male subjects of 7 and 16 years of age) were found passing the parasite egg in their urine. The
intensity of infection was low in both individuals settling at only 2 and 6 eggs per 10 ml of urine.
Analysis of their birth places and history of
Observation of blood microfilariae 187
────────────────────────────────────────────────────────────
Table 2: Results of urine examination by filtration and reagent methods
community
Alt (m)
Study
Subject
No. Examined by
Urine
Reagent
filtration
Strip
Melkaman
Gomgom
Genale
Burkitu
Erba
Mena
Meda
Angetu
Soda Wolmel
Uko Negesso
Oborso
Kerjule
Ela Bidre
Bidre
Berisa
Total(%)
1100
1200
1200
1300
1350
1350
1450
1500
1500
1600
1600
1600
1650
1700
1780
-
R
R
S
R
R
R
R
S
S
R
S
R
R
S
R
-
77
29
140
33
28
43
12
21
5
38
16
42
27
13
15
539
Haem = haematuria;
R = Residents;
77
29
140
33
52
43
12
21
5
38
16
42
27
13
15
539
No (%) positive
Urine
Reagent strip
filtration
haem.
2(2.6)
5(10)
0
1(2)
0
2(4)
0
2(4)
0
25(50)
0
3(6)
0
0(0)
0
1(2)
0
0(0)
0
1(2)
0
1(2)
0
2(4)
0
3(6)
0
2(4)
0
2(4)
2(0.4)
2(0.4)
protein
2(12.5)
0(0)
2(12)
3(18)
2(12.7)
2(12.5)
1(6.2)
1(6.2)
0
0
0
0
1(6.2)
2(12.5)
0
16(3.0)
S = Students
travel showed that they had been in Somalia for over 5 years. The prevalence of haematuria and
proteinuria detected by the reagent strips were also very low except at Erba where 50% was positive
for microhaematuria (Table 2).
Malacological Finding: Of the 9 fresh water bodies surveyed only 3 contained one or another kind of
snails. The snails recovered included: Biomphalaria pfeifferi (at Shewe and Burkitu streams), Bulinus
forskalii (at Genale River) Lymnea natalensis (at Shewe and Wolmel Rivers) and small planorbids (at
Shewe River) (Table 4). The environmental characteristics (water temperature and pH) were within
normal values for breeding of the snails.
Discussion
It is not surprising to find that intestinal parasites are highly prevalent and widespread in the Delo
Awraja since in Ethiopia and wherever poverty, ignorance, lack of safe water supply, inadequate
sanitation and poor hygiene prevail such parasites are expected to be ubiquitous. Of interest is,
however, the discovery of endemic localities of intestinal schistosomiasis in this hithertofore isolated
part of Ethiopia. Previous survey of few communities in Bale Region (2) led to a general thinking that
the Region would be free from schistosomiasis.
It is possible that local transmission of Schistosoma mansoni is taking place in the five communities
(Meda, Genale, Burkitu, Sodo Wolmel and Mena) where the disease prevalence exceeded 5%. The
high prevalence (48%) observed in Meda where agricultural development is underway by Norwegian
Church Aid programme is of interest. Its proximity to the Ganale basin where large scale irrigation
scheme is envisaged also gives adequate reason for worry in terms of disease spread. In general, the
finding of positive patients and snail intermediate host of intestinal schistosomiasis in a number of
communities warrants initiation of schistosomiasis surveillance and/or control measures in Delo
Awraja, particularly if the development plans have to be realized.
Urinary schistosomiasis seems to be absent in the surveyed communities at least at the moment as
indicated by the results of both urine examination and reagent strips.
188 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 3: Findings of the malacological survey in Delo Awraja
Temperature Waterbody Air
Alt (m)
Water
Genale River
1200
30
Wolemel River
1200
29
Burkitu Stream
1300
20
Gobale River
1500
24
Ridimo River
1500
22
Angetu Stream
1500
22
Shewe River (Meslo)
1500
19
Kerjule Stream
1600
18
Oborso Stream
1600
16
1 = Biomphalaria Pfeifferi
2 = Bulinus forskalii
pH water Snail species
20
25
20
24
22
22
19
18
16
7
6
6
6
6
7
7
7
7
2
Negative
1,3
Negative
“
“
1,3,4
Negative
“
3 = Lymnaea natalensis
4 = Small planorbids
However, the lower portion of the Awraja bordering Somalia need to be surveyed to confirm the
absence of urinary schistosomiasis. The two positive patients diagnosed at Melkaaman have history of
travel to Somalia where urinary schistosomiasis in known to be endemic (11). Hence, they may be
harbouring the Somalian strain of S. haematobium. Since the community is located at a lower altitude
(about 1000 m) and is close to a permanent water body (Welmel River) the presence of infected
individuals if coupled with the presence of the right snail species in the area may lead to the
establishment of S. haematobium transmission. Refugees returning from Somalia should therefore be
screened for S. haematobium infection and positive patients treated to avoid any possible introduction
of a foreign parasite strain into this ecologically receptive area.
The relatively high positivity rate (50%) of microhaematuria among those examined at Erba could
not be explained. Perhaps menustrating female subjects were examined although other causes of
microhaematuria, which warrants further investigation, should not be ruled out.
The high prevalence of Ascaris and Trichuris at intermediate elevations (around 1500 m.a.s.l) in the
area is quite typical of similar rural ecological settings in Ethiopia (12). However, the dominance of
hookworm infection demands special attention since it causes chronic blood loss and depletion of the
body’s iron store, leading to iron-deficiency anemia particularly in children (13). Future study should
aim at identifying the species of hookworm occurring in the area. According to Shibru Tedla and
Leykun Jemaneh (14) and Leykun Jemaneh and Shibru Tedla (15) both Ancylostoma duodenale and
Necator americanus may occur as sympatric species. Periodic de-worming of children with
broadspectrum anthelminthic drugs coupled with health education on personal and environmental
sanitation and improvement of water supply should be given high priority to control morbidity due to
the soil-transmitted intestinal helminth parasites.
Acknowledgement
This study was supported by the Norwegian Church Aid of Ethiopia (NCA/ET) as part of a
comprehensive health survey in Delo Awraja. The moral and Technical support of the technical staff
of the Institute of Pathobiology, Addis Ababa University, is also highly appreciated.
References
1. Kloos H, Hailu B, Lo CT, Teklemariam A and Shibru T. Schistosomiasis in Ethiopia. Soc Sci
Med 1988;26:803-827.
2. Institute of Pathobiology. Progress Report Number 4. 1986;4:116.
3. Kloos H. Water resources development and schistosomiasis in the Awash Valley, Ethiopia. Soc
Sci Med 1985;20:609-625.
4. Simon PE, Assefa N, Furu P. Intestinal schistosomiasis among children in a village of Wonji Sugar
Observation of blood microfilariae 189
────────────────────────────────────────────────────────────
Estate, Ethiopia. East Afr Med J. 1990;67:532-537.
5. Awash T, Fletcher M. A parasitological and malacological survey of schistosomiasis manosoni in
the Beles Valley, northwestern Ethiopia. J Trop Med Hyg 1990;93:12-21.
6. Central Statistical Office. Addis Ababa, Ethiopia. 1994.
7. Water Resources Development Agency (WRDA). Three Year plan (1986-88). Addis Ababa,
Ethiopia. 1986.
8. Ministry of Agriculture. Unpublished data. 1980.
9. Ritchie LS. An ether sedimentation technique for routine stool examination. Bull US Army Med
Dep. 1948;8:326-9.
10. Feldmeier H, Doehring E, Daffalla AA. Simultaneous use of sensitive filtration technique and
reagent strips in urinary schistosomiasis. Trans Roy Soc Trop Med Hyg. 1982;76:416-21.
11. World Health Organization. Atlas of the global distribution of schistosomiasis. WHO:
1987;257:62.
12. Shibru T, Teklemariam A, Hailu B, Lo CT. Intestinal helminthiasis in Ethiopia. In: Proceedings
of a Symposium on Human Schistosomiasis in Ethiopia (eds).
13. Stephenson LS. Helminth parasites, a major factor in malnutrition. Wld Hlth Forum 1994;15:169171.
14. Shibru T, Leykun J. Distribution of Anchylostoma duodenale and Necator americanus in Ethiopia.
Ethiop Med J. 1988;23:149.
15. Leykun J, Shibru T. The distribution of Necator americanus and Ancylostoma duodenale in school
populations, Gojjam and Gonder Administrative Regions. Ethiop Med J 1984;22:87.
Original article
Determinants of infant and early childhood mortality
in a small urban community of Ethiopia: a hazard
model analysis
Assefa Hailemariam1 and Makonnen Tesfaye 2
Abstract: By applying Cox's proportional hazard model regression analysis to data collected using a
retrospective survey conducted in Sebeta, a town 25 Km west of Addis Ababa, the capital city of
Ethiopia, the paper examines the factors impinging on the survival of infants and children between 1 3 years of age. It is shown that for higher order births (more than 5), for births to young women (under
20 years of age), and for those to older women (more than 34 years of age), the risk of dying at infancy
is higher. The risk of infant mortality is also high for births with short previous birth intervals. In fact,
the length of the previous birth interval is found to be the single most important factor affecting the
chances of survival during infancy. It is further shown that education of mother, occupation of father,
household income, source of drinking water, availability of latrine, and survival status of older sibling
have direct effect on infant mortality. Among these, source of water and availability of latrine are
identified as having significant effects on infant mortality even after controlling for the effects of other
variables. During early childhood, however, the effects of age at maternity, birth order and preceding
birth interval becomes trivial. Following birth interval appears to have a strong effect on the chances
of survival during early childhood. Household income, religion and survival status of the previous
sibling are found to have significant effects on early childhood mortality. The findings provide solid
ground to support strategies to broaden MCH/FP services, environmental health and income generating
scheme to reduce the risk of death for infants and children. [Ethiop. J. Health Dev. 1997;11(3):189200]
Introduction
Infant and child mortality have long been used as indicators of the level of socio-economic
development of a nation. Most of the developed countries have registered low levels of infant and
child mortality rates. In the developing countries, particularly, in sub-Saharan Africa, although
significant achievements have been made, infant and child mortality still stand at high levels. In
Ethiopia, in 1990, about 10% of the infants died before their first birthday and more than 15% of the
children died before their fifth birthday (1).
Also various studies (2-9) show high levels of infant and childhood mortality in the country. This high
level of mortality may be associated with demographic, socio-economic, and environmental factors
such as ethnicity, housing condition, crowding, availability of latrine, and early termination of breast
feeding (9). However, most of the studies conducted so far examined the effects of the socio-economic
and environmental variables and did not consider the inpact of the demographic factors (age at
maternity, birth interval, birth order, and survival status of preceding sibling mainly due to the general
lack of data. Some studies (2, 3, 6) identified such variables as sex of child, mother’s place of
residence, education, religion, ethnicity, marital status, income, and environmental sanitation as
important determinants of infant and childhood mortality. Others (7, 9) identified additional factors
such as care at delivery, duration of breast feeding, child nutrition, and paternal education.
______________________________________
1
From Demographic Training and Research Center, Institute of Development Research, Addis Ababa
University, P. O. Box 1176, Addis Ababa, Ethiopia, and 2Population Studies and Analysis Center,
Central Statistical authority, P.O. Box 1143, Addis Ababa
Other factors such as maternal age, birth order, and birth interval are also shown to have significant
Determinants of child mortality 191
────────────────────────────────────────────────────────────
impact on the chances of infant and child survival (10, 11). Cultural values and norms, as well as
community level or environmental variables among others, are also known to influence the chances of
infant/child survival (12, 13). However, in Ethiopia, indepth studies dealing with demographic, socioecomonic and environmental correlates of infant/child mortality are lacking. Identifying the socioeconomic, environmental and demographic determinants of infant and early childhood mortality is
believed to assist in the design of programs for lowering the risk of high infant and child mortality that
prevails in the country. In view of the strong association between infant and child mortality and
fertility, lowering infant and child mortality may subsequently lead to fertility decline (15, 16).
This study, therefore, attempts to narrow the gap in our present understanding of the effect of
demographic, socio-economic and environmental factors on infant and early childhood mortality by
controlling for social and demographic characteristics of the mother and the child.
The objectives of this paper are : (i) to analyze the separate and combined effects of the
demographic variables (age at maternity, birth order and length of the preceding and subsequent birth
interval) on infant and early childhood mortality in the absence and presence of other socio-economic
and environmental factors and (ii) to determine the relative importance of each of the demographic
variables.
Methods
Data: The study is based on data obtained from Sebeta Town, about 25 Km west of Addis Ababa,
between June and July 1992. All females of reproductive age who were permanent residents of the
town at the time of the survey were included in the study. Consequently, a total of 3,140 women living
in 2,134 households were interviewed. Female interviewers conducted the interview using structured
questionnaires. Two sets of questionnaires, namely, the household and individual questionnaires were
used. The household questionnaire was used to develop and collect inforrmation on socioeconomic
and environmental background characteristics of the households, such as source of drinking water, type
of toilet facilities, and availability of radio in the household.
The individual questionnaire was used to obtain information on women's characteristics, such as
age, marital status, education, work status, occupation, religion, ethnicity, husband's background
characteristics, household income, the number of children ever born, the number of children dead,
births in the last 12 months, birth history of each live birth, breastfeeding status, and duration of breast
feeding.
Since age in months is an important requirement for studies which include birth intervals, children
with missing month of birth were dropped from the analysis. The differential in reporting month of
birth among sub-groups of the women was examined and found to be insignificant, implying that the
exclusion of these from the analysis would not introduce serious bias in the study. Although children
whose mothers have died have higher mortality risks, in this study, it is assumed that mortality among
children whose mothers have died is similar to that of children whose mothers are alive (17).
Data reporting was considered to be poor for older women (those aged 40 years and older)
compared to younger women. Reporting of children ever born and children dead was better for the
most recent births (10-15 years prior to the survey). However, age data indicated the presence of age
heaping, age shifting and age under-reportings especially for older women. In order to minimize the
magnitude of error and bias in the results, data analysis is limited to births that occurred in the 15 years
preceding the survey. Of 6405 live births to all women, only 5385 (3, 114 births for infant mortality
and 2, 271 for early childhood mortality) were included in the analysis due to the exclusion of births
with missing month of birth, and births which occurred before 15 years prior to the survey date.
Statistical methods: Several models are available for handling studies of this kind (linear
regression, logistic regression, etc.). However, such models do not handle censored cases (cases with
incomplete exposure). The proportional hazard model, first developed by Cox (17), is a well known
model for handling censored cases. It is a special case of the more general survival model in that it
192 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
combines aspects of the life table and regression analysis and allows the formulation of relations
between a set of covariates and the survival function as in conventional multiple regression. It is used
in the analysis of survival data when mortality risks vary among individuals.
The model assumes that, at a given age (or duration since the start of life), the force of mortality is a
constant (specific to that age) multiplied by a proportional factor which is determined by the
characteristics of the individual. Specifically, at duration d for an individual i with a known set of
explannatory variables Zi=(z1,z2,...zz) the hazard (risk) function is given by λ(t;Z)=λ(d)expß’Zi where
λ(d) is the baseline hazard; and ß=(ß1,ß2,...ßk) is a vector of parameters. The factor expßj is the relative
risk associated with the jth explanatory variable. Parameters are estimated using maximum likelihood
procedure. This is the model employed for examining the determinants of infant and child mortality in
this study.
STATA, a standard statistical software, was used for data analysis. The data required for estimation
has the form (ti,δi,Zi), where ti represent time to death or censorship since entry into the study (in this
case, number of months child stayed alive); δi indicates whether child was excluded from the study
population due to death or censorship. For infant mortality analysis, δi=0 or 1, depending on whether
the infant died before age one or survived to age one. For child mortality analysis also, δi=0 or one
depending on whether the child died while aged 1-3 years or survived to age three. Zi is a vector of
explanatory variables.
The demographic variables included in the model are birth order, age at maternity, preceding birth
interval, and following birth Interval. Birth order indicates the order of birth of the index child and it is
treated as a categorical variable in the model with three categories representing births of order 2 to 3,
4, and 5 or above. Births of order 4, indicated the lowest death rate and are used as a reference
catagory.
First order birth are known to be at a higher chance of mortality, mainly due to the detrimental
biological influence of being the first birth. It appears that toxaemia of pregnancy is more frequently
associated with first births (19). The increased risk of mortality among first born children may also be
due to the young age of mother at their birth. However, in this study, since the multi-variate analysis
includes preceding birth interval as one of the main explanatory variables, first births are excluded
because they are preceded by no one.
Age at maternity refers to the age of the mother at the birth of the child and it is also grouped into
three, namely, less than 20 years (early), 20-34 years (intermediate), and 35 years or more (old).
Maternal age of 20-34 is used as the reference category. Preceding birth interval is the length of the
interval between the birth of the index child and its older sibling. Three categories of preceding birth
interval, (less than two years, 2-3 years, and 3 years and more) are included. The reference category is
the 2-3 years birth interval. ‘Following birth interval’ refers to length of time between the birth of the
index child and the one following it. This variable is used only in the early childhood mortality
analysis and it is treated as a dummy variable (1 if<18 months, 0 therwise).
A number of demographic, socio-economic, cultural and environmental variables expected to have
close relationship with infant/child mortality were used as control variables. These variables were
included to see if they modify the effects of maternal factors on infant/child mortality. These include
survival status of previous birth, parental education (i.e., educational levels of mother and father),
occupation of father, work status of mother, place of work of mother, mother’s ethnic back ground,
religion of mother, source of water, availability of latrine, and household income. Place of birth of
mother (urban or rural) and sex of the child were not used here because infant/child mortality differed
very little by place of birth of mother or by sex of child.
All of the control variables were treated as categorical. Three categories of education were used for
father's and mother's education: no schooling (illiterate and informal education), elementary (grades 16), and high (junior and above) . The educational category 'high' was used as a reference category.
Three categories of father's occupation were employed: professional-secretarial, sales-service, and
Determinants of child mortality 193
────────────────────────────────────────────────────────────
others (production workers, laborers and all others). The first group, i.e., the professional group was
used as a reference category. Two groups of work status of mothers were considered: working and
non- working. Household income was categorized into three: less than Birr 100, Birr 100-299, and
Birr 300 or more. The category ‘Birr 300 or more’ was used as a reference category.
Ethnicity and religion of mother are the two
socio-cultural variables used as control variables. The two major ethnic groups in the study area,
that is, Oromo and Amhara, and all other groups together, were compared. The Amhara had the
lowest infant/child mortality and this group was used as a reference category. Religion of mother was
grouped into Orthodox, Protestant/Catholic, and Muslim. Followers of Orthodox Christianity had the
largest number of cases. This group was used as a reference category.
Environmental contamination was controlled using source of water and availability of latrine.
Private pipe, shared pipe, public tap, and well/river were the four categories for source of water.
Private pipe was used as a reference category. Latrine was categorized as private, shared and none.
Those with private latrine were used as control category. Survival status of older sibling at the time of
survey was included to control for child mortality pattern in the family. In the analysis, the effects of
variables of main interest were first considered and then the effect of these variables was examined
after adjusting for the effects of the control variables.
Results
Population Characteristics: At the time of the survey (June 1992), the Town had a total pupulation
of 11624. This excludes those living in unconventional living quarters. Of these, 5436 were males and
6188 were females giving a sex ratio of 88 males per 100 females. More than half of the population
(55.1%) belonged to the economically active age group. The dependent population (children under 15
and adults 60 yrs and above) constituted 44.9%. There were more females in the economically active
age group compared to males (57.3% against 52.5%). However, most women were not economically
active. At the time of the survey, nearly three quarters of the women had no work other than household
chores. Women of reproductive age (15-49 years) constituted a little over one quarter of the total
female population (Table 1).
Table 1: Distribution of the population of Sebeta town by broad age groups and sex
Age Group
0-14
15-59
60+
Total
Male
No.
2334
2852
250
5436
%
42.9
52.5
4.6
00
Female
No.
2364
3548
276
6188
%
38.2
57.3
4.5
100
Total
No.
4698
6400
526
1624
%
40.4
55.1
4.5
100
Table 2 presents some socio-economic background characteristics of the respondents.
Data on marital distribution show that less than 40% of women in their reproductive years were in
marriage and 15.8% were either divorced or widowed while 46.5% were never married at the time of
the survey. Although the majority of the women were literate, only 26.8% had attained secondary or
higher education. The educational distribution of currently married husbands on the other hand shows
that men had better education than women. While the percentage of literate among the female
population was 78.5, the corresponding percentage for males was 90.6. In all the other educational
categories also the proportion for males is much higher than that for women.
194 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 2: Percentage distribution of woman aged 15-49 years by selected socio-economic characteristics
Variable
Distribution
Percentage
N
Marital Status
Single
46.5
1460
Married
37.8
1186
Widowed
5.4
169
Divorced/Separated
10.3
325
Education of Woman
Illiterate
Read and Write
Elementary
Junior
Secondary
Above Secondary
Not Stated
Education of Husband
21.5
16.4
19.4
15.5
25.2
1.6
0.4
674
516
608
488
792
51
11
Illiterate
Read and Write
Elementary
Junior
Secondary
Above Secondary
Not Stated
Work Status of Woman
9.3
18.3
23.7
15.8
22.5
8.7
1.7
110
217
281
188
268
103
20
Working
Not Working
Place of Birth
27.0
73.0
847
2292
Urban
Rural
59.2
40.2
1857
1263
A large majority of the respondents had no work other than household chores, only 27% reported as
working at the time of the survey. Nearly 60% had urban origin while 40% were migrants to the town
from rural areas.
Infant mortality: In this section, we first present the level of infant mortality in the town before
applying the hazard model. Table 3 presents the levels of infant and child mortality estimated using
the Trussell version of the orginal Brass technique of estimating infant and child mortality from
information on children ever born and children surviving classified by age of mothers (20). It may be
observed that infant mortality rate for the town at the time of the survey was about 86 per 1000 live
birth. This indicates that the level of infant mortality was relatively lower in the town compared to
other urban areas in the country at that time. For instance, the level of infant mortality estimated for
urban Ethiopia using the 1990 National Family and Fertility survey was 93.6 per 1000 live births (1).
The 1994 Census for Urban Oromyia resulted an adjusted infant mortality rate of 93 per 1000 live
births and 78.5 for Addis Ababa (21,22).
Table 3: Average infant, child(1-4) and under-five mortality per 1000 live births, Trussell, north model; Sebeta, 1992
Rates per 1000 live births
Infant mortality rate (1q0)
86
Child mortality rate (4q1)
57
Under five mortality rate (5q0)
138
Tables 4 through 8 show results of the proportional hazard model applied to the data. Table 4
presents seven hazard models of infant mortality. The first three models estimate the univariate effects
of variables of main interest, namely, birth order, age at maternity, and preceding birth interval.
Models 4-6 include two variables at a time, and model 7 includes all three variables simultaneously
Determinants of child mortality 195
────────────────────────────────────────────────────────────
and is considered the full model. In this and subsequent tables the effects of the various covariates are
expressed as relative risks.
The effects, or the relative risks, are calculated as expßij, where ßij is the estimated coefficient for the
ithcategory of variable j. For a given variable, relative risks compare mortality risks for different
categories with that of the reference category. For example, in the full model (model 7) of Table 4, the
relative risk of 1.781 for age at maternity of less than 20 years means that mortality rate for infants
born to mothers aged less than 20 years is 1.781 times higher than that for infants born to mothers in
the reference category, that is,
Table 4: Relative effect of variables of main interest (maternal factors) on infant mortality.
Variables
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
Model 7
Birth order
2-3
1.4189
1.216
1.415
1.254
4
1.000
1.000
1.000
1.000
5+
1.60*
1.603*
1.553*
1.544*
Age at maternity
<20 years
1.844
2.016
1.697
1.781
20-34
1.000
1.000
1.000
1.000
35+
1.272
1.121
1.039
1.211
Preceding birth
Interval
<2 years
1.755
1.731
1.668
1.644
2-3 years
1.052
1.004
1.341
1.018
>3
1.000
1.000
1.000
1.000
Note 1 In this and the following tables, the symbols *, ,  indicate level of significance (t test, two side) at 10, 5 and 1 percent level,
respectively
Note 2 Model 1-3 estimate univariate effects; model 4-6 include two variables at a time while model 7 (full model) contain all variable
simultaneousely.
women aged 20-34 years. The symbols indicating level of significance refer to the departure of the
relative risk from unity, the value of the relative risk for the reference category.
The single-factor effect (models 1-3) indicates that, when the effect of other factors is not taken into
account, births of order 2-3, and higher (5 and above) seem to have higher mortality risks compared to
births of order 4. Births of order 2-3 had 42% higher risk of dying while those of order 5 or higher had
60% higher risk compared to births of order 4. However, the effect is not statistically significant.
Infants born to mothers under 20 years of age had a significantly higher risk of dying (84%)
compared to those born to mothers in the reference category (aged 20 - 34 years). Infants born to older
mothers (35 years or more) were also at a higher risk of dying (27%). A birth within two years of the
birth of the index child had a significantly higher risk of dying at infancy (at 0.01 level) compared to
preceding birth interval of more than three years. Whereas, a birth that occurred 2-3 years after the
birth of the index child had no significant impact on infant mortality compared to that born after an
interval of more than three years (Table 4).
The two-factor effect (models 4-6) indicates that the effect of births of order 2-3 compared to births
of order four declined when age at maternity is controlled. However, controlling the effect of age at
maternity or the length of preceding birth interval did not alter the risk of dying at infancy for higher
order births (birth order 5 or higher). This suggests that the relatively higher risk of dying for infants
of order 2-3 is mainly due to the young age at maternity. Controlling for the effect of the preceding
birth interval appears to have reduced the effect of maternal age on infant mortality. Birth order and
preceding birth interval affect the chance of infant survival independently of each other.
Results in the full model (model 7) show that the independent effect of the variables persisted even
after making adjustment for the effects of any two of the other variables. Although the magnitude of
the effects are reduced, they are still significant. After controlling for the effect of age at maternity and
length of preceding birth interval, births of higher order have 54% higher risk of dying before the first
birthday than births of order 4. When the effects of birth order and length of preceding birth interval
are controlled, children born to mothers under 20 years of age have 78% higher risk of dying compared
196 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
to those born to mothers aged between 20 and 34 years. Children born within two years of the
preceding birth have 64% higher chance of dying at infancy, even after controlling for age and birth
order, compared to those born after three or more years of the pervious child.
Log likelihood ratio tests are used in order to assess the significance of the contribution of variables
as well as to determine their order of importance. Table 5 presents the change in -2 x log likelihood (2LL) when a backward selection procedure is employd. In step 1 ‘-2LL’ of a model containing all of
the demographic (maternal) factors is presented. In step 2 a model containing maternal age and birth
interval is compared to the ‘Full model (a model containing all variables). The increment in the
goodness of fit of the model (χ2= difference in -2LL) indicates that birth order has a significant impact
on infant mortality rate at 10 percent level. Similarly, in step 3 and 4 models which exclude,
respectively, maternal age and birth interval are compared to the full model. The result shows that age
at maternity and length of preceding birth intervals have significantly increased the goodness of fit of
the model.
Table 5: Log likelihood ratio test of significant of variables: infant mortality
step
Model
-2* Log
X2
df
p
Likelihood
1
Full Model
2277.46
2
ß2=0
2280.41
2.95
1
0.05<p<0.1
3
ß2=0
2282.47
5.01
1
0.025<p<0.01
4
ß3=0
2285.76
8.3
1
p<0.0005
A Comparison of -2 x log likelihood of the models shows that the preceding birth interval is the
most important variable, followed by age at maternity in affecting the chances of survival at infancy.
Birth order has the least but statistically significant effect on chances of survival at infancy.
Table 6 presents nine hazard models after controlling for certain variables. Model 1 contains the
effects of birth order, age at maternity and length of the preceding birth interval. Models 2-5 include
the socio-economic variables (mother’s education, work status and household income). Survival status
of previous sibling is also included in addition to the variables of main interest. In models 6 and 7, the
variables are entered in blocks, ethnicity and religion in model 6, and source of water and availability
of latrine in model 7. Model 8 is the full model and model 9 considers only control variables.
Education and occupation of father are not entered at this stage because the numbers of cases were not
the same for the two variables and this made comparison rather difficult. These will be considered
later in the paper. Place of work is also dropped from the model because of strong colinearity with
work status of mother.
The results show that none of the control variables included in the analysis appear to explain the
effect of higher birth order on increased infant mortality. Controlling maternal education, availability
of latrine and household income, indicates marginal change on the effect of birth order. The other
variables do not appear to affect the impact of birth order on the survival chances at infancy. The
higher chance of dying during infancy for higher order births lost its statistical significance after
controlling for the effect of maternal education.
Model 4 shows that some of the effect of young maternal age on infant survival is explained by
household income. The effect of young maternal age is stronger in low income households. Infants
born to mothers in households whose income is less than Birr 100 had higher risk of dying compared
to those born to mothers living in households with higher income. Controlling effects of other
variables did not change the effect of age at maternity on infant mortality.
Analogous to birth order, statistical control of variables considered here shows that the effect of the
length of preceding birth interval is invariant under most conditions. Only survival status of previous
sibling seem to explain a little of the effect of preceding birth interval on infant mortality. This and the
bi-variate results show that the negative impact of short preceding birth interval appears to increase if
previous child has died. The change in the effect of preceding birth interval is small when all other
Determinants of child mortality 197
────────────────────────────────────────────────────────────
variables are controlled simultaneously.
Table 6: Relative risks of maternal factors when control is made for some background variables, infant mortality: Hazard model.
Variables
Birth order
2-3
5+
Age at maternity
Model 1
Model 2
Model 3
1.254
1.537*
1.300
1.467
1.244
1.559
<20 yrs
35+ yrs
Preceding birth
1.781
1.211
1.549
1.840
1.644
1.018
1.633
1.027
Model 4
Model 5
Model 6
Model 7
Model 8
1.267
1.609*
1.269
1.498
1.254
1.519
1.208
1.614*
1.215
1.634
1.800
1.208
1.6432*
1.179
1.661
1.223
1.807
1.177
1.760
1.288
1.679
1.302
1.644
1.107
1.707
1.041
1.558
1.018
1.622
1.001
1.613
0.903
1.531
1.012
Model 9
interval
<2 yrs
2-3 yrs
Education of
mother
No school
Elementary
1.548*
1.020
1.043
1.548
1.275
1.313
1.231
1.237
Work status of
mother
Working
1.312
Incom(Birr)
<100
2.347
1.399
1.384
100-299
1.678
1.165
1.159
1.648
1.901
1.408*
1.094
1.057
0.995
1.004
1.049
1.174
1.672
1.483
0.935
1.446
0.914
1.407
2.801
1.361
2.455
0.707
1.301
2.409
0.761
1.439
1.686
2225.0
1.303
1.473
2198.8
1.308
1.498*
2214.0
Preceding sibling’s
survival status
Dead
Ethnicity
Aroma
Others
Religion
Muslim
Prot/Catholic
Water source
Pipe-com
Public tap
Well/river
Latrine
Common
None
-2LL
2.106
2274.5
2257.7
2275.0
2258.6
2266.3
2271.1
Note 1. Reference Categories are not shown.
Note 2. Model 1 estimate effects of main variables; Model 2-4 in addition, include one background variables at time. Model 5 and 6 add background
variables in block on top of main variables, while model 8 (ful model) contain main varables and background variables simultaneously. Model 9 considers
the effect of bacground variabls in group.
When the control variables are considered, women’s education, other than its effect through
maternal factors, appears to have an independent effect on the chances of survival at infancy.
However, when either source of water, latrine or household income is added into the model, education
lost its role as an important variable in explaining the variability in infant mortality. Ethnicity, source
of water, latrine, and household income influence the risk of infant survival by operating through
mechanisms other than maternal factor. However, the importance of these variables in affecting the
chances of survival at infancy declines with the introduction of maternal education into the model. For
instance, the magnitude of the effect of ethnicity and income declined when maternal education was
198 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
introduced and vanished when all other variables were simultaneously controlled.
Once the effects of maternal factors are taken into accont, work status of mother and religion are not
found to be important determinants of the survival status of infants. The single factor effect, however,
indicates that infants born to working mothers had higher mortality risks relative to non-working
mothers. This may be because working women have less time for infant feeding (including time for
breastfeeding) and caring compared to non-working women.
Determinants of child mortality 199
────────────────────────────────────────────────────────────
Another variable which has direct and significant effect on infant survival is the survival status of
the preceding sibling. When the preceding birth has died, the chance of dying of the index child is
about two times that of a child whose older sibling has survived. However, when the effects of other
variables are controlled, this high effect is reduced but still remained significant.
The effect of maternal and socio-economic factors on the risk of infant survival was also considered
by restricting the analysis to currently married mothers. This was done so as to control for some of the
fathers background characteristics. Table 7 presents the hazards of main variables along with
education, and occupation of father.
Note that the significant effect of young age at maternity (under 20) is lost when the analysis is
restricted to currently married mothers. When father's education is controlled, the effects of birth order
and preceding birth interval on the chances of infant survival did not change, but that of age at
maternity increased. Father's education appears not to have any significant impact on infant survival.
Table 7: Relative effects of maternal factors, when father’s background is controlled, on infant mortality
Variables
Model 1
Model 2
Model 3
Birth order
2-3
1.833*
1.8507*
1.8765*
5+
2.111
2.0883
2.0256
Previous birth
Model 4
1.879*
2.043
interval
<2 years
2-3 years
Age at ,atermotu
1.805
0.981
1.809
0.969
1.775
0.957
1.783
0.956
<20 years
>34 years
Father’s education
1.5902
1.097
1.6524
1.318
1.547
1.976
1.581
1.061
No schooling
1.317
1.057
Elementary
0.961
0.791
Occupation of father
Sales-service
2.254
2.341
Prod-others
1.936
1.936
-2LL
1663.3
1642.1
1640.6
Note 1. Reference categories are not shown
Note 2. Model 1 estimate effects of main variables; Model 2 and 3 in addition, include one background variables at a time, while model 4 full model contain
main variables and background variables simultaneously.
Father's occupation appears to have very little impact in explaining the effects of high birth order on
infant mortality. Only a slight change in the effects of preceding birth interval and age at maternity is
observed when father's
occupation is controlled. Father’s occupation has its own highly significant effect on infant mortality
rate. Infants born to fathers working in the sales-service sector and production/laborers-others are,
respectively, 2.3 and 1.9 times more likely to die compared to infants born to fathers engaged in
professional-secretarial activities. Occupation of father remains important even after controlling for
the effects of other socio- economic variables such as education of father and household income (see
Table 7). When all other control variables are included in the model simultaneously, however, the
effect of father's occupation is not only reduced in size but also lost its statistical significance.
Child mortality: As stated earlier, the length of following birth interval is used as a covariate
in this analysis and this has reduced the number of cases to 1,250 only. Table 8 presents hazards of
child mortality. Model 1 considers the effects of maternal factors alone. Model 2-4, in addition,
include the effects of education of mother, household income and survival status of preceding child.
Model 5 is the full model and model 6 considers the effects of control variables only.
200 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 8 shows that the effect of age at maternity, which was strong in the case of infant mortality,
has disappeared for child mortality. High order births and short preceding birth intervals have
relatively high but non-significant effect on child mortality. Of the four variables of main interest, only
length of following birth interval appears to have strong and highly significant effect on child
mortality. The single factor effect of length of following birth interval is such that the chance of dying
during childhood of an index child born within 18 months after the birth of his/her older sibling is 3.8
times higher than that of a child born after 18 months.
The effect of maternal factors on child mortality is not influenced when the control variables are
entered into the model sequentially. In fact, the magnitude of the effect of short preceding birth
interval is
Table 8: Relative risks of maternal factors when control is made for some background variables, child mortality: Hazard model.
Variable
Multi-variate effects
1
2
3
4
5
1.138
1.310
1.158
1.447
1.189
1.356
1.276
1.523
1.191
1.367
1.177
1.193
<20 yrs
35+
Preceding birth interval
1.0953
0.3654
1.118
0.308
1.190
0.295
1.041
0.305
1.115
0.306
1.385
0.317
<2yrs
2-3 yrs
Following birth interval
1.519
1.373
1.362
1.454
1.386
1.478
1.414
1.487
1.292
1.465
1.082
1.407
<=18 Month
Education of mother
No school
Elementary
Education of father
3.797
3.869
3.796
3.870
3.738
4.589
Uni-variate
Effect
Birth order
2-3
5+
Age at maternity
1.776
1.255
6
1.061
1.214
0.968
1.218
No school
Elementary
Occupation of father
Sales-serv
Prod-other
Work status of woman
1.203
1.312
1.157
1.411
0.739
1.292
0.809
1.358
Working
Income (Birr)
<100
100-299
Survival of older sibling
Dead
Ethnicity
Aroma
Others
Religion
1.741
1.688
6.043
2.921
6.334
3.102
2.766
6.409
3.151*
2.666
2.42
1.815
1.767
1.730
1.564
Muslim
2.552
2.555
Prot/Chatholic
2.234
2.284*
Water
Pipe-com
0.952
0.936
Public tap
0.467
0.569
Well/river
1.293
1.031
Latrine
Common
1.464
1.221
None
1.211
0.952
-2 x Long likelihood
467.2
465.4
452.8
562.1
20.0
438.9
Note 1. Reference Categories are not shown
Note 2 figures in Column 2 are the univariate effects of main variables; column 2-2-7 present the multivariate effects. Model 1 estimate the combined
effects of main variables. Models 2-4 in addition, include one background variables at a time, while model 5 (full model) contain main variables and
background variables simultaneously. Model 6 considers the effect of background variables in group.
Determinants of child mortality 201
────────────────────────────────────────────────────────────
Reduced when survival status of the preceding sibling is controlled. However, the reduction in the
effect is negligible. The death of the preceding sibling is expected to raise the survival chance of the
index child by removing the competition for food and maternal care. The net effect of the length of the
following birth interval is almost the same as that obtained in the absence of the control variables.
The control variables, taken as a group, are important in explaining the variability in early childhood
mortality. However, the variation is statistically singificant only for household income, religion and
survival status of previous sibling. Other things being equal, children born in households with a
monthly income of less than Birr 100 are about 6.3 times more likely to die during early childhood
compared to those children born in households with a monthly income of more than Birr 300. The
corresponding risk for children born in households with a monthly income of Birr 100-299 is about
3.10 times higher (Table 8).
When the effects of all other variables are taken into account, ethnicity loses its significance.
Although it is difficult to explain why religion becomes important during early childhood period than
at infancy, the finding suggests that net of other effects, children born to Muslim mothers are 2.6 times
more likely to die compared to those born to Orthodox mothers and those born to Protestant/Catholic
mothers are 2.3 time more likely to die compared to those born to Orthodox mothers.
Discussion
The analysis generally suggests that the hazards to infant and child survival arise from young and
old age at maternity, short birth intervals and higher birth order. Births to young (under 20) and old
(35 and older) mothers are at a greater risk of dying during infancy than those to mothers aged 20-34
years. The gap is wider for young mothers and more pronounced during infancy. Births within two
years after the birth of the preceding child had significantly higher risk of dying at infancy compared to
births that occurred three or more years after the birth of the preceding birth. High order births (five
and higher) were found to be at a higher risk of dying during infancy compared to births of order 4.
Length of following birth interval was also found to have a considerable effect on childhood mortality.
The likelihood of dying before age three of children born within 18 months of the birth of their older
sibling is 3.8 times higher than those of children born after 18 months. This high negative effect of
short subsequent birth interval persists even after the effects of other variables are taken into account.
When the relative importance of maternal factors is considered, the preceding birth interval is the
most important determinant of infant mortality followed by age at maternity and birth order,
respectively. For early childhood mortality, however, the situation is different. Following birth
interval is the single most important maternal factor affecting the survival status during early
childhood.
The control variables have differing impact during infancy and early childhood. While source of
water and latrine are very important during infancy, household income and survival
status of preceding child appear to be influential during early childhood period. Education of woman
has its own direct significant effect on infant mortality although this effect is reduced when source of
water, latrine and household income are controlled. Source of water, availability of latrine, ethnicity,
and occupation of father are shown to have independent effect on infant mortality.
Mortality during infancy and early childhood can be significantly reduced if births to very young
mothers, to mothers aged 35 years or older, births that occur within two years after the birth of the
previous child and births of order five and higher are prevented. The strong association between
infant/child mortality and age at maternity, birth order and birth interval suggests that high risk births
can be prevented by expanding family planning, and reproductive health services.
The study also shows that socio-economic variables, especially those relating to improved
environment and higher household income play important roles in lowering infant and childhood
202 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
mortality. Availability of latrine and access to safe water were found to be important factors affecting
infant and child mortality. Greater attention should therefore be given towards increasing access to
safe drinking water, female education and improving household income.
This study has attempted to provide some insight about the determinants of infant and early
childhood mortality. A better understanding of the determinants of infant and early childhood
mortality may be achieved in future studies if more health related information on child birth and on
child care practices are collected. More specifically, it would be useful to collect detailed information
on such variables as birth attendant, access to, and utilization of, maternal and child health services,
treatment of umbilical cord, birth weight, nutrition status of mother, food preparation practices,
personal hygiene, vaccination, and treatment during illnesses as well as duration and intensity of
breastfeeding.
References
1. Central Statistical Authority, 1990 National Family and Fretility Survey Report. Addis Ababa, 1993
2. Assefa H. Infant and early childhood mortality in Shewa Region: An investigation into the Levels
and Differentials. M.Sc. Thesis in Demography, Addis Ababa University, Addis Ababa. 1991.
3. Yohannes K. Correlates of Intant and Child mortality in Addis Ababa, M.A. Thesis in Population
Studies, University of Ghana, Legon. 1990.
4. Abate M. Mortaltiy in rural Ethiopia: Levels, trends and differentials. PhD Thesis, University of
Pennsylvania., Philadelphia. 1988.
5. Abdulahi H. Prenatal, Neonatal and Infant Mortality in Ethiopia. In procedings of the East Africa
Workshop on Research and Intervention Issues Concerning Infant and Child Health. IDRC, Ottawa,
1988.
6. Kassahun D. Infant and Child Mortality Differentials in Rural Ethiopia, M.A. Thesis in Population
Studies, University of Ghana, Legon. 1987.
7. Tesfayesus M. Mortality Levels and Differentials in Ethiopia: with reference to Metu, Alemaya and
Addis Ababa. A Thesis submitted for the Degree of Doctor of Philosophy, University of Dar-esSelam. 1985.
8. Nebiat T. Health, Nutrition and Problems of Maternal, Infant and Child Mortality in Ethiopia. In
Proceedings of Conference on Population Issues in Ethiopa’s National Development. ONCCP, Addis
Ababa. 1989;II:505-546.
9. Shamebo D. Epidemiology for public health research and action in a developing society; the
Butajirra Rural Health Project in Ethiopia. The Ethiopian Journal of Health Development 1994;8
Special Issue.
10. Bhakta G, Kim S and Abdul Kashem M. Socio-economic Demographic and environmental
determinants of infant mortality in Nepal. Journal of Biosocial Science. 1991;23(4):425-435.
11. Pant PK. Effect of education and household characteristics on infant and child
mortality in Urban Nepal. Journal of Biosocial Science. 1991;23(4):437-443.
12. United Nations. 1985. Socio Economic Differentials in Child Mortality in Developing Countries.
New York (UN publication, Sales No. E. 85.XIII.7).
13. Holian J. Community level determinants of infant mortality in Mexico Journal of Biosocial
Science, 1987;20(1):67-77.
14.Central Statistical Authority Population and Housing Census of Ethiopia 1984, Analytical Report at
National Level, Addis Ababa 1991.
15. Assefa Hailemariam. An overview of the determinants of high fertility in Ethiopia. Ethiopian
Journal of Dev. Research, 1992;14(2):1-30.
16. Assefa Hailemariam. Fertility Levels and Trends in Arsi and Shewa Regions of Central Ethiopia.
Determinants of child mortality 203
────────────────────────────────────────────────────────────
Journal of Biosocial Science., 1991;23(4):387-400.
17. Brass W. Methods for Estimating Fertility and Mortality from Limited and Defective Data.
Laboratories for Population Studies. The University of North Carolina Chapel Hill. 1975.
18. Cox D.R. Regression Models and life tables. Royal Statistical Journal. 1972;34B:187-220.
19. Federici N. and Terrenato L. ‘Biological Determinants of early life Mortality’ in Biological and
Social Aspect of Mortality and the length of life. Proceedings of a Seminar at Fiuggi Italy. May
1980;13-16 Liege:IUSSP, 1980.
20. United Nations Manual X: Indirect Techniques for Demographic Estimation. Department of
International Ecomomic and Social Affairs. Population Studies. 1983 No. 81.
21. Office of Population and Housing Census Commission. The 1994 Population and Housing Census
of Ethiopia: Results for Oromiya Region 1994;vol. I part III.
22. Office of Population and Housing Census Commission and analytical report at national level.
Central Statistical Authority. The 1994 Population and Housing Census of Ethiopia: Results for Addis
Ababa. Volume I. Statistical Report 1994.
Original article
Clinical characteristics of orbital tumours as seen in
a tertiary eye center
Mussie Haile1 and Wondu Alemayehu1
Abstract: Clinical files of 93 patients with orbital tumours were examined. Primary orbital tumours
accounted for 55% of cases and secondary tumours for 45%. The most frequently diagnosed primary
orbital tumours were mucoceles (15%), lacrimal gland tumours (7.5%), and rhabdomyosarcoma
(6.5%). The most common secondary orbital tumours were retinoblastoma (23.6%), and squamous
cell carcinoma (10%). Commonly found physical signs, radiographic and ultrasonographic findings
are reported. Most frequently employed surgical techniques for removal of the tumours was
exenteration (64%). Orbitotomy through various approaches was employed in 36% of the patients.
Agreement was obtained between the clinical and histopathologic diagnoses for the commonly
encountered orbital tumours. Poor outcome is expected, because of late presentation in most patients.
Discussion and recommendations on further management of cases is given. [Ethiop. J. Health Dev.
1997;11(3):201-206]
Introduction
The orbit can harbour primary tumours arising from orbital structures, it can be secondarily involved
from eyelid and ocular tumours, or it can also be a seat for distant metastasis. Generally cystic,
vasculognic and inflammatory tumours predominate the picture of primary orbital tumours in all age
groups (1, 2, 7, 11). Retinoblastoma in children and eyelid tumours and uveal melanomas in adults are
the most common causes of secondary orbital tumours elsewhere. In African studies the picture is
similar with the exception of rarity of uveal melanoma as a cause of secondary orbital tumours.
The source of pathology specimens for orbital tumours is generally from exenteration of the orbit or
by total or subtotal excision of the tumour mass through various surgical approaches (1, 2, 7).
Various orbital tumours the patterns of which tend to vary with the age of the patient and
geographical location of the study area have been described by numerous investigators (1, 10, 12, 13).
The sources of materials of most of the reports on incidence of orbital tumours are from pathology
units and few studies integrated the clinical characteristics with the histo-pathologically verified
diagnosis (2, 3, 4, 5).
Although the pathologic patten of orbital tumours has been described in this tertiary eye center 14
years ago, the clinical characteristics of the different orbital pathologies were not studied (13).
The purposes of this retrospective study are to:
1. describe the pattern of orbital tumours by age, sex, clinical impression and histopathologic
diagnosis,
2. study the clinical characteristics of the different orbital tumours and compare with
histopathologic diagnosis,
3. study the malignancy rate of orbital tumours.
______________________________________
1
From the Department of Ophthalmology, Addis Ababa University, P. Box 33456, Addis Ababa,
Ethiopia
Methods
The study was conducted at Menilik II Hospital, Department of Ophthalmology which is a teaching
and national referral center under the Medical Faculty of Addis Ababa University, Addis Ababa,
Clinical characteristics of orbital tumours as seen in a tertiary eye center 205
────────────────────────────────────────────────────────────
Ethiopia.
We looked for the clinical records of 143 patients with orbital tumours on whom surgery was
performed and histopathologic verification of their tumours was made. Pathology files of the study
population found at Tikur Anbessa Hospital, Addis Ababa University, Medical Faculty, were identified
and studied in detail.
Patients who underwent orbital surgery from 1st July 1990 up to 30 th June 1994, the surgical
specimens of whom were examined histo-logically and patients with both primary orbital tumours and
lesions originating from eyelids and eyeball with secondary orbital involvement were included in the
study. The study was limited to analyze four years records as data were incomplete prior to this
period.
Patients with lesions confined to eyeball and eye lids without orbital involvement, dermoids and
orbital inflammatory conditions not subjected to surgery were excluded from the study.
All available clinical information in each patient is recorded. The variables that were looked for
are: age and sex of the patients, duration of illness, mode of proptosis (position of globe), visual acuity,
general physical examination, ophthalmoscopic and extraocular motility findings, amount of proptosis
(Hertel exophthalmometry), x-ray characteristics, ultrasonographic picture, type of surgery and
histopathologic diagnosis. The classification system of orbital tumours given by Jerry A. Shields is
used in the present study to fill and analyze the data available (2, 3).
Results
During the period 1st July 1990 to 30 th June 1994 a total of 143 patients were operated for orbital
tumours with their specimens subjected to histological examination. Clinical data and histopathologic
reports of 93 patients were available for study.
The frequency distribution and demographic data of patients with orbital tumours are shown in
Table 1. Primary orbital tumours accounted for 55% of cases and secondary orbital tumours for 45%
of all cases. The most frequently diagnosed primary orbital tumours were mucoceles 14 (15%),
lacrimal gland tumours 7 (7.5%) and rhabdomyosarcomas 6(6.5%). The most common secondary
orbital tumours were retinoblastoma 22(23.6%) and squamous cell carcinoma 15(16.1%). Forty eight
percent of cases occurred in the right orbit and 52% in the left. The male to female ratio was 1:2.
Table 1: Frequency distribution, age range, means of age and duration of illness of patients with orbital tumour
Pathologic diagnosis
Cases
Age
Mean Age
mean duration
Range (yrs)
No
%
(Yrs ±SD)
(Month)
Retinoblastoma
22
23.6
0.7-7
3.3±1.3
10
Squamous cell ca.
15
16.1
22-75
47±15.9
18
Mucoceles
14
15.0
17-70
40±19.1
36
Rhabdomyosarcoma
6
6.5
4-20
9± 6.1
2
Lacrimal gland tumours
Pleomorphic adenoma
Carinoma
Malignant melanoma
Neurofibroma
Hemangioma
Meningioma
Lymphoma
Adipose tumours
Others*
Total
3
4
5
4
4
4
2
2
8
93
3.2
4.3
5.4
4.3
4.3
4.3
2.2
2.2
8.6
100.0
16-50
22-60
28-70
4-63
18-50
11-45
35-50
50-60
32±17.0
45±16.3
54±16.2
32±24.1
35±13.4
26±15.9
42±10.6
55± 7.0
76
11
9
16
96
55
21
106
* One each case of: Pseudotumour, Fibrohistocytosis, Varix, Angiofibroma, Basal cell ca., Granuloma, Neuroblastoma, Hydatid cyst.
The age range and mean ages at presentation for orbital tumours are also shown in Table 1. The
mean age at presentation for retinoblastoma was 3.3±1.3 SD yrs, for mucoceles 40±19.1 SD yrs and
206 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
squamous cell carcinomas 47±15 SD yrs. Patients with pleomorphic adenoma of lacrimal gland had
lower mean ages (32±17 SD yrs) as compared to those of lacrimal gland carcinoma (45.5±16.3 SD
yrs). The mean duration of clinical history was least in patients with rhabdomysacrcoma (2 months)
and retino-blastoma (10 months).
The clinical data on patients with orbital tumours is shown in Table 2. Eighty nine percent of
patients presented with proptosis. Positive family history was found only in one patient who had
squamous cell carcinoma. None of the patients with retinoblastoma had positive family history of the
disease.
Table 2: Clinical data on patients with orbital tumours
Signs
Cases
No
Proptosis
Present
83
Absent
10
Total
93
Globe status*
%
89
11
100
Intact
Destroyed
Total
Eyelid involvement
49
44
93
53
47
100
Present
Absent
Total
Ocular motility
43
50
93
46
54
100
Restricted
Normal
Total
* Patients with intact globes only.
33
9
42
79
21
100
The globe was destroyed at presentation in 47% of patients. All of the 22 patients with
retinoblastoma and 13 of the 15 patients with squamous cell carcinoma (86.7%) had ruptured globes at
the time of presentation.
Eyelids were infiltrated by tumours in 46% of patients. Thirteen out of 22 patients with
retinoblastoma (59%) and 13 of the 15 patients with squamous cell carcinoma (86.7%) had lid
involvement at presentation. Out of 42 patients who had records of ocular motility examination
33(79%) had restricted ocular motility in one or more directions.
Details on position of globe at presentation and visual acuity levels are as shown in Table 3. Out of
the 49 patients with intact globes, record on globe position was obtained in 46 patients. Four (9%)
patients had normal globe position and the remaining 42 patients (91%) had globe displacement in one
or more directions. Together with the proptosis, abnormal globe displacement observed was, down
and out (28%), vertically down (15%), horizontally out (13%) and axial displacement (11%). Seventy
four percent of patients were blind in the involved eye.
Clinical characteristics of orbital tumours as seen in a tertiary eye center 207
────────────────────────────────────────────────────────────
Table 3: Details on position of globe and visual acuity levels
Globe Displacement
No.
Down and Out
13
Down
7
Out-horizontally
6
Axial
5
Normal
4
Others
11
Total
46
Visual Acuity (n=93)
NLP* - 3/60
70
>6/36 = 6/24
12
6/18
12
Total
93
%
28
15
13
11
9
24
00
74
13
13
100
The amount of proptosis was quantified by means of Hertel’s exophthalmometry in 29 of the 49
patients with intact globes. The mean difference in exophthalmometric readings between the involved
eye and fellow eye of the entire group was 6.3 mm, being 7.5 mm for mucoceles and 7.6 mm in
meningiomas.
Table 4: Mode of surgery in 93 patients with orbital tumours
Type of Surgery
No.
%
Exenteration
Total Exenteration
37
40
Lid Saving
22
24
Orbitotomy
Brow Incision
Total
24
26
Trans Conjunctival
Coronal Flap
Kronlein
4
3
3
93
4
3
3
100
Ophthalmoscopic examination was possible in 32 patients. In the rest of the patients fundus
examination was not possible either because of destroyed globes (44 patients), opaque media (10
cases) or unknown reasons (7 cases). The fundus appeared normal in 20 patients (63%). The most
common abnormal fundus finding was choroidal striae (16%). Other abnormal fundus findings include
optic nerve pallor (3 cases or 9%), choroidal striae and optic nerve pallor (2 cases or 6%) and
papilloedema (1 cases 3%). Orbital x-ray was obtained in 58 (62%) of the patients. The commonest
x-ray abnormality found was increased soft tissue density which, in isolation or in combination with
other abnormalities, was seen in 45% of the patients. Orbital widening was seen in 36% of cases either
as an isolated finding (12%) or in combination with other x-ray findings (24%). Bone destruction with
sinus involvement was seen in 16% of patients. Two patients had enlarged optic canals.
Table 5: Mode of surgery for common orbital tumours
Diagnosis
Retinoblastoma
Exenteration
Squanous Cell Ca
22
Mucocele
15
Lacrimal Gland Tumours
Rhabdomyosarcoma
3
Malignant Melanoma
6
Neurofibroma
5
Hemangioma
2
Meningioma
Total
2
Orbitotomy
14
4
4
4
2
208 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Solid tissue pattern was the most frequently noticed ultrasonographic finding (63%). Cystic picture
was noticed in 11 (34%) patients. Infiltrative pattern in one patient (3%). The surgical procedures
performed on patients with orbital tumours are given in Tables 4 and 5. The most common surgical
procedure performed was exenteration (37 or 40%) or partial, lid saving exenteration (22 or 24%).
Orbitotomy through different approach accounted for 34 (36%) of the operations. The specific type of
surgery performed for common orbital tumours is shown in Table 5. Agreement between clinical and
histopathological diagnosis of the commonly diagnosed orbital tumours is depicted in Table 6. The
agreement is higher for retinoblastoma, squamous cell carcinoma, mucocele and rhabdomyosarcoma.
Table 6: Agreement between clinical and histopathological diagnosis of the common orbital tumours
Disease Condition
Clinical Diagnosis
Histologically Confirmed
Missed No.
Retinoblastoma
22
21
1
Squamous cell Ca
17
15
2
Mucocele
Rhabdomyosarcoma
Lacrimal Gland Tumours
Malignant Melanoma
Hemangioma
Retinoblastoma
18
8
6
4
8
22
14
6
7
5
4
21
4
2
4
2
1
1
Discussion
The comprehensive classification system of orbital tumours and pseudotumours was given by
Shields et al (2). This classification system was utilized in the present series. In the review of orbital
tumour by Shield’s et al, cystic, inflammatory, lacrimal and secondary orbital lesions (in that order)
dominated the picture.
The single most common cystic lesion in the series conducted by Shields was dermoid cyst
accounting for 80% of the cystic lesions. Because of lack of definition of exact tumour site of dermoid
tumours, this tumour was excluded from our study. Excluding dermoid cyst the second most common
cystic tumour was mucocele. This condition was the most common benign lesion observed in the
present series.
The relative frequency of orbital tumours obtained in the present study agrees with other studies
conducted on the subject in developing countries (4, 6, 13). Exception to this is the fact that
significant proportions of the patients with orbital tumours observed in Nigeria had extension of
Burkitt’s lymphoma from maxilla. Rapid regression of Burkitt’s lymphoma with cytostatic drugs can
be achieved and, as treatment of this condition is in pediatric hospitals, the patients could have escaped
histopathological examination of their orbital tumours. In our series retinoblastoma and squamous cell
carcinoma were observed to be the most common secondary orbital tumours. This is in agreement
with other studies in developing countries (1, 4, 6). In contrast to this in a series conducted by Shield
et al, basal cell carcinoma and Uveal melanomas were most common origins for secondary orbital
tumours. Both of these conditions are said to be rare in non-white populations.
A recent study on the pattern of orbital tumours in children over the past 60 years has shown that the
number of secondary tumours has decreased significantly. This reflects the improved early diagnosis
of diseases with potential secondary orbital involvement. This study also signifies the necessity for
periodic evaluation of orbital tumours for changing pattern of diseases in response to improved early
detection and management of the tumours (5).
Previous studies conducted in developing countries indicated that patients with retinoblastoma
usually present in advanced stages. The mortality rate of patients with orbital involvement has been
shown to be 100% (6). The present study reconfirmed this fact. Early detection and management of
the disease is of paramount importance in the effort to improve vision and survival of the patient (8, 9,
11). A report by Abramson et al indicated that early detection of retinoblastoma with implementation
of less destructive modes of treatment such as cryo and laser photocoagulation has led to a decrease in
Clinical characteristics of orbital tumours as seen in a tertiary eye center 209
────────────────────────────────────────────────────────────
the frequency of enucleation and prolonged the survival of patients with the disease (8).
In the present study, none of the patients with retinoblastoma had a positive family history. These
patients probably had the non- hereditary form of the tumour. This form of the tumour is further
evidenced by the unilateral nature of the tumours and presentation at a higher mean age than the usual
presentation for the hereditary forms (9, 14).
All of the patients with secondary orbital tumours and most of those with primary tumours that were
reported in this series came to the hospital with an advanced stage of the disease. Fifty percent of the
patients had destroyed globes with eyelid involvement and three quarters with visual acuity of counting
fingers at three meters or less.
In the analysis of 20 patients with exophthalmos Frieberg and Associates have shown that the
choroidal folds they observed occurred with anteriorly located orbital tumours and in patients with
greater amount of exophthalmos (14). In the present study, although only a small number of patients
were noticed to have choroidal folds, it was the most common Ophthalmoscopic abnormality detected
and the patients had either anterior location or large proptosis.
In this study plain x-ray films and B-scan Ultrasonography have been observed to be important
ancillary tools in the evaluation of orbital tumours. In areas where computerized tomography or MRI
is not available these diagnostic tools are important in the management and follow-up of patients with
orbital tumours.
Unlike the studies from areas where early detection and less drastic modes of treatment are
employed, our study shows that the great majority of the surgeries were done for palliation of advanced
orbital tumours. Orbitotomy with resection of the tumour or drainage of cyst was performed in only
1/3 of the patients. Late presentation of the tumours in far advanced disease state made them less
amenable for non-destructive forms of treatment.
This study showed an agreement between the clinical and histopathological diagnosis for the
commonly seen orbital tumours. A proportionally small number of cases were missed clinically. A
higher index of awareness saves time and unnecessary cost spent during the preoperative evaluation of
the patients and narrows the areas where the pathologist should be looking for.
Conclusion
Delay in presentation of patients with malignant orbital tumours is observed in this series. These
conditions should be diagnosed early and prompt referral and management of cases should be made.
Complete evaluation of proptosis should be performed, including clinical history, physical and
radiological examination to facilitate early diagnosis and management.
References
1. Jones IS, Jakobiec FA, Nolan BT. Patient examination and introduction of orbital disease. In:
Duane TD, Jaeger EA editors. Clinical ophthalmology, Philadelphia: Harper and row, 1989;2(21):2125.
2. Shields JA, Bakewell B. Classification and incidence of space occupying lesions of the orbit. Arch
ophthalmol. 1984;102:1606-1611.
3. Shields JA, Bakewell B. Space occupying orbital masses in children. Ophthalmology
1986;93:379-384.
4. Templeton AC. Orbital tumours in African children. Br. J. Ophthalmol. 1971;53:254-261.
5. Kodsi SR, Shetlar DJ, Campbell RJ. A review of 340 orbital tumours in children during a 60 year
period. Am. J. Ophthalmol 1994;117:177-182.
6. Abiose A, Adido J, Agarwil SC. Childhood malignancies of the eye and orbit in Northern Nigeria.
Cancer 1985;55:2889.
7. American Academy of Ophthalmology Orbit. Basic and clinical science course, section 8, San
210 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Francisco, California, American Academy of Ophthalmology 1989;52-87.
8. Abramson DH, Marks RF. The management of unilateral Retinoblastoma without primary
enucleation. Arch. Ophthalmol. 1982;100(8):1249-52.
9. Halk BG, Siedlacki A, Ellsworth RM. Documented delays in the diagnosis of Retinoblastoma.
Ann Ophthalmol. 1985;17:731-2.
10. McFadzean RM, Gowan ME. Orbital tumours - a review of 34 cases J R Coll Surg Edinb.
1983;28(6):361-4.
11. Kulshrestha OP, Arora I, Shukla Y. Experience with orbital tumours. Indian J Ophthalmol. July
1983;31(4):313-5.
12. Kennedy RE. An evaluation of 820 orbital cases. Trans Am Ophthalmol soc. 1984;82:134-153.
13. G. Hiwot T, Alemu B, Quana’a P. (Abstract) Oculo Orbital tumours: A ten year analyses from
Menelik II Hospital. Ethiopia Med J. 1980;18(2):129.
14. Friberg TR, Grove AS. Chorodial folds and refractive errors associated with orbital tumours.
Arch. Ophthalmol. April 1983;101:598-603.
Original article
Quality of family planning services at the Family
Guidance Association of Ethiopia (FGAE) Clinic:
The clients’ perspective
Antenane Korra1
Abstract: The study was carried out to describe some aspects of the quality of family planning
services of the FGAE headquarters clinic as perceived by the clients and measure the clients waiting
time. A total of 400 clients were systematically selected and interviewed. Results of the study
indicated that good ranges of contraceptive method-mix are made available in the clinic and clients
calimed to be well informed about the methods, including their possible side effects and
contraindications as well as how the methods work. Nearly all interviewed clients claimed that they
had enough time to discuss about problems and were cordially treated by service providers. The
average length of a visit to the clinic was estimated to be an hour and 33 minutes and clients had to
spend 55 minutes in the waiting room. Further, 89% of the clients expressed their satisfaction with the
overall services of the clinic and the majority stated that they would recommend and encourage others
to get family planning services from the clinic. In conclusion, the quality of care in the clinic is found
to be good from the clients perspective and no major constraints were identified. It is recommended
that the family planning IEC program of the clinic should be improved, i.e, clients have to get
information about the available methods with equal vigor. A mechanism has also to be established to
follow up defaulters. [Ethiop. J. Health Dev. 1997;11(3):207-212]
Introduction
The quality of family planning services is a subject of increasing interest to family planning service
providers and organizations responsible for financing and promoting family planning services. Quality
of services is important both for its intrinsic value-high quality service is inherently more desirable
than a lower quality service- and for its instrumental value, ie, higher quality service should be
associated with, or result in low complication rates, better acceptance, higher continuation rates, and
declining fertility rates overtime (1).
Most family planning programs set quantitative targets to reach their ultimate goals of reducing
population growth and improving people’s health. The number of new acceptors, continuing users, the
couple year of protection (CYP) generated, and contraceptive prevalence rate are the most common
measurements of success. Until recently, little attention has been paid to measure the quality or the
impact of family planning service delivery on clients initial acceptance, satisfaction, correct method
use, follow-up, clinic visits or continued use (2).
During its long years of services, the Family Guidance Association of Ethiopia (FGAE) has been
exerting considerable efforts to ensure quality family planning services through its service delivery
outlets. Since the founding of FGAE clinic at the headquarters, attempts have been made to provide
quality services for clients who seek family planning and reproductive health services.
To this end, assessing the overall quality of family planning services of the FGAE family planning
clinic by considering some of the quality of care elements of identified by Judith Bruce (3) seems to be
essential to maintain and/or further strengthen its service delivery endeavors.
______________________________________
1
From the Family Guidance Association of Ethiopia, P.O. Box 5716, Addis Ababa, Ethiopia
This study, therefore, attempts to analyze the quality of family planning services in FGAE
212 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
headquarters clinic using data generated from a survey conducted to serve the purpose. First, aspects
of service quality, viz, availability of contraceptive methods, information provided on their use,
interpersonal relations and mechanisms to encourage continuity, are described. Then, contraceptive
choice by women, including providers and other influences on the selection of the method is addressed.
A client flow analysis is also performed to measure the client waiting time in the clinic.
Methods
Data to assess clients' perspectives on the quality of services during clinic sessions were collected by
client interviews at exits. A systematic random sampling scheme was used to select respondents. Out
of the first five clients who attended the clinic one was selected randomly and then every fifth client
who had received family planning services was interviewed until a total of 400 interviews were
completed. These exit interviews were geared toward finding out how clients perceived the quality of
services of the clinic. A sample size of 400 was determined primarily on logistical feasibility and by
considering desired levels of accuracy in estimates of proportions.
A structured questionnaire was developed and administered to clients through enumerators. Five
enumerators were involved to perform the exit interviews after providing them a one day intensive
training. In addition to the questionnaire, a client flow form was designed and utilized to carry out the
clients flow analysis.
Data entry and analysis, which includes manual editing, coding and cleaning of data as well as
production of the required tables and statistical analysis were made before and after data entry.
Verification of the edited questionnaires during data entry was made. The data entry was made using
Statistical Package for Social Sciences (SPSS/PC+) and the same software was used for the production
of tables and other relevant statistical analysis.
Results
Background characteristics of the sampled clients showed that more than half of the interviewed
clients were in the age range of 25 to 35 years and their mean age was estimated to be about 30 years.
The overwhelming majority (88.9%) are christians and about 79% attained at least elementary
schooling. Majority of the interviewed clients were in union at the time of the study (Table 1).
Quality of family planning services at the FGAE clinic 213
────────────────────────────────────────────────────────────
Table 1: Characteristics of Sampled Clients
Characteristics
Age
15-24
25-34
35-49
Mean age
Frequency
Percent
72
212
117
30
18.0
52.8
29.4
Christian
Muslim
Educational Level
354
43
88.9
10.8
Illiterate
Read and Write
Elementary
Junior Secondary
High School
Above Grade 12
Marital Status
63
20
101
42
114
60
15.8
5.0
25.3
10.5
28.5
15.0
49
327
17
6
12.3
82.0
4.3
1.5
Religion
Single
Married
Divorced
Widowed
As shown in figure 1, about 94% of the clients reported that they have been informed about family
planning methods. Out of these 93.8% were informed about the pills, 84.2% about IUD, 72.9% about
injectables, and 68% about the condom. The finding also indicated that new clients seem to be least
informed about permanent (VSC), and long acting (Norplant) contraceptive methods. On the average,
clients have been informed about four of the available methods.
Figure 1: Percentarge of Respondents Informed about Contraceptives by Method
About three-fourths of the clients reported that they have chosen the contraceptive methods
themselves, and 22.9% claimed that their choice was influenced by recommendations from service
providers. The contraceptive use experience of the clients revealed that prevalence of pills is found to
be high followed by injectables (figure 2).
214 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Figure 2: Percentage of Clients by Method they have Ever Used
The study further showed that nearly all clients included in the exit interview received information
pertaining to side effects and contraindications as well as how each method works. They were also
informed how to use the methods and what to do if problems arose (Table 2).
Table 2: Distribution of Clients Response about the Information Given
Information
Percent
Cases
How the Method Works
96.0
382
How to Use the Method
96.2
383
Side effects and Contraindications
95.0
378
What to do if Problems arose
94.7
377
Results of the exit interview on the client’s views on service providers show that all clients but one
claimed that the time they spent with service providers was about right. Nearly all (99.5%) of the
clients noted that they have received information and services they wanted from the clinic. All the
clients who participated in the exit interview also reported that they were politely treated by service
providers. They also claimed to be comfortable in the examination rooms.
Further, 89% of the clients had expressed their satisfaction with the overall service provision
mechanisms of the clinic. About 91% stated that they would recommend and encourage friends and
relatives to get family planning services from the FGAE clinic (Table 3).
Table 3: Distribution of Clients View on Service Providers
Clients Views
Frequency
Enough time with providers
397
Treated with respect
399
Received information and services
397
Feel Comfort in exam rooms
383
Satisfied with the services
356
Encourage others to get services
362
Percent
99.7
100.0
99.5
95.8
89.0
90.5
The analysis also showed that 22.5% of the clients had been served by the clinic for less than a year,
and 46.1% have been continuously visiting the clinic for more than two years. On the average, the
clients included in the study had been served by the clinic for about four years. Among the
interviewed clients, almost all (99.2%) claimed to be told by providers when to return to the clinic and
have scheduled for next appointments (not shown).
Results of the clients flow analysis showed that clients spent an average of an hour and thirty three
minutes in the clinic. New clients spent longer time (two hours and 35 minutes) than revisits (one hour
and 21 minutes). The analysis also revealed that IUD acceptors had to spend more time than
injectables and pills acceptors (Table 4).
Table 4: Average Time Spent in the Clinic
Average Time
Category
Hours
Minutes
Quality of family planning services at the FGAE clinic 215
────────────────────────────────────────────────────────────
Clients
New
Revisit
2
1
35
21
Method Used
Pills
Injectables
IUD
Total
1
1
2
1
2
10
44
04
33
35
It was also found out that clients spent an average of 38 minutes with clinic staff and they had to see
four or five staff in the clinic during each visit. In general, clients spent an average of about 55 minutes
in the waiting room and between contacts, they have to spend 18 minutes in the waiting room.
Discussion
There are powerful arguments and evidences to indicate that providing a choice of methods
improves program performance and individual satisfaction. Providing a choice of methods increases
the effectiveness of family planning programs due to the reason that individuals and couples pass
through different stages in their reproductive life and, therefore, overtime, their needs and values will
change. Multiple methods provide for switching for individuals who find their initial choice
unacceptable or unhealthful, and the availability of a variety of methods makes it more likely that,
given erratic contraceptive supplies, at least services for some methods will be available (4).
This element of quality of care was assessed based on data obtained from exit interviews and
observations. At the time of the study, about 10 different contraceptive methods (Oral pills, Condom,
Intra Uterine Device (IUD), Injectables, Cream jelly, Foam, Foam tablet, Diaphragm, Norplant and
VSC) were made available in the FGAE clinic. This shows that there is a reasonably good mix of
contraception in the clinic. Although preference for pills and injectables was prevalent, all the
methods mentioned are used by the clients.
About 93% of the clients reported to have received information about two or more methods of
contraception, and 87% of them named at least three methods. An average of about four methods were
identified by the interviewed clients. Although the information provided on methods like permanent
and long acting contraceptive methods are not sufficient, clients of the FGAE clinic seem to have
information on a good range of contraceptives. Free and informed choice of contraceptive methods
seems to be available in the clinic. However, it is worth to mention that VSC and Norplant are the most
recently introduced contraceptive methods in the clinic. A similar study in Kenya showed that
information was provided on a reasonable range of methods, with the exception of permanent methods
(5).
It has been hypothesized elsewhere that if the contraceptive methods are not explained sufficiently
and their side effects appreciated, users are much more likely to discontinue using them. In this regard,
about 96% of the clients have been informed about how contraceptive methods work and about their
correct use. Accordingly, 95% of them stated that they have been informed about the possible side
effects and contraindications of the methods. The findings generally showed that clients of the FGAE
clinic claimed to be well informed and counselled about the available contraceptive methods.
Client-provider relationships form part of a set of indicators that measure the clients' attitudes
toward their interactions with service providers. Such attitudes are hypothesized to contribute to the
client's overall satisfaction with the services (6). Despite the impression that information pertaining to
clients' attitudes towards their interaction with the service providers, in most cases is highly susceptible
to a courtesy bias, results of the exit interview on this aspect indicated that the clients relations with
service providers were generally rated as good. Further, the majority expressed their satisfaction with
the overall services of the clinic.
Mechanisms to encourage continuity indicate a program's concern and ability to promote continuity
of contraceptive use, whether well informed users manage that continuity on their own or the program
has formal mechanisms to ensure it (4). The need for return will depend upon the personal
characteristics of the client as well as the type of method prescribed (6). The findings of the study
portrayed that clients have been advised to be back for the services and scheduled appointments were
given, although a follow up mechanism has not yet been established. Therefore, efforts need to be
exerted to put a mechanism in place to follow up defaulters and the reasons cited by these group of
clients could help clinic personnel to review the service delivery endeavors.
Reducing client waiting time addresses a common problem for managers of family planning
programs. Long waits in clinic waiting room can create barriers that prevent clinic services from
reaching family planning clients. Analyzing client waiting time and then developing a program to
reduce long waits would improve client satisfaction, strengthen organizational capabilities, and
ultimately increases staff productivity.
According to the results of the client flow analysis, clients of the FGAE clinic were spending an
average of an hour and thirty three minutes in the clinic at the time of the study. New clients seem to
spend longer time than revisits and IUD acceptors had to spend more time than injectables and pills
acceptors. The observed long wait for the case of new clients could be explained by the requirement of
new clients to get information pertaining to different aspects of family planning before getting their
contraceptive, and they have to spend some time on IEC. For the case of IUD acceptors, on the other
hand, the relatively long waits could be attributed to the time service providers spend to make
extensive medical check-ups before the insertion.
In sum, the clients who have been served by the clinic, experienced relatively, short waiting time
than other similar health institutions. For instance, a study on the client's view of high-quality of care
in Santiago, Chile showed that the average length of a visit to the clinic was
an hour and 40 minutes and between examinations clients have to spent 29 minutes in the waiting room
(7).
Out of the total of clients interviewed, 24.8 percent reported that they have received family planning
services from other health institutions. In comparison, most clients stated that the FGAE family
planning clinic provides good medical examination before dispensing any contraceptive method, and
some said a variety of contraceptive methods are available and have a free informed choice. They
further mentioned that, in other health institutions, the availability of different types of contraceptives
is skimpy and do not offer a free informed choice of methods. A relatively shorter time is required to
complete the services in FGAE clinic than other health institutions is also considered as a good aspect
of the clinic.
In conclusion, the FGAE clinic provides quality family planning services to different categories of
clients and no major client complaints were identified. In line with these, the majority of clients
expressed their satisfaction with the overall services of the clinic. However, there is a need of
improving the information dissemination endeavor to address all the methods with equal vigor.
Mechanism should also be established to follow-up defaulters.
Acknowledgement
I would like to acknowledge the International Planned Parenthood Federation (IPPF) for its
financial support. The technical and logistical support of the Family Guidance Association of Ethiopia
(FGAE) in facilitating the study is also gratefully acknowledged. I am also indebted to Dr. Taye
Tokon, Head of the Medical and Clinical Department of FGAE, for his constructive comments and
keen interest for the study. The contribution of the FGAE clinic staff in expediting the data collection
process for the study is highly appreciated. The Views expressed in this paper are those of the author
and are not attributed to either IPPF or FGAE.
References
1. James EV, Robert M, Pamina G. Measurement of the Quality of Family Planning Services, 1992.
2. Young Mi Kim, Jose R, Kim W, et.al. Improving the Quality of Service Delivery in Nigeria.
Studies in Family Planning, 1992;23(2):118-127.
3. Bruce Judith. Fundamental Elements of the Quality of Care: A Simple Frame work. Studies in
Family Planning, 1990;21(2):61-91.
4. Bruce J and Jain A. Improving the Quality of Care through Operations Research. In
Operations Research: Helping Family Planning Programs Working Better, 1991;259-282.
5. Robert AM, Louis N, Margaret MG and Andrew F. The Situation Analysis Study of the Family
planning Program in Kenya. Studies in Family Planning, 1991;22(3):131-143.
6. Betrand Magnani and Knowles. Handbook of Indicators for Family Planning Program Evaluation,
1994.
7. Hernan V. The Client's View of High-Quality Care in Santiago, Chile. Studies in Family Planning,
1993;24(1):40-49.
Original article
Xerophthalmia at a welfare home in Jimma town
Negussie Zerihun1
Abstract: One hundred and seven children under sixteen years of age were examined between May
1994 and Octobr 1995; of whom 39 (36.5%) had evidence of clinical xerophthalmia. Twenty three
(58.97%) of these xerophthalmic children were males (M:F=1.4:1). Thirty six children (92.3%) were
over seven years of age. No significant association was seen between gender and xerophthalmia
(p=0.94). Respiratory tract infections and diarrhoea were seen in 30.8% of xerophthalmic and 35.3%
of non-xerophthalmic children, respectively, with no significant difference in morbidity pattern
between the two groups (p>0.5). Seven of the xerophthalmic children were wasted while two were
stunted. Chronic malnutrition (stunting) was significantly associated with xerophthalmia (p<0.05).
Night blindness and bitot's spots disappeared within three weeks of initial vitamin A administration.
Bitot's spots couldn't vanish completely in two children. It is recommended that the welfare home
administration has to provide the children with cheap and locally available vegetables that are rich in
vitamin A; and in the long run become self-sufficient by developing its own garden. [Ethiop. J. Health
Dev. 1997;11(3):213-218]
Introduction
Vitamin A deficiency (VAD) is the main cause of preventable childhood blindness (1,2). Its severe
form affects up to 500,000 young children every year; most of whom are in the developing world (3).
Although the deficiency state can affect any age group, the most susceptible are pre-school children(4).
Displaced children from natural calamities, may be at a greater risk of VAD (5). The same may apply
to institutionalised children if their diet is deficient in the nutrient.
VAD is known to be an important public health problem in Ethiopia (6-8). Nutritional education of
the public has been advocated as an important preventive tool against childhood blindness. The impact
of dark-green leafy vegetables(dglv) and fruits in the prevention of vitamin A deficiency has been a
universally accepted knowledge until recently. Some works have now imposed serious challenges by
questioning the efficacy of plant sources of provitamin A to improve vitamin A status (9,10).
The impact of vitamin A on childhood morbidity has been another controversial issue. The
association between vitamin A deficiency and increased childhood morbidity is well documented(1114). Hence, it was generally agreed that supplementation of the vitamin would reduce childhood
morbidity from diseases like diarrhoea and respiratory tract infections. But reports have appeared to
demonstrate that supplementation has little lowering effect on childhood morbidity(15-18).
The use of vegetables and fruits in the prevention of xerophthalmia will, however, continue
specially in developing countries where retinol-rich foods are difficult to come by.
The present study was initiated by an encounter, during a routine out-patient activity at Jimma Eye
Unit (JEU), of three children with mild xerophthalmia on a single morning. All of them were from a
near-by welfare home. With the belief that more cases of VAD may be available at the home, the
study had been undertaken with the following objectives:
______________________________________
1
From the Department of Ophthalmology, Jimma Institute of Health Sciences, P.O. Box 378, Jimma,
Ethiopia.
a. to study the prevalence of VAD in children under sixteen years of age who are residents of the
Xerophthalmia at a welfare home in Jimma town 219
────────────────────────────────────────────────────────────
home.
b. to compare morbidity patterns in the previous one month between xerophthalmic and nonxerophthalmic children
c. to study the regression pattern of the signs and symptoms of VAD
d. to come up with a feasible long-term solution to the problem of VAD at the home.
Methods
There were a total of one hundred and seven children under sixteen years of age at the home all of
whom were taken for the study. The names, ages, gender, weights, morbidity pattern during the
previous month, history of eye trauma, eye infection or application of traditional eye medications were
inquired and responses obtained from older children and/or foster mothers recorded. Diarrhoea was
defined as the passage of liquid stools at least three times a day. Respiratory infection was diagnosed
if the child had cough with fever for at least three days. Urinary infections were documented if there
was a history of dysuria or frequent urination with fever and/or chills; while skin infection implied any
dermatological condition, including acne-like lesions, scabies or boils.
Weights and heights (lengths) were each measured twice to the nearest 100 grams and 0.5 cms,
respectively, by two different recorders trained for the task. The average measurement was taken for
each child.
Night blindness, which has an Amharic equivalent of "dafint", is well known in the area. It was
thoroughly explained to older children and foster mothers. Older children were asked if they find it
difficult to walk about, or play, at dusk; and if younger children stumble over objects at conditions of
reduced illumination. The presence or absence of night blindness was documented only after its
meaning was well understood. When any uncertainty was noted, the case was automatically recorded
as negative (i.e no night blindness). Visual acuity (v/a) was recorded in cooperative children over four
years of age using the Snellen E-chart.
Examination of the anterior segment of the eye was done using a torch and a magnifying loupe (2x
magnification). Children suspected of clinical VAD were subjected to fluorecein staining of the cornea
for subsequent slit lamp examination at JEU. Direct ophthalmoscopy was done in xerophthalmic
children after the application of two drops of 1% Cyclopentolate drops into each eye.
Nutritional status was evaluated according to weight-for-height and height-for-age, standard
deviation scores or z-scores using the Anthro Version 1.01 anthropometry Soft Ware (Nutrition, CDC
and WHO; Dec., 1990). Children with ZWH score of <=-2 were considered wasted and those with
ZHA score of <=-2 were considered stunted.
Clinical xerophthalmia was diagnosed if a child had a history of night blindness(XN), or when
ocular examination revealed Bitot's spots (XIB), or corneal xerosis (X2). Conjunctival xerosis (X1A)
occurring together with night blindness was also considered diagnostic of clinical xerophthalmia.
Serum retinol determination was not done.
Xerophthalmic children were given three doses of vitamin A (retinol palmitate 200,000 IU) on days
one, two and seven. They were then put on prophylactic doses of vitamin A (200,000 IU) every six
months for the next eighteen months. The rest of the study population got a single dose followed by a
six monthly administration. Children under one year of age received half the dose. Response to
therapy in terms of amelioration of night blindness and disappearance of Bitot's spots was checked
weekly for the first two months, fortnightly for the next four months and monthly for the remaining
twelve months. Statistical analysis was carried out using Epi-info version 6.
Results
Of the total one hundred and seven children examined, sixty four were males and forty three females
(Table 1). Thirty nine children (36.5%) had clinical xerophthalmia, of whom only three were seven
years of age or under whereas the majority (92.3%) were older than seven years (Table 2). Twenty
220 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
two other children had conjunctival xerosis only and three others had corneal scarring that could not be
attributed to VAD alone.
Table 1: Age and sex distribution of children at the Jimma welfare home, 1995
Age(years)
Under 4
4 to 7
8 to 11
12 to 15
Total
Male
Xeroph.
1
15
7
23
Non-xeroph
6
7
8
20
41
Female
Xeroph
1
1
7
7
16
Total
Non-xeroph
4
9
7
7
27
11
18
37
41
107
Twenty three of the xerophthalmic children were males (58.97%) and sixteen were females
(41.03%) with a male to female ratio of 1.4:1. No statistically significant association was seen
between xerophthalmia and gender (p=0.94). Ninety children had visual acuity of  6/18 in the better
eye. V/a couldn't be determined in 17 children who were either uncooperative or too young.
Table 2: Signs and symptom of VAD in children at the Jimma walfare home, 1995
Age (years)
XN
XN+XIA
XIB
0-3
1
4-7
2
8 - 11
1
2
19
12 - 15
2
1
10
Total
3
3
32
X2
1
1
Total
1
2
23
13
39
Respiratory tract infections (RTI) and/or diarrhoea were responsible causes of morbidity in 30.8%
of xerophthalmic and 35.3% of non-xerophthalmic children. No statistically significant difference was
seen in morbidity pattern between the two study groups (p>0.5) (Table 3).
Seven of the 39 (17.9%) xerophthalmic and fourteen of the 68 (20.6%) non-xerophthalmic children
had wasting while six (15.4%) of the xerophthalmic and two (2.9%) of the non-xerophthalmic children
were stunted. One xerophthalmic child had both wasting and stunting. Xerophthalmia was associated
with chronic malnutrition (P<0.05). No significant difference was seen in the prevalence of acute
malnutrition in the two groups of children (p=0.94) (Table 4).
Table 3: Morbidity pattern among the study populations, Jimma, 1995
Causes of Morbidity
Xerophth Children
Non-xeroph Children
No
%
No
%
R.T.I &/or diarrhoea
12
30.8
24
35.3
Skin infections
2
5.1
1
.5
U.T.Infections
1
2.6
2
2.9
None
24
61.5
41
60.3
Total
39
100
68
100
*
**
X2*
0.23
1.21
0.01
0.02
P-value
0.23
1.21
0.01
0.02
Mantel-Haenszel test
Fisher exact 2-tailed values
Night blindness disappeared within a week of initial therapy. Ten children (25.6%) and twelve
others (30.8%) with Bitot's spots showed complete disappearance of the lesions within two and three
weeks, respectively, of initial treatment. Two children had marked shrinkage of Bitot's spots that
failed to go away within eighteen months (Table 5). Two children in the xerophthalmic group (who
had responded completely) and one child in the non-xerophthalmic group developed Bitot's spots
about 56 weeks after the beginning of the study despite the biannual vitamin A administration.
Xerophthalmia at a welfare home in Jimma town 221
────────────────────────────────────────────────────────────
Table 4: Nutritional status of children at the Jimma welfare home, 1995
Nutritional status
Xeroph. group
Non-Xeroph.
X2*
group
Normal
Wasted
Stunted
Wasted and stunted
*
Continuity correction
25
7
6
1
52
14
2
-
1.316
0.006
3.895
0.08
p-value
0.251
0.094
0.048
0.777
Table 5: Disappearance time of XN, XIB and X2, in children at the Jimma welfare home, 1995
Disappearance time
Number of children
%
Under one week
3(a)
7.69
Between 1 and 2 weeks
10
25.64
Between 2 and 3 weeks
12
30.77
Between 3 and 4 weeks
8(b)
20.51
Between 4 and 6 weeks
2
5.13
Between 6 and 8 weeks
2
5.13
Persistent
2(c)
5.13
Total
39
100.0
(a) all with XN
(b) 2 children with XIB had recurrence
(c) These children had marked shrinkage of Bitot's spots that failed to disappear in toto.
Discussion
This study may remind us of Bloch's observation of children in Danish orphanages where the
occurrence of malnutrition and xerophthalmia was associated with consumption of diets deficient in
milk and milk products (19). The higher prevalence of clinical VAD among older children may be due
to two factors. Firstly, the comparatively higher number of children over seven years of age (72.9% of
the study population) might have affected the result. Secondly, the feeding practice at the home is
responsible. There are basically three routine feeding profiles noted:
a. children two years and under are fed on different milk formulae (about four to six times a day)
that contain considerable amounts of vitamin A; in addition to gruel made from cereals, and,
occasionally, eggs.
b. children three to five years of age are fed on mainly porridge made from cereals, milk (about
three times a day) and "injera" (special bread made from a tiny seed - "Eragrostis teff") with
sauce made from legumes, a variable amount of pepper and oil and, occasionally, eggs.
c. older children (> 5 years of age) are fed almost exclusively on "injera" and sauce.
Foods rich in Vitamin A, like dark-green leafy vegetables (dglu) and fruits, have not reached the
home for over two years. Retinol-rich foods like eggs and milk are not available especially for older
children.
Kale (Brassica carinata) or "Habesha Gomen" is amply available in and around Jimma and it is not
difficult to grow. Papaya, mangoes and carrots are grown in the zone. These vitamin A-rich foods,
especially "Habesha Gomen", could be grown at the home to provide the children with some of the
nutritive requirement. The information that dark-green leafy vegetables (dglv) may have little to
contribute in the prevention of VAD has to be considered seriously.
It implies the need for total revision of the issue of dietary prophylaxis of VAD using dglv. This has
again a grim implication for poorer communities that can't afford retinol-rich foods. They may have to
depend on the health sector for their periodic supplies of vitamin A megadoses.
The male gender has been associated with a higher risk of developing VAD (4,20-22). No such
association was seen in the present study probably because of the small size of the study population. It
is known that children with xerophthalmia are prone to diarrhoea and respiratory tract infections (1114). The lack of association of clinical VAD with morbidity, in the present study, may be attributed to
222 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
the presence of sub-clinical VAD among some of the non-xerophthalmic children hence masking the
possible difference in morbidity. The small population size and morbidity under-reporting might also
have contributed. Hence, it will be difficult to make conclusions, from this study, regarding morbidity.
Failure of Bitot's spots to disappear completely or their reappearance after some time despite
vitamin A supplementation has been documented (23-24). Two children in the xerophthalmic group
and one in the non-xerophthalmic group developed Bitot's spots in the second year of initial
supplementation. Two of these children had wasting and one was stunted. Other possible causes of the
lesions like trauma, trachoma, pemphigus, kerato-conjunctivitis sicca, collagen diseases, etc..were not
apparent in these children. Hence other than VAD, malnutrition along with other factors may be
responsible for the recurrence of the lesions.
It is recommended that the welfare home administration can tackle the problem through two
measures. These are:
1. Short term measures:
a. improving diet at the home by supplying the kitchen with cheap and locally available sources of
the vitamin.
b. periodic supplementation of mega doses of vitamin A (at least twice a year). This may be done
with the help of Jimma eye unit and the Zonal Health Department.
2. Long term measure:
The home should be self-sufficient by developing its own garden. There is adequate land to grow
vegetables and fruits.
Acknowledgemtnt
This study was supported by the Research and Publication Office of the Jimma Institute of Health
Sciences to which I am grateful. My special thanks goes to Dr. Tekletsion W/mariam for his support
and encouragement. I am also thankful to Ato Temam Hussein, Dr. Amha Mekasha, Dr. Afework
Assefa, Dr.Tesfaye Getaneh, and Dr. Damtew W/mariam for their contributions to the completion of
this work and to Ato Fassil Tessema for the statistical work. The staff at the welfare home deserve my
gratitude for their unrelenting cooperation.
References
1. Prevention of Blindness. World Health Statistics Annual. 1990;16-18.
2. Foster A. Childhood Blindness. Eye 1988;2(suppl.):527-36.
3. Thylefors B. The W.H.O's Programme for the Prevention of Blindness. Int'l ophthalmology
1990;14:211.
4. Oomen HAPC. Clinical Epidemiology of Xerophthalmia in Man. Am J Clin Nutr.
1969;22(8):1098-2005.
5. Pizzarello LD. Age specific xerophthalmia rates among displaced Ethiopians. Arch Dis Child
1986; 61:1100-03.
6. De Sole G, Yigzau B, Bekalu Z. Vitamin A deficiency in Southern Ethiopia. Am J Clin Nutr 1987;
45:780-4.
7. Lintjorn B. Xerophthalmia in the Gardula area of South West Ethiopia. Ethiop Med J
1983;21:169-73.
8. Wolde Gebriel Z, West CE, Gebru H, Tadesse A, Fissiha T, Gabre P, et al. Interrelationship
between vitamin A, iodine and iron status in school children in Shoa Region, Central Ethiopia. British
J Nutr 1993; 70(2):593-607.
9. De Pee S, West Ce, Muhilal Karyadi D, Hautvast JGA. Lack of improvement in vitamin A status
with increased consumption of dark-green leafy vegetables. Lancet 1995;346:75-81.
10. Solomon, NW., and Bulux, J. Plant sources of provitamin A and human nutriture. Nutr Revs 1993;
51:199-204.
Xerophthalmia at a welfare home in Jimma town 223
────────────────────────────────────────────────────────────
11. Sommer A, Katz J, Tarwotjo I. Increased risk to respiratory disease and diarrhoea in children with
pre-existing mild vitamin A deficiency. Am J Clin Nutr 1984;40:1090-5.
12. Bloem MW, Wedel M, Egger RJ. Mild vitamin A deficiency and risk to respiratory tract diseases
and diarrhoea in preschool and school children in Northeastern Thailand. Am J. Epidemiol
1990;131:332-39.
13. Milton RC, Reddy V, Naidu An. Mild vitamin A deficiency and childhood morbidity-an Indian
experience. Am J Clin Nutr 1987;46:827-29.
14. Kothari G. The effect of vitamin A prophylaxis on morbidity and mortality among children in
urban slums in Bombay. J Trop Pediatr. 1991;37:141.
15. Arthur P, Kirkwood B, Ross D, Morris S, Jyapong J, Tomking A. et al. Impact of vitamin A
supplementation on childhood morbidity in Northern Ghana. Lancet 1992;339:361-62.
16. Ghana VAST Study Team. Vitamin A supplementation in northern Ghana: effects on clinic
attendances, hospital admissions, and child mortality. Lancet 1993;342:7-12.
17. Vijayaraghavan K, Radhaiah G, Pranksam BS, Sarma KV, Reddy V. Effect of massive dose
Vitamin A on morbidity and mortality in Indian children. Lancet 1990;336:1342-45.
18. Rahmathullah L, Underwood BA, Thulasiraj RD, Milton RC. Diarrhoea, respiratory infections, and
growth are not affected by a weekly low dose vitamin A supplement: a masked, controlled field trial in
children in Southern India. Am J Clin Nutr 1991;54:568-77.
19. Sommer A. Field Guide for the detection and control of xerophthalmia, W.H.O. 1978.
20. Solon FS, Popkin BM, Fernandez TL, Latham MC. Vitamin A deficiency in the Philippines. A
study of xerophthalmia in Cebu. Am J Clin Nutr 1978;31:360-8.
21. Brink EW, Perera WDA, Broske SP, Cash RA, Smith JL, Sauberlich HE. et al. Vitamin A status
of children in Sri Lanka. Am J Clin Nutr 1979;32:84-91.
22. Roels OA, Debeir O, Trout M. Vitamin A deficiency in Ruanda Urundi. Trop. Geogr. Med
1958;10: 77-92.3.
23. Sinha DP, Bang FB. The effect of massive doses of Vitamin A on the signs of vitamin A deficiency
in pre-school children. Am J Clin Nutr 1976;29:110-115.
24. Sommer A, Emran N, Tjakrasudjatma S. Clinical characteristics of vitamin A responsive and non
responsive Bitot's spots. Am J Ophthalmol 1980;90(2):160-71.
Original article
Intestinal helminth infections among the current
residents of the future Finchaa Sugar plantation
area, Western Ethiopia
Hailu Birrie1, Girmay Medhin1, Berhanu Erko1, Gedlu Beshah2, Teferi Gemetchu1
Abstract: In a cross-sectional survey of helminth infections made in February 1995 in the future
Finchaa Sugar Project area, Finchaa Valley, Western Ethiopia, Ascaris lumbricoides and hookworms
were found to be the most prevalent reaching, on average, 28% and 20%, respectively, among the
populations living in seven camps. Schistosoma mansoni also reached 22% and 30% in two of the
camps. Other parasites which were present at lower prevalences were Trichuris trichiura, Taenia
saginata, Entrobius vermicularis, Fasciola hepatica, and Hymenolepis nana. The geometric mean egg
counts per gram of faeces (epg) of A. lumbricoides, S. mansoni, hookworms and T. trichiura were 977,
141, 126 and 65 respectively. Both prevalence and intensity of infection of the last four parasites were
highest among those below 15 years of age except hookworm which appeared to be more prevalent
among the teenagers. All ages combined, only A. lumbricoides was more prevalent among the females
(P<0.05). The frequency distribution of A. lumbricoides, S. mansoni, and hookworm egg counts
showed that the parasites are highly over-dispersed with the majority of the sample population
producing none or few eggs, and a small portion producing relatively high numbers of eggs. Also, the
ratios of variance: arithmetic mean egg counts were large for the young age groups indicating a high
degree of aggregation of the parasites in the community and adding more evidence to the generally
held view about the frequency distribution of helminth parasites in the human population. The
possibility of increased transmission of the parasites due to irrigation development and their potential
adverse effects on the population is discussed and possible control measures suggested.[Ethiop. J.
Health Dev. 1997;11(3):219-228]
Introduction
Intestinal helminth infections are known to be persistently ubiquitous in the developing world
especially in poorest communities where the sanitary conditions are a lot to be desired (1,2). Inspite of
the availability of cheap and effective drugs for their control, worldwide prevalence rates of helminth
parasitism have not changed since the turn of the century (3). In 1990 it was estimated that at least
1000 million are infected with geohelminths alone (4). The main reason for this persistent ubiquity is
the fact that intestinal helminths frequently rank low in the list of priorities in public health programs
because the effects of helminth infections on the human population cannot be measured unambiguously
in terms of mortality figures (5). However, even if the mortality figures directly related to intestinal
helminth infections may be low, the absolute numbers are still reported to be fairly high because of the
high prevalences of infections in the developing countries (2,4). Current estimates suggest that
hundreds of thousands of avoidable deaths occur each year due to helminthiasis (4). Furthermore,
recent studies have convincingly demonstrated that helminth infections have detrimental effects on
human nutrition (4,5); they indirectly rob of his energy and ambition; interfere with his nutrition; and
ultimately make him more susceptible to other diseases (6). These findings and the availability of
broad spectrum and effective anthelminthic drugs at affordable cost have led to renewed interest to
control intestinal helminth infections through ______________________________________
1
From the Institute of Pathobiology, Addis Ababa University P.O. Box 1176 Addis Ababa, Ethiopia
and 2Medical Service of Finchaa Sugar Project, Ethiopia.
chemotherapy (2,4,5). To optimize the cost-effectiveness of chemotherapy, it is now recommended to
Helminth infection Finchaa sugar plantation 225
────────────────────────────────────────────────────────────
target it against the heavily infected individuals or sectors within the population (4,5,7-9). In order to
do this, population-based measurement of the intensity of infection and identification of high risk
groups are emphasized (7,8). More reliable estimation of the intensity of infection could be made
using direct estimates of gastrointestinal nematodes worm burdens obtained by anthelminthic expulsion
(10). However, since this procedure is logistically difficult as well as unpleasant in community
applications (11-13) the intensity of infection is usually indirectly estimated through faecal egg counts
using quantitative coprological techniques (13).
The distribution of heavily infected individuals in the population has been examined in a number of
countries (11). The major geohelminths of man are known to be highly over-dispersed, the large bulk
of the heavy infections occurring in a minority of individuals (14,15). However, although most
literatures show that the large majority of these heavily infected individuals constitute the younger age
group(s), there are conflicting reports on the sex and age patterns of infection (13). The issue is further
complicated by provision of evidence in support of predisposition to heavy infection due to genetic
and/or environmental factors (10,14,18,19). Hence, it is difficult to extrapolate results of studies
conducted in other countries for designing national control strategies. One needs to establish the
distribution of these heavily infected individuals in the population living under specific environmental
and socio-cultural conditions in order to devise appropriate community control strategies.
A number of surveys on human helminth parasites in Ethiopia have been carried out (20-32).
Hence, the species, prevalence and distribution of intestinal helminth parasites are well known for most
parts of Ethiopia. However, little is known about the distribution of the worm burden and intensity of
infection in the population. The main reason for the gap in knowledge about the intensity of infection
in Ethiopia was largely due to lack of diagnostic field tools for quantitative determination of infection.
With the adoption of the Kato’s thick smear method (33) in 1979 for a large scale use in Ethiopia it
became possible to quantify infection due to intestinal schistosomiasis among schoolchildren in
Ethiopia and the populations of some communities (34). However, due to the longer time required to
count the eggs under the microscope, this quantitative diagnostic method has not been extended to
other intestinal helminths. Recently, quantitative examination of intestinal helminths among
schoolchildren in the Kolfe Elementary School, Addis Ababa, revealed astounding intensities of
infection due to Ascaris and Trichuris in children (30). Spurred by this finding it was decided to
extend measurement of the intensity of infection to the whole population in order to determine the
intensity and distribution of worm burden, as indirectly gauged by faecal egg counts, of intestinal
helminths in the population. For this purpose, the labour population of the Finchaa Sugar Project
Area, of Western Ethiopia was selected.
Methods
The study area and population: The study was conducted in the Finchaa Sugar Project Areas,
Wellega Region, Western Ethiopia (Figure 1) in February 1995. The Project Area, lying in the
Finchaa River valley envisage to cover about 20,000 ha of land to grow sugar cane using sprinkler
irrigation system at a cost of over 300 Million USD (35). The previous State Farm area is now
absorbed by the Project. Currently, intensive construction of irrigation networks and residential houses
are underway and a sugar factory is soon to be erected.
The population, comprising workers of the previous State Farm, their families and new arrivals in
search of job opportunities, already surpasses 10,000 people. When completed, it is expected that the
Finchaa Sugar Project will
226 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Figure 1: Sketch-map of the study area.
ultimately create jobs for over 7,000 workers who may in turn create an additional population of
25,000 people or more (35). Currently, the people are living in seven camps, each with a population of
a little over 1000 except Camps 5 and 7 which have over 2000 people. Most of the houses are with
thached-roof and are constructed without plan. Water is provided at few standpipes in each camp but
the majority of the residents still depend, for almost all purposes, on the Finchaa River and a number of
smaller streams flowing close to the camps. Only very few households have pit latrines and the
sewerage system is still in its infancy. With the completion of the construction of the residential
houses and the beginning of the Sugar Factory, installation of piped water supply and sanitary facilities
is expected. However, it is not certain how many of the poor and unemployed families could benefit
from the planned improvements of water supply and sanitary facilities.
Stool examination: The number of households in each camp was estimated based on the information
provided by the local Sugar Project authorities. Based on the estimates, 10% of the households were
selected by systematic random sampling technique. Once a household was selected all members were
summoned for stool examination. For each individual a single slide was prepared using the Kato’s
thick smear method (with 41.7mg template (36)). The eggs of all intestinal helminths encountered
under the microscope (using the 10 x eyepiece magnification) were counted using a multiple tally
counter. For each parasite the number of eggs counted per slide was converted to eggs per gram of
faeces by multiplying by a factor of 24.
Statistical analysis: the data were coded and entered into a computer using the dBase IV version
2.0. and analyzed using SPSS statistical package. The intensity of infection, (estimated as exp[log
c)/n], where n is the number of individuals examined and c is the faecal egg count (epg) for each
individual), was expressed as the geometric mean of egg counts on the positive subjects. The
variance: mean (arithmetic) ratios were calculated as a measure of aggregation of parasites within sex
and age classes. The prevalences and intensities were compared with X2 tests and with Student’s ttests, respectively.
Helminth infection Finchaa sugar plantation 227
────────────────────────────────────────────────────────────
Results
The size, age and sex composition of the sample population are presented in Table 1. Males and
females are equally represented except for a slight under representation of males in the 15-19 years age
group. Altogether, 1321 persons (662 males and 659 females) were examined in the seven camps.
Table 1: The age and sex distribution of the sample population.
Age group
Male
Female
Total
0-4
104
93
197
5-9
140
146
286
10 - 14
117
104
221
15 - 19
29
42
71
20 - 24
52
59
111
25 - 29
36
82
118
30 - 34
55
48
103
35 - 39
46
47
93
40 - 44
36
18
54
45 - 49
26
14
40
50+
21
6
27
Total
662
659
1321*
* The ages of five persons missing
The intestinal helminths diagnosed were Schistosoma mansoni, Ascaris lumbricoides, Hookworm
species, Trichuris trichiura, Hymenolepis nana, Taenia saginata, Fasciola hepatica and Enterobius
vermicularis. However, only the results of the first four parasites are presented here since the rest
occurred in very low prevalence rates.
Table 2: Prevalence of intestinal helminths among the population in the Finchaa Sugar Project area, Western Ethiopia, 1994.
Camp
Pop*
No.
Exam.
Percent positive**
Geometric mean EPG
6
7
8
9
10
11
12
162
3699
765
1193
961
978
542
44
432
187
200
104
204
145
Sm
2
30
8
22
4
11
1
Al
34
27
23
40
14
26
30
Ho
2
33
13
4
21
17
18
Tt
2
2
3
2
6
4
14
Sm
24
198
62
122
69
66
72
Al
777
883
1102
908
1338
1123
1074
Ho
120
143
68
265
64
148
125
Tt
72
107
27
215
56
66
51
Total
M
8210
-
1326
667
16
17.7
28
25
20
21
4.5
3.6
142
140
968
887
125
117
64
59
F
659
15
31
19
4.6
146
1040
135
69
* - Total population according to local census in 1993
S.M - Schistosoma mansoni,
Tt = Trichuuris trichiura
A.l - Ascaris lumbricoides
Ho. - Hookworms,
** - intestinal helminths which are not shown here because of their extremely low prevalences are: Hymenolepis nana, Fasciola hepatica, Taenia saginata and
Enterobius vermicularis
All camps combined, the average prevalences of A. lumbricoides, hookworms and S. mansoni were
28%, 20% and 16%, respectively, (Table 2) while the corresponding geometric mean egg counts per
gram of faeces were 968, 125 and 142, respectively. However, S. mansoni was highest in prevalence
(30%) and intensity of infection (198 epg) in Camp 7 followed by Camp 9(22% and 122epg) while A.
lumbricoides was over 20% in six of the camps, reaching 40% in Camp 9. The intensity of infection of
Ascaris was however, highest (1338 epg) in Camp 10 where the prevalence was lowest (14%)
indicating that the few infected persons could have high worm burdens. Hookworms were most
prevalent in Camp 7(33%) followed by Camp 10(21%) but the intensity of infection was highest (265
epg) in Camp 9 where the prevalence was lowest (4%), again showing that the few infected persons
could have heavy worm burdens.
Analysis of age-specific prevalences and intensities of infection of S. mansoni showed that children
228 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
in the 10-14 years age-group are most affected (Fig. 2). All ages combined, there were no significant
differences in both prevalence and intensity of infection due to sex (P>0-05).
Figure 2: Prevalence and intensity of infection of S. mansoni by sex and age
Analysis of prevalence and intensity of infection of A. lumbricoides showed that children 0-9 years
of age are most affected with heavier load among the under-fives (fig. 3). All ages combined, more
females than males appear to be infected (P<0.05) but the intensity of infection was not significantly
different (P>0.05).
Figure 3: Prevalence and intensity of A. lumbricoides infection by sex and age
Hookworms were more or less equally distributed among all age-groups and sexes. A rise in
prevalence and intensity of infection among the 45-50 years (Fig. 4) is due to the small number
examined in that age-group. All ages combined, there were no significant differences in infection due
to sex (P>0.05).
Helminth infection Finchaa sugar plantation 229
────────────────────────────────────────────────────────────
Figure 4: Prevalence and intensity of hookworms infection by sex and age
The prevalences of heavy infections for both S. mansoni and A. lumbricoides were very low in the
population; only 5% and 6% of the infected population harboured moderate to heavy worm loads of S.
mansoni and A. lumbricoides, respectively (Table 3). That of hookworms was not analyzed since the
diagnostic technique is not absolutely reliable for quantification of hookworm eggs.
Table 3: Classification of intensity of infection of S. mansoni and A. lumbricoides in the Finchaa Sugar Project Areas, western
Ethiopia, 1995.
Egg range
All Camps
No. exam.
%
S. mansoni
Negative (0 egg)
1109
84
Low (<200 EPG)
143
11
Moderate (201-800 EPG)
52
4
Heavy (>800 EPG)
21
1
Total
1325*
100
A. lumbricoides
Negative (0 egg)
Low (<5000 EPG)
Moderate (5001 - 20,000 EPG)
Heavy (>20,000 EPG)
Total
956
290
58
21
1325*
72
22
4
2
100
* one person each missing
The frequency distribution of egg counts showed that A. lumbricoides, S. mansoni and hookworms
are overdispersed with the majority of the sample population producing none or few eggs and that only
a small number of individuals are excreting more than 150 epgs (figures 2-4). Analysis of variance:
mean ratios as a measure of the degree of aggregation (13), showed large ratios within age-groups and
sexes (Table 4), suggesting a high degree of aggregation of egg counts in the infected population. A.
lumbricoides appears to be more aggregated among the 0-9 years, especially females, while S. mansoni
is more aggregated among the 5-14 years, especially males and hookworms among the 10-19 years of
age. All ages combined, A. lumbricoides was more aggregated in the female while S. mansoni and
hookworms appear to be more aggregated in the males (Table 4).
Discussion
Ascariasis and hookworm infection are the predominant and widespread intestinal parasitic
problems in the camps of the Finchaa Sugar Project. The wide distribution of these two helminth
parasites may be related to the favourability of the environmental conditions such as dampness created
by the irrigation activities in progress, for their development, survival and transmission. It may also
reflect the defecating habits of the population in the immediate vicinity of homes due to inadequate
sanitary facilities in the camps. It is, however, of interest that T. Trichiura, which has a similar life
cycle and mode of transmission with A lambricoides, occurred in low prevalence and intensity of
infection (data not presented).
It is difficult to compare the prevalences and intensity of infections of A. lumbricoides and
hookworms in this population with those reported from other parts of Ethiopia since almost all of the
previous studies were qualitative and most were limited to school- children (20-32). However,
compared to few population-based studies which used qualitative techniques (24, 25, 27-29), our
findings indicate that the overall prevalences of both
230 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 4: Variance to mean ratios for Ascaris lumbricoides hookworms and S. Mansoni egg counts within sex and age classes in
Finchaa Sugar Project Area.
AgeAscaris
Hookworms
S. mansoni
group(yrs)
M
F
T
M
F
T
M
F
T
0-4
5-9
10 - 14
15 - 19
20 - 24
25 - 29
30 - 34
35 - 39
40+
Total
26263
13471
17508
7615
1245
21096
3824
3088
2715
8654
68933
67593
11902
15377
11702
11454
3790
12162
18898
47452
53264
53368
14787
10758
10696
18031
3619
11640
14484
37016
439
479
940
160
19
314
37
354
647
558
414
278
253
843
259
199
509
295
244
343
420
394
673
757
204
223
257
305
502
456
54
161
7085
605
478
8
38
435
65
5495
27
7937
2655
179
5
81
567
9
4703
35
6448
5353
452
452
66
40
466
47
5086
parasites are quite high in the project area. This raises a serious concern since these parasites have
been associated with decreased work capacity and productivity in both children and adults, increased
maternal and foetal morbidity and mortality, premature delivery and low birth weights, and increased
susceptibility to other infections (1,2,5,6). Evidence abounds to show that intestinal helminths,
especially Ascaris, seriously impair the mental and physical development of children (37) through
depletion of micronutrients required for growth and development. Hookworm infection causes iron
deficiency anaemia both in adults and children; pregnant women are particularly susceptible as iron
demands are increased to meet the physiological requirements of the growing foetus and maternal
tissue (38). Anaemia is always associated with a diminished capacity for sustained work and exercise.
Hence, with the present overall prevalences of hookworms and schistosomiasis mansoni in the Finchaa
Sugar Project area, both of which cause anaemia, the health and the productivity of adults whose
livelihood and contribution to the economy depend on hard physical work will be severely impaired.
For helminth parasite populations the intensity is the central statistics determining both the
morbidity of infection and the dynamics of transmission (8). The most convenient means of estimating
the intensity of infection is to quantify the density of eggs in faeces on the assumption that this is
directly proportional to the number of worms in the intestine (37, 38). The significance of the results
of the present study is the fact that it has indicated the distribution of intensity of infection, as gauged
by faecal egg count, of helminth parasites in the study population. The intensity of infection of
intestinal helminthic infections in general and their epidemiological characteristics in the adult
population in particular have not been well studied in Ethiopia. Only recently did Dagnew et al (39)
attempt to determine the intensity of A. lumbricoides in a small rural village in northwest Ethiopia
using the Stoll’s dilution technique. From our study it is evident that infections with intestinal
helminths are persistent throughout adult life. However, as indicated by the geometric mean egg
counts, the variance: and mean ratios, the infection is aggregated in the younger age-groups, A.
lumbricoides in the 0-9 years, S. mansoni in the 5-14 years and hookworms in the 10-19 years of age.
These results add evidence to the findings of the frequency distributions that the major geohelminths of
humans are over- dispersed in their distribution in the population, i.e., only a minority of individuals in
the population excrete large amounts of eggs (13-15, 32). Hence, chemotherapy-based control
programme in the Finchaa Sugar Project Area should focus on these high risk age-groups. Special
attention should be paid to the screening and treatment of the labour force and the child-bearing female
population since helminths in general, and schistosomiasis and hookworms in particular, cause iron
deficiency anaemia leading to impaired health and reduced productivity. Deciding on which strategy
(mass, selective or targeted) to use and at what interval to deliver the chemotherapy remains to be a
challenge. In Nigeria, mass chemotherapy has been reported to be cost-effective as regards lowering
the intensity of A. lumbricoides in a high endemicity area (40, 41). However, Guyatt et al. (42), after
analysing the cost-effectiveness of several control programmes, have shown that child-targeted
Helminth infection Finchaa sugar plantation 231
────────────────────────────────────────────────────────────
treatment can be more cost-effective than population treatment in reducing the number of diseases,
especially if repeated at a two yearly interval by covering at least 90% of the children.
The long-term reduction of soil-transmitted geohelminths is greatly dependent on the safe disposal
of human faeces because of the problem of re-infection. Improvement in sanitation, such as by the
provision of latrines, is likely to result in a progressive decline in the abundance of the parasites in the
host population (43). Feachem et al. (44) suggest that helminth infections are more sensitive to
improvements in sanitation facilities than are other intestinal organisms. Although the prevalence of
soil-transmitted helminths has been reported to be influenced less by water supply (43,44), provision of
piped water supply (private or communal) for all households should not be forgotten since this has
been proved to be useful in the control of faecally-transmitted parasites (45). Last but not least, all
measures must be accompanied with basic hygienic education, especially in schools.
Quantitative assessment of hookworm burden relies on faecal egg counts but epg are subject to
individual variation and density- dependent depression of fecundity (46-52). Egg output is a poor
indicator of worm burden but does pin-point very heavy and light infections. Recently, a more
sensitive diagnostic technique, known as the Agyina (53) method, has been developed for more
accurate assessment of hookworm burden and this needs to be adopted in future hookworm research in
Ethiopia. Unless preventive measures, such as provision of sanitary facilities, are implemented the
problem of hookworm infection may be exacerbated with the development of large scale irrigated
agriculture and agglomeration of population in the Finchaa Sugar Project area since the parasites are
also known to flourish in agricultural areas, especially in plantation agriculture, with excess water
resources which play an important role in their transmission and maintenance (46). For example, both
schistosomiasis and ancylostomiasis have been important diseases in areas of Egypt which, without
irrigation, would be much too dry for the survival of the parasites (46).
Since no species identification was performed, it is not known which species of hookworms is (are)
responsible for the infection in the study area. It has been reported that both Ancylostoma duodenale
and Necator americanus are sympatric in lowlands where the soil types are sandy-clay-loam and
sandy-loam (51). The project area is located at an altitude of 1200 metres above sea level with a clayloam soil type (51). Hence, it is likely that either both species of hookworms or A. duodenale, which is
more of a lowland parasite in Ethiopia (32), may be present in the Finchaa Valley. Future studies
should include species identification since A. duodenale causes more harm to the host (51, 52). Not
least, A. duodenale can infect humans equally successfully by percutaneous, oral, transmamary and
perhaps transplacental routes (52).
At present intestinal schistosomiasis appears to be limited to few camps in the project area. The
occurrence of the disease among school- children and labourers of the previous State Farm in the
Finchaa Valley was reported in 1990 (54). At that time, the prevalences of schistosomiasis in camps 7,
8, 9 and 10 were 29.5%, 6.5%, 16.7% and 3%, respectively. The current prevalences for the same
camps are 30%, 8%, 22% and 11%, respectively, showing a rapid increase in some of the camps in
about five years. Hence, if the necessary precautions are not taken, the situation may get out of control
with the development of irrigated sugar cane production which will create perennial water pockets that
favour increased propagation of host snails and parasite transmission. Schistosomiasis has now
undergone unprecedented increase in the Wonji and Metehara Sugar Estates, Southeastern Ethiopia,
from only a rare disease in the early 1960s (55, 56). Irrigation-based agriculture is labour-intensive
and calls for the concentration of people; this in turn leads to heavy use and pollution of canals and
ditches and as a result, “man-made schistosomiasis” spreads in the labour population (31, 55). The
development of irrigation schemes is the main cause of the spread of schistosomiasis in Africa (31);
they offer good breeding grounds for intermediate host snails. Hence, to control morbidity and curb
the problem of schistosomiasis from increasing in magnitude in the future Finchaa Sugar Project area,
mass chemotherapy and snail control using Endod (Phytolacca dodecandra) berries should be started
in the endemic camps (where prevalence in humans is 10% or above) before the development of the
232 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
irrigation scheme is in full swing. In the rest of the camps, indirect case detection and treatment and
regular monitoring of snails and their infection should be started.
Control measures should include periodic mass de-worming of children using broad spectrum
anthehlminthics, such as Albendazole which has been proven to have high efficacy on both Ascaris and
hookworms (2, 5). A recent study at the Kolfe Elementary School in Addis Ababa has shown that a 6monthly de-worming programme significantly reduces both prevalence and intensity of infection of
geohelminth parasites (unpublished data).
Acknowledgement
The financial support to undertake the study was obtained from the Finchaa Sugar Project. The
technical staff of the Department of Medical Parasitology, Institute of Pathobiology, especially of Mr.
Abraham Redda, Mr. Negash Gemeda, Ms. Etetu Mamao and Ms. Fantu Assefa, are highly
appreciated.
Reference
1. World Health Organization. Report of a WHO Expert Committee. Public health significance of
intestinal parasitic infections. Bull Wld Hlth Org 1987;65:575-588.
2. Savioli L, Bundy DAP and Tomkins A. Intestinal parasitic infections: a soluble public health
problem. Trans Roy Soc Trop Med Hyg. 1992;86:353-354.
3. Crompton DWT, Nesheim MC and Pawlowski ZS. Ascariasis and its Prevention and Control.
Taylor and Francis, London, New Tork and Philadelphia 1989. pp406.
4. World Health Organization. Informal consultation on intestinal helminth infection, Geneva, 9-12
July 1990. WHO/CDS/IPI/90.1.
5. Anderson TJC, Zizza CA, Leche GM, et al. The distribution of intestinal helminth infections in a
rural village in Guatemala. New Inst Oswaldo Cruz, Rio de Janeiro 1993;88:53-65.
6. World Health Organization. Soil-transmitted helminths. Report of a WHO Expert Committee.
WHO Tech Rep Ser 1964;No. 277.
7. Anderson RM and May RM. Helminth infections of humans: mathematical models, populations
dynamics, and control. Adv Parasitol 1985;24:1-101.
8. Anderson RM and Medley GF. Community control of helminth infections in man by mass
chemotherapy. Parasitology 1985;90:627-660.
9. Forrester JE, Scott ME, Bundy DAP and Golden MHN. Clustering of Ascaris lumbricoides and
Trichuris trichiura infections within households. Trans Roy Soc Trop Med Hyg 1988;82:282-288.
10. Bundy DAP and Medley GF. Immuno-epidemiology of the human geo-helminthiasis: ecological
and immunological determinants of worm burden. Parasitology 1992;104-105.
11. Croll NA, Anderson RM, Gyorkos TW, et al. The population biology and control of Ascaris
lumbricoides in a rural community in Iran. Trans Roy Soc Trop Med Hyg. 1982;76:108-197.
12. Bundy DAP, Thompson DE, Cooper ES, et al. Population distribution of Trichuris trichiura in a
community of Jamaican children, Trans Roy Soc Trop Med Hyg 1985;77:232-237.
13. Bundy DAP, Thompson DE, Cooper ES, et al. Rate of expulsion of Trichuris trichiura with
multiple and single dose regimens of Albendozole. Trans Roy Soc Trop Med Hyg 1985;77:232-237.
14. Schad GA and Andreson RM. Predisposition to hookworm infection in man. Science
1985;228:1537-1540.
15. Anderson RM. (ed). Population dynamics of Infectious diseases: Theory and Applications.
Champman and Hall (London). 1982.
16. Croll, NA and Chadirian E. Wormy persons: Contributions to the nature and patterns of over
dispersion with A. lumbricoides, Ancylostoma duodenale, N. americanus and T. trichiura. Trop Geogr
Helminth infection Finchaa sugar plantation 233
────────────────────────────────────────────────────────────
Med 1981;33:241-248.
17. Guyatt HL and Bundy DAP. Are wormy people parasite prone or just unlucky? Parasitol. Today
1990;6:282-283.
18. Bundy DAP, Cooper ES, Thompson DE, et al. Age-related prevalence and intensity of Trichuris
trichural infection in St. Lucia. Trans Roy Soc Trop Med Hyg 1987;81:85-94.
19. Forrester JE, Scott ME, Bundy DAP, et al. Predisposition of individuals and families in Mexico to
infection with A. lumbricoides and T. trichiura. Trans Roy Soc Trop Med Hyg 1990;84:272-276.
20. McConnell E, Armstrong MC. Intestinal parasitism in fifty communities on the Central Plateau of
Ethiopia. Ethiop Med J 1976;14:159-168.
21. Kloos H, Lemma A. Kirub B, et al Intestinal parasitism in migrant farm labour populations in
irrigation schemes in the Awash Valley, Ethiopia, and the major labour source areas. Ethiop Med J
1980;18:53-62.
22. Tedla S. Intestinal helminthiasis of man in Ethiopia. Helminthologia 1986;23:43-48.
23. Lo CT, Ayele T and Birrie H. Helminth and snail survey in Harerege Region of Ethiopia with
special reference to schistosomiasis. Ethiop Med J 1989l;27:73-83.
24. Zein AZ and Assefa M. The prevalence of intestinal parasites among farming cooperatives,
Gondar Region, north western Ethiopia. Ethiop Med J 1985;23:159-167.
25. Erko B and Tedla S. Intestinal helminth infections at Zeghie, Ethiopia, with emphasis on
schistosomiasis mansoni. Ethiop J Health Dev 1993;7:21-26.
26. Tedla S. Hookworm infection in several communities in Ethiopia. East Afr Med J 1986;63:134139.
27. Tesfa Yohannes TM and Ayele T. Intestinal helminthic infections in Lake Zway islands, central
Ethiopia. Ethiop Med J 1983;21:149-153.
28. Gundersen SG and Birrie H. Schistosoma mansoni and other intestinal parasites in the Blue Nile
Valley of western Ethiopia. Ethiop Med J 1988;26:157-165.
29. Birrie H, Erko b and Tedla S. Intestinal helminthic infections in the southern Rift Valley of
Ethiopia, with special reference to schistosomiasis. East Afr Med J 1994;71:447-452.
30. De Carneri I, Di Matteo L and Tedla S. A comparison of helminth infections in urban and rural
areas of Addis Ababa. Trans Roy Soc Trop Med Hyg 1992;86:540-541.
31. De Carneri I, Di Matteo L, Tedla S and Jemaneh L. Schistosomiasis mansoni and
geohelminthiases in school children at Wonji, Upper Awash Valley, Ethiopia. La Medicina Tropicale
1993;9:99-103.
32. Tedla S and Jemaneh L. Distribution of Ancylostoma duodenale and Necator americanus in
Ethiopia. Ethiop Med J 1985;23:149-158.
33. Peters P, El Alamy M, Warren K and Mohmoud A. Quick Kato smear for field quantification of
Schistosoma mansoni eggs. Am J Trop med Hyg 1980;29:217-219.
34. Birrie H, Ayele T, Tedla S and Abebe F. Transmission of Schistosoma mansoni in three ecological
settings in Ethiopia. I. Epidemiological aspects. Ethiop J Health Dev 1993;7:63-69.
35. Tate and Lyle Technical Services. Updating and revision of project Documents 1986;17p Mimmo
document Finchaa Sugar Project.
36. World Health Organization. Bench Aids for the Diagnosis of Intestinal Helminths. Geneva:
Programme of Intestinal Parasitic Infections, Division of Communicable Diseases, WHO 1992.
37. Guyatt HL, Bundy DAP, Medley GF and Grenfell BT. The relationship between the frequency
distribution of Ascaris lumbricoides and the prevalence and intensity of infection in human
communities. Parasitology 1990;101:139-143.
38. Chai JY, Kim KS Hong ST, et al Prevalences, worm burden and other epidemiological parameters
of Ascaris lumbricoides infection in rural communities in Korea. Korean J Parasitol 1985;23:241-6.
234 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
39. Dagnew M, Hailu T, Tessema S, et al. Intensity of infection and reinfection rate of A. lumbricoides
in a rural farming village, Dembia District, northwest Ethiopia. Ethiop Med J 1995;33:243-349.
40. Crompton DWT. Actions for the control of soil-transmitted helminthiases in Nigeria. Parasitol
Today 1990;313-315.
41. Holland CV and Asaolu SO. Ascariasis in Nigeria. Parasitol Today 1990;6:143-147.
42. Guyatt HL, Chan MS, Medley GF and Bundy DAP. Control of Ascaris infection by
chemotherapy: which is the most cost-effective option? Trans Roy Soc Trop Med Hyg 1995;89:16-20.
43. Arfaa F, Sahba GH, Farahmandian I and Jalali H. Evaluation of the effect of different methods of
control of soil-transmitted helminths in Khuzestan, southwest Iran. Am J Trop Med Hyg 1983;77:515521.
44. Feachem RG, Guy MW, Harrison S. Et al. Excreta disposal facilities and intestinal parasitism in
urban Africa: Preliminary studies in Botswana, Ghana and Zambia. Trans Roy Soc Trop Med Hyg
1983;77;515-521.
45. Manson PR. Patterson BA and Loewenson R. Piped water supply and intestinal parasitism in
Zimbabwean Schoolchildren. Trans Roy Soc Trop Med Hyg 1986;80:88-93.
46. Palmer DR and Bundy DAP. Epidemiology of human hookworm and Ascaris lumbricoides
infections in rural Gambia. East Afr Med J 1995;527-530.
47. Stephenson LS. Helminth parasites, a major factor in Malnutrition. World Health forum
1994;15:169-172.
48. Torlesse H, Crompton DWT, Savioli L and Albonico M. Hookworm disease and pregnancy in
tropical Africa. Trans Roy Soc Trop Med Hyg 1996;89:595.
49. Schad GA, Nawalinski TA and Kochar V. Human ecology and the distribution and abundance of
hookworm populations. Human Ecol Infect Dis 1983;8:187-223.
50. World Health Organization. African Conference on Ancylostomiasis. WHO Tech Rep Ser
1962;255:5-30.
51. Hoogland KE and Schad GA. Nevator americanus and Ancylostoma duodenale: Life history
parameters and epidemiological implications of two sympatric hookworms of humans. Exp parasitol
1978;44:36-49.
52. Crompton DWT. Hookworm disease: current status and new directions Parasitol Today 1989;5:12.
53. Ramson M, Gemetchu T, Birrie H. Et al Aginya’s modification to Kato thick smear in field work.
European Conference on Tropical Medicine October 22;26:1995. G112.
54. Birrie H, Tedla S, Erko B, Berhe N and Abebe F. Schistosomiasis in Finchaa River Valley,
Wellega Region, western Ethiopia Ethiop J Health Dev 1993;7:9-15.
55. Simonsen PE, Nega A and Furu P. Intestinal schistosomiasis among children in a labour village of
Wonji Sugar Estate, Ethiopia East Afr Med J 1990;8:532-538.
56. Tedla S and Yimam M. Schistosomiasis at the Wonji/Shoa Sugar Estates during two decades
(1964-1985). Afr J Sci Technol 1986;1:1-18.
Original article
Health risk assessment of a planned irrigation
scheme along the Genale River, South Ethiopia
Hailu Birrie1, Teferi Gemetchu1, Fekade Balcha1, Getahun Bero 2, Meshesha Balkew 1 and Girmay
Medhin1
Abstract: Health risk assessment of a planned irrigation scheme along the Genale River, South
Ethiopia, was conducted in June 1994 covering parasitological, entomological, malacological and
environmental/ ecological parameters. Malaria due to P. falciparum and P. vivax appears to be the
main health problem. The implicated vectors are Anopheles gambiae and A. pharoensis. P. martini, the
vector of visceral leishmaniasis in the Aba Roba focus, Southern Ethiopia, is present in the study area
although the status of visceral leishmaniasis here is uncertain. Intestinal parasites, including
schistosomiasis, were very low in prevalence. The possible health consequences of the planned
irrigation scheme and precautions to be taken are discussed. [Ethiop. J. Health Dev. 1997;11(3):229233]
Introduction
Water resources development, especially in arid and semi-arid environments, may bring about
profound ecological changes which in turn may lead to the aggravation of existing diseases and their
vectors and/or introduction of new ones (1,2). Many of the most serious diseases that affect mankind
depend in one way or another on water for their transmission (2). In particular, the transmission and
spread of parasitic diseases such as malaria and schistosomiasis are invariably associated with water
impoundment (2). In Ethiopia, the development of irrigated agriculture in the Awash River Basin and
Finchaa Valley is a vivid example of how such activities may lead to almost insurmountable
schistosomiasis problem (3). For the control of the existing or prevention of the anticipated diseases
and their vectors, it is strongly advised to undertake ecological and health risk assessment surveys of
the envisaged development schemes at the initial stage.
The Oromo Relief Association (ORA), Addis Ababa, Ethiopia, is planning to undertake a stage
by stage development of irrigation schemes at Genale village located along the Genale River for
settlers, the majority of whom are Oromo refugees returning from neighboring Somalia. The plan of
ORA is to install pumps along the river to irrigate initially about 100 ha of land. The main objective of
the present study was to make health risk assessment of the development scheme planned by ORA.
Hence, parasitological, entomological, malacological and environmental/ ecological surveys were
conducted at the site in June 1994.
Methods
Study area and populations: The Genale project site is at Genale bridge (Fig.1) and is located at a
distance of 47 km east of Negele Borena town. The elevation is about 1000 meters above sea level in
______________________________________
1
From the Institute of Pathobiology, Addis Ababa University, P.O.Box 1176, Addis Ababa, Ethiopia
and 2Oromo Relief Association (ORA), Addis Ababa, Ethiopia
236 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
the basin. The communities involved are Genale village (population of about 2000) and Genale Dume
village (population of about 1000) although more settlement is envisaged in the area. Genale village is
located on the western side of the river on the Sidamo side while Genale Dume is located on the
eastern side of the river on the Bale side (Fig.1).
Figure 1: Sketch map of Genale project site
Parasitological examination: In each village every 5th house was selected by systematic random
sampling method and all residents in the selected houses were summoned for parasitological
examination.
Fresh stool specimens were collected on pieces of plastic sheets and about 1 g each transferred into
screw-capped vials pre-filled with 10% formalin. In the laboratory, the specimens were examined by
the formol-ether concentration technique (4).
Urine specimens were collected in wide-mouthed plastic vials and examined on the spot for
Schistosoma haematobium by the filtration technique (5). Examination of urine samples for
schistosomiasis haematobium was undertaken because most of the people were refugees who just
returned from somalia where the disease is known to be endemic (6).
Blood smears, thin and thick, were prepared by finger pricking of febrile and malaria suspected
cases and were examined on the spot. For confirmation, duplicate smears were also fixed with 70%
ethyl alcohol and re-examined after staining with Geimsa.
Vector survey: Mosquitoes were collected from within houses and animal quarters using suction
aspirators and killed with chloroform. They were then transported to the laboratory where they were
identified by species using appropriate keys.
Sandflies were caught using CDC light traps, sticky plastic sheets (smeared with edible oil on both
sides) and suction aspirators from likely habitats such as pipe-type termitaries, huts, tree holes, animal
burrows and rock crevices. The specimens collected on sticky sheets were recovered with camel-hair
Health risk assessment, irrigation, Genale River, South Ethiopia 237
────────────────────────────────────────────────────────────
brush and were successively washed in 10% Savlon and later in 2-3 changes of distilled water. These
specimens and those collected by other methods were then preserved in 70% ethanol and transported to
the laboratory in Addis Ababa where they were first rehydrated in water for 24 hrs and then cleared in
Nesbitt's solution for another 24 hrs. They were then mounted on a slide in chloral hydrate and
identified to species using appropriate keys.
Volunteer human baits were used for catching blackflies (Simulium spp.) at selected sites along the
bank of the Genale River.
For collection of tsetse flies, biconical traps, baited with acetone and cow urine as odor attractants,
were put up for 24 hrs at selected sites along the Genale River.
Search for snails was conducted along the banks of Genale River. Collected snail shells were
identified using appropriate keys.
Environmental/ecological observation:
Environmental characteristics such as presence/ absence of vegetation, other water bodies, termite
hills, animal burrows, rock crevices and open defecation sites were observed and recorded.
Results
Schistosomiasis mansoni: Altogether, 912 individuals from the two villages were stool-examined.
Of these only 7 (0.8%) individuals, aged above 20 years and who reported to have come from Meda
Wallabu (Bale Region) and Negele Borena (Sidamo Region), were found positive for S. mansoni ova
in their stools.
Table 1: Age-specific prevealence of intesitnal parasites among residents of Genale villages, June 1994.
Age
No.
Al.
TT
HW
TS
HN
ST
SM
EH
EC
(yr)
Ex.
0- 4
219
34(15.5)
8(3.7)
7(1.1)
3(1.4)
6(2.7)
1(0.5)
14(6.4)
30(13.7)
5- 9
224
52(23.2)
10-14
119
24(20.2)
15-19
52
7(13.5)
20-24
51
6(11.8)
25-29
44
6(12.5)
30-34
54
8(13.6)
35-39
38
4(10.5)
40-44
28
1(3.6)
45-49
15
3(20)
50+
68
3(4.1)
Total
912
141(17.1)
Al = Ascaris lumbricoides
Tt = Trichuris trichiura
Hw = Hookworm
Ts = Taenia saginata
Hn = Hymenolepis nana
24(10.7)
12(5.4)
2(0.9)
10(8.4)
11(9.2)
1(0.8)
6(11.5)
8(15.4)
4(7.7)
4(6.7)
2(3.9)
3(5.0)
1(2.1)
1(2.3)
4(8.3)
3(5.1)
2(3.4)
5(8.5)
3(7.9)
3(7.9)
3(7.9)
1(3.6)
2(7.1)
1(6.7)
(6.7)
5(6.8)
4(5.4)
8(10.8)
66(7.7)
60(7.0)
34(4.0)
Sm = Schistosoma mansoni
Eh = Entameoba histolytica
Ec = E coli
Ib = Iodamoeba butschellii
St = Strongyloides stercolaris
10(4.5)
7(5.9)
2(3.8) 1
1(2.6)
1(1.4)
27(3.1)
1(0.4)
1(0.8)
(1.9)
0(0.0)
1(1.7)
1(2.6)
5(0.6)
(1.9)
2(3.9)
0(0.0)
2(3.4)
2(5.3)
1(3.6)
(0.8)
34(15.2)
16(13.4)
11(21.2)
6(10.0)
10(20.8)
7(11.9)
5(13.2)
5(17.9)
(33.3)
13(17.6)
123(14.3)
50(22.3)
34(28.6)
18(34.6)
8(35.3) 1
13(29.5)
12(22.2)
8(47.4)
1(39.3)
4(26.7) 2
28(37.8)
231(26.3)
IB
9(4.1)
21(9.4)
12(10.1)
6(11.5)
14(23.3)
8(16.7)
4(6.8)
3(7.9)
5(17.9)
2(13.3)
11(14.9)
95(11.1)
Schistosomiasis haematobium: Only one boy, aged 10 years and who reported to have returned with
his parents from Somalia, was found positive for S. haematobium ova in his urine.
Other intestinal parasites: The species and the age specific prevalence of other intestinal parasites
in the communities surveyed are presented in Table 1. Among helminths Ascaris lumbricoides was the
most prevalent followed by Trichuris trichiura and hookworms. All helminth parasites except
hookworms, S. mansoni, and Taenia saginata were more significantly (p<0.01) prevalent among
children aged below 15 years. Protozoan cysts were more or less equally distributed in all age groups.
Malaria: Of 32 febrile and suspected malaria cases examined 10 (31.3%) children aged below 15
years were found positive for malaria (4 P. falciparum and 6 P. vivax).
Entomological findings: Anopheline mosquitoes : A. gambiae s.l., A. pharoensis, and A. marshali
238 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
were collected from within soil cracks (outdoor), animal shelters (indoor) and human baits (outdoor).
The sandflies collected were Phlebotomus martini and P. rodhaini and a number of Sergentomyia
species, mostly from shafts of termite hills.
No blackflies and tsetses were collected during the survey.
Molluscan vectors: Only shells of Bellamiya spp. and bi-valves were collected along the banks of
the river.
Environmental/ecological observation: The banks of the Genale River contained little or no microvegetation and the water was fast flowing and appeared to carry lots of silt. A small seasonal stream
entering into the River had sandy bottom and was dry at the time of the study.
Table 2: Age and Sex specific prevalence of malaria at Genale project site (both villages combined), June 1994.
Age (yr)
No. Exam
No. Positive
P. falciparum
P. vivax
Male
Female
Male
Female
Male
Female
<1
1
0
0
1
0
0
1- 4
2
2
1
0
0
1
5- 9
7
3
0
1
2
0
10-14
8
5
0
1
1
1
15+
2
2
0
0
0
0
Total
20
12
1
3
4
2
The reverine macro-vegetation was grass woodland dominated by a number of Acacia spp.
Furthermore, the area is characterized by abundant pipe-type termitaries.
Table 3: Sandfly fauna in the project site of Genale River basin, June, 1994.
Metohod
Sticky trap
CDC light trap
Aspirator
Phlebotomus martini
19
8
P. rodhaini
2
Sergentomyia
vorax
antennata
bedfordi
clydei
garnhami
ghesquerei
inermis
ingrami
kirki
kitonyii
magna
multidens
schwetzi
suberectus
Total
2
1
191
1
55
5
3
63
4
84
22
452
3
1
3
10
7
1
3
1
1
1
2
41
3
1
4
Total for all methods
27
2
5
2
194
1
65
12
1
6
1
67
4
85
25
497
Although there were no piped water supply and sanitary facilities in any of the villages, open
defecation was observed in very few sites, the probable reason being people defecating in the bushes
away from homes.
Discussion
The absence of snail hosts and the extremely low prevalence of schistosomiasis indicate that the
disease has not yet established in the Genale project site. Absence of snails along the banks of Genale
River may be related to the velocity and silt content of the water and lack of aquatic microvegetation.
However, irrigation development may create a conducive habitat for the breeding of snail hosts in the
future. Hence, with irrigation related environmental/ecological modification and increased population
movement to the newly developing area, schistosomiasis mansoni may become established. Regular
Health risk assessment, irrigation, Genale River, South Ethiopia 239
────────────────────────────────────────────────────────────
malacological and parasitological surveillance must be instituted in time to prevent it.
Although the altitude of the Genale River basin is about 1000 meters above sea level (m a.s.l.), at
the project site its microclimate, especially as it descends towards the Somali border, may make the
ecology receptive to the transmission of urinary schistosomiasis if the actual and/or potential snail
hosts are introduced. Hence, regular malacological and parasitological monitoring is required in
conjunction with that of schistosomiasis mansoni. Particular attention should be given to people
returning or coming from neighboring Somalia since the Ethiopian snail hosts of Schistosoma
haematobium are known to be susceptible to the Somalian strain of the parasite in the laboratory (7).
As regards malaria, both of the vectors of malaria in Ethiopia - A. gambiae s.l. and A. pharoensis,
and the parasites- P. falciparum and P. vivax, are present in the area. The creation of perennial
waterbodies resulting from irrigation and drainage canals will certainly create more conducive
breeding habitats for the mosquito vectors which may lead to perennial transmission of malaria unless
control is strengthened. The concern should also be viewed from the increasing occurrence of drug
resistant malaria parasites in Ethiopia (8).
The presence of A. gambiae-complex, the most important anopheline mosquito transmitting malaria
in Ethiopia (9), is of concern, since this mosquito species, by virtue of its endophilic and exophilic
tendencies, is known to show varying biting and resting habits thereby posing difficulty in its control.
This calls for the need of supplementing intra-domiciliary application of insecticides with other vector
control measures such as environmental management and larviciding. The fact that mosquitoes were
collected from animal shelters may help as zooprophylaxis as this may reduce mosquitoes entering
human habitation. Hence, the residents should be taught, in addition to other measures, to breed and
keep cattle close to their habitations. On the other hand, A. pharoensis commonly breeds in, among
others, irrigated areas (10). Hence, intermittent drainage of the canals will be important for its control.
The presence of P. martini, the known vector of visceral leishmaniasis in the Aba Roba focus,
southern Ethiopia (11) and in neighbouring Kenya (12), in the study area shows that the disease, if
introduced, could become a public health problem. According to local information the disease is said
to be endemic in Filtu district, about 150 km east of Genale project site.
No blackflies and tsetse flies were caught although the Genale River basin appears to be
ecologically conducive for their breeding and transmission of both onchocerciasis and
trypanosomiasis. The Genale River is a fast flowing waterbody bisecting a grassed woodland located at
a relatively low elevation of 1000 m a.s.l or less. The absence of the vectors during the study may
simply be related to season. Hence, more longitudinal entomological study and parasitological survey
are necessary to monitor the situation.
Compared to many areas of Ethiopia (13,14) the prevalence of intestinal parasites in the Ganale
project site are very low. The present status may be due to high temperature which results in the
desiccation of the parasite eggs, larvae and cysts. With increased population settlement and
modification of the microclimate by the irrigation to be developed the parasites may increase in
prevalence. Hence, the residents should be taught about the importance of personal and environmental
sanitation for their control.
Acknowledgment
The expenses for the study were covered by the Oromo Relief Association (ORA). The technical
assistance of Ato Abraham Redda and Ato Tadese Chane, both of whom are chief technicians of the
Institute of Pathobiology, Addis Ababa University, are highly appreciated.
References
1. Smith CEG. Changing patterns of disease in the tropics. In: Man-made lakes and human health.
Stanley NF and Alpers MP. (eds.). London, New York, San Francisco. 1975;345-363.
240 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
2. Hunter JM, Rey L and SCott D. Man-made lakes-man-made diseases. Wrld Hlth Forum
1983;4:177-181.
3. Kloos H. Water resources development and schistosomiasis in the Awash Valley, Ethiopia. Soc Sci
Med 1985;20:609-625.
4. Ritchie LS. An ether sedimentation technique for routine stool examination. Bull US Army Med
Dep 1948;8:326-9.
5. Feldmeir H, Doehring E and Daffalla AA. Simultaneous use of sensitive urine filtration technique
and reagent strips in urinary schistosomiasis. Trans Roy Soc Trop Med hyg 1982;76:416-21.
6. Arafa I. Studies on schistosomiasis in Somalia. Am J Trop Med Hyg 1975;24:280-3.
7. Hailu B, Nega B, Shibru T and Negash G. Schistosoma haematobium infection among Ethiopian
Prisoners of war (1977-1988) returning from Somalia. Ethiop Med J 1993;31:259-264.
8. Assefa N. Malaria. In: The Ecology of Health and Diseases in Ethiopia. (eds.). Kloos, H. and Zein,
A.Z. Westview Press. Boulder. Sanfrancisco. Oxford. 1993;341-352.
9. Gebre Mariam N. Malaria. In: Zein Z.Z. and Kloos H. eds. The ecology of health and diseases in
Ethiopia. Addis Ababa: Ministry of Health. 1988;136-150.
10. O'Conner CT. The distribution of anopheline mosquitoes in Ethiopia. Mosquito News
1967;27:42-54.
11. Gebre Michael T. Studies on the phlebotomine vectors of visceral leishmaniasis in southern
Ethiopia. PhD Thesis. 1992. University of London. U.K. 298pp.
12. Perkins P V, Githure J I, Mebratu Y. et al. Isolation of Leishmania donovani from Phlebotomus
martini in Baringo District, Kenya Trans Roy Soc Trop Med Hyg 1988;82:695-700.
13. McConnel E and Armstrong JC. Intestinal parasitism in fifty communities on the central high
plateau of Ethiopia. Ethiop Med. J. 1976;14:159-68.
14. Tedla S, Ayele T, Birrie H and Lo CT. intestinal helminthiasis in Ethiopia In: Ayele T, Lo CT and
Birrie H. eds. Proceedings of a symposium on human schistosomiasis in Ethiopia, November 1-4,
1982. Addis Ababa.
Original article
Establishment of two lines of Ethiopian isolates of
Plasmodium falciparum in vitro
Moges kassa1 and Robert Mshanan2
Abstract: An attempt was made to establish Ethiopian isolates of Plasmodium falciparum in tissue
culture flasks. Two lines, FCA-1 and FCA-2, were derived from two patients infected with falciparum
malaria in North and South Shewa, Ethiopia, respectively. Parasites were initiated into culture in
tissue culture flasks. Both lines grew very slowly for the first four weeks but increased their
multiplication rates and became established between the 5th and 7th weeks in culture. Both lines
produced gametocytes. Infected red blood cells cryopreserved after five weeks were easily recovered.
Sufficient parasite materials have been preserved in liquid nitrogen for later use and/or for supply to
researchers in other laboratories. [Ethiop. J. Health Dev. 1997;11(3):235-239]
Introduction
The in vitro culture system for continuous cultivation of human malaria parasite, Plasmodium
falciparum, was standardized in 1976 (1, 2). This has contributed much to the investigation of new
approaches to malaria chemotherapy, immunology, biochemistry, molecular biology, and vaccine
development, in the past twenty years. The first continuous cultures of P.falciparum were established
from the blood of an Autus trivirgatus monkey infected with P.falciparum by Trager and Jensen (1).
Later this in vitro method was adopted for the establishment of new lines of P.falciparum directly from
human infections (3).
Several workers have shown that the availability of P.falciparum from continuous cultures is
extremely useful in the screening of new antimalarial drugs (4); in investigations of the mechanisms
involved in development of drug resistance (5); and in studies of drug susceptibility and resistance of
P.falciparum in endemic areas (6). In vitro establishment of isolates of P.falciparum from different
endemic areas of Ethiopia is important in this regard.
In this study we report establishment of
Methods
Medium: The medium was prepared according to the protocol developed by Trager and Jensen (1).
That is, a stock solution was prepared by the addition of 10.4 g of RPMI-1640, 5.94 g of HEPES
powder and 0.9 ml of gentamicin (10 mg/ml) to 960 ml of distilled water. The solutions were allowed
to dissolve for at least 4 h on a magnetic stirrer and filtered using 0.22 micron millipore filter. This
can be used for washing cells. The medium was supplemented by the addition of sterile 5% Sodium
bicarbonate at the concentration of 4.2 ml to 96 ml of the RPMI-HEPES solution and 15% heat
inactivated human type AB+ serum.
Blood cells and serum: Blood cells (O+) and serum (AB+) were obtained through the kind
cooperation of the Ethiopian Red Cross Society.
______________________________________
1
From the Ethiopian Health and Nutrition Research Institute, P.O.Box 1242, Addis Ababa, Ethiopia,
and 2Armauer Hansen Research Institute, Addis Ababa
two lines of Ethiopian isolates of P.falciparum from North and South Shewa zones.
Blood cells: The blood cells were transferred aseptically into 50 ml centrifuge tubes and stored at
4oC. To prepare cells for culture, 20 ml of stored blood cells was washed twice by centrifugation
(2000 rmp, 10 min) in 2-3 volumes of stock RPMI solutions. The supernatant and buffy coat were
removed. This wash was repeated once in complete medium. The cells were finally resuspended in an
equal volume (50% cell suspension) of complete RPMI-1640 medium and stored at 4oC.
Serum: The serum was obtained from volunteers with no malaria history and who had not taken
antimalarial drugs for the last three months at the time of collection. The serum used in the culture was
242 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
heat inactivated (56oC, 40 min) and centrifuged (3500 rpm, 20 min) to remove the clotting factor. It
was then filter-sterilized (0.45 micron) and stored at -20oC. The serum was thawed and kept at
4oC before use.
Parasites: The blood samples infected with P.falciparum were collected at the Tekelehaimanot
Clinic, Addis Ababa, in April, 1994. After confirmation that patients were infected with P.falciparum
mono-infection from microscopic examination of stained thick blood films, parasitemias of the
patient's blood were determined by counting the infected RBCs on stained thin blood films. Blood
samples were obtained just before the patients were treated with antimalarial drugs. About 4 ml venous
blood was withdrawn from each patient in a sterile heparinized venoject. The samples were taken to
the laboratory (AHRI) in 1-2 h time and immediately washed as described for the blood cells above.
Sufficiently washed type O+ RBC were added to yield a parasitemia of 0.2-0.5% in preparation for
addition to culture.
Culture maintenance: P.falciparum isolates were cultured continuously according to Trager and
Jensen (1), with modification using tissue culture flasks as described by Desjardins et al. (4) at the
laboratory of AHRI. Parasites were grown in 50 ml tissue culture flasks, each containing 10 ml of
complete medium and about 0.5 ml of diluted parasitized blood to give a final blood cell suspension of
approximately 5% with an initial parasitemia of 0.2-0.5%. The flasks were set in a CO2 (5%)
incubator at 37o C. Complete RPMI-1640 medium was usually changed every 24 h and every 12 h
when the parasitemia was greater than 5%. Thin blood films were made every 48 h and stained with
Giemsa. The cultures were diluted by adding washed human red cells (O+) every four days and when
the parasitemia was greater than 5% to reduce the parasitemia back to 0.2-0.5%.
Results
As suggested by Jensen and Trager (3), the system of designation for culture lines established in our
laboratory was: FC to indicate P.falciparum cultures followed by a letter to indicate the institution
where the line was initiated in culture, followed by a number to indicate the chronological order in
which lines were placed in culture. The lines we established were initiated at Armauer Hansen
Research Institute (AHRI) on the same day on April 7, 1994. So they are designated as FCA-1 and
FCA-2. FCA-1 was collected from a girl (age 25) who contracted the disease from Debre-Berhan
(lowland) and FCA-2 from a boy (age 21) who contracted the disease from Ziway, Ethiopia.
Plasmodium falciparum, in vitro, Ethiopia 243
────────────────────────────────────────────────────────────
Table 1: Early in vitro development of Ethiopian isolate of P.falciparum, line FCA-1 in tissue culture flasksa
Days in vitro
R
0
2
4
6
6c
8
10
12
12c
14
16
16c
18
20
20c
22
24
24c
26
28
50
8
0
12
4
8
9
24
10
16
22
6
9
13
3
8
10
0
9
28
Parasite stages per 10,000 erythrocytesb
T
2N
0
15
18
14
7
25
13
16
8
10
28
8
12
36
6
13
23
7
15
32
>2N
0
9
6
4
0
3
10
9
3
8
11
3
5
12
7
0
14
4
0
3
Culture was begun on 7 April 1994
R, rings; T, Trophozoites; 2N, binucleated stage; > 2N, Schizonts with more
c
Count after the addition of fresh erythrocytes, mean of 3 seprate counts.
0
4
10
16
4
8
12
17
5
9
17
5
14
20
3
15
19
6
12
18
Total
50
36
34
46
15
24
38
66
26
43
78
22
40
81
19
36
66
17
36
79
Rate of
increase
1.4X
1.6X
1.6X
1.7X
1.7X
1.8x
1.8X
2.0X
1.9X
1.8X
2.1X
2.3X
a
b
than two nuclei. Mean of 3 different culture flasks.
Tables 1 and 2.- show the development in vitro of FCA-1 and FCA-2 lines of P.falciparum,
respectively, for the first 28 days. Both lines were initiated into culture at the same time with an initial
parasitemia of 0.5% for FCA-1 and 0.6% for FCA-2, and were handled in exactly the same way using
the same medium, cells and serum.
As illustrated in Tables 1 and 2, the early developmental rate of FCA-1 and FCA-2 is almost the
same: the initial parasitemia dropped to 0.34 and 0.30%, respectively, in the first two cycles (96 h), but
the parasitemia increased two days later (3rd cycle), at which time parasites were diluted with fresh red
cells (O+) reducing the parasitemia to 0.15% (FCA-1) and 0.17% (FCA-2). Although there was an
increase in the rate of multiplication, both lines grew at a slow rate for the first 28 days.
Table 2: Early in vitro development of Ethiopian isolate of P.falciparum, line FCA-2 in tissue culture flasksa
Days in vitro
Parasite stages per 10,000 erythrocytesb
R
T
2N
>2N
Total
Rate of
increase
0
62
0
0
0
62
2
6
18
6
8
38
4
6
16
3
5
30
6
15
12
5
9
41
1.4X
6c
6
4
2
5
17
8
11
8
0
7
26
.5x
10
14
10
5
9
38
1.5x
12
19
22
8
14
63
1.7x
12c
6
10
2
4
22
14
10
14
4
8
36
1.6x
16
20
16
7
19
62
1.7x
16c
4
5
2
8
19
18
11
8
4
10
33
1.7x
20
16
20
6
16
58
1.8x
20c
3
6
2
5
16
22
12
7
0
10
29
1.8x
24
16
11
6
18
51
1.8x
24c
3
6
1
4
14
26
8
10
2
6
26
.9x
28
12
18
0
22
52
2x
a
Culture was begun on 7 April 1994
b
R, rings; T, Trophozoites; 2N, binucleated stage; > 2N, Schizonts with more than two nuclei. Mean of 3 different culture flasks.
c
Count after the addition of fresh erythrocytes, mean of 3 separate counts
Developmental rate after seven weeks and 4 months is presented in Table 3, indicating an improved
rate of multiplication for both FCA-1 and FCA-2 after 7 weeks in culture. The multiplication rate of
FCA-1 (26x) was relatively higher than that of FCA-2 (14x). Interestingly FCA-2 (7 fold per cycle)
increased its rate of multiplication after 4 months and exceeded the rate of multiplication of FCA-1
(6.8 fold per cycle).
In general, both lines became established by increasing their rate of multiplication significantly from
26x (FCA-1) and 14x (FCA-2) in two cycles (96 h) after 7 weeks to 6.8x and 7x per cycle respectively
after 4 months, but growth of the asexual blood stages was not synchronous (Tables, 1,2 & 3).
Gametocyte production was evident in both lines starting from one week in in vitro culture, the
percentage (0.01%) of gametocyte was, however, very low.
Table 3: Developmental rates of two lines of Ethiopian isolates of P.falciparum in tissue culture flasks - line-FCA-1 and FCA-2 after 7 weeks and after 4 months in
vitro
Strain
Parasite stages per 10,000 erythrocytesa
R
T
2N
>2N
Total
Rate of increase
after 2 Cycles (96 hr)
FCA-1
7 weeks in vitro
Meanb of "0" time count
6
12
5
15
38
Meanc of 96 h count
346
70
65
02
983
26 X
4 months in Vitro
b
Mean of "0" time count
8
7
2
4
21
Meanc of 96 hr count
425
375
36
102
965
46X
FCA-2
7 weeks in vitro
Meanb of"0" time count
14
12
7
20
53
Meanc of 96 h count
285
138
62
276
761
14X
4 months in Vitro
b
Mean of "0" time count
6
8
0
4
18
Meanc of 96 hr count
372
410
22
94
898
50X
a
R, rings; T, Trophozoites; 2N, binucleated stage; > 2N, Schizonts with more than two nuclei.
b
Mean of 3 separate counts.
c
Means of 3 different culture flasks.
Discussion
There is a considerable number of culture- adapted lines of P.falciparum now available for research
purposes, including research in malaria chemotherapy (1, 3, 5, 6, 7). The
successful establishment of two lines of Ethiopian isolates of P.falciparum we are reporting here may
also be relevant as an additional source of culture-adapted parasite materials in this regard.
The successes and disappointments in an attempt to cultivate new strains of P.falciparum were
discussed by Jensen and Trager (3).
In the present work we have attempted to establish a large number of isolates from different
endemic areas of Ethiopia. However, in the process, we lost a lot of our cultures mainly due to
contamination. But we have been able to establish two lines (FCA-1 and FCA-2) of Ethiopian isolates
of P.falciparum from two different zones. The early and late development of our culture lines was not
different from that described by Trager and Jensen (3). Gametocytes were freqently and routinely seen
in asexual stage cultures of freshly isolated P.falciparum (8). However, considerable varations have
been found among isolates of P.falciparum in their capacity to produce gametocytes during in vitro
cultures (9). Gametocyte production in such culture system is believed to be induced due to lack of
nutrients and accummulation of metabolic waste products. Thus the low level of gametocytes in our
culture system could be due to more frequent change of medium and dilution of cultures with fresh
RBCs.
Parasite materials frozen after five weeks in culture were found to be recoverable. It was observed
that reducing the haematocrit to about 3-4% is necessary when retrieving frozen parasite materials in
preparation for culture. It has been suggested that some human sera, especially in areas endemic for
malaria, are unsuitable for use in P.falciparum cultures (10). Thus, the interview approach used to
avoid the presence of immune serum in our culture may be of value in this regard. Enough copies of
the parasite materials have now been preserved in liquid nitrogen for later use or supply to workers in
other laboratories. We suggest the establishment of new isolates from different endemic areas since
studies of their drug response characteristics could play an important role to fully understand the
malaria problem in these areas.
Acknowledgement
This work at Armauer Hansen Research Institute (AHRI) was supported by the Ethiopian Science
and Technology Commission. We wish to thank AHRI, the Department of Traditional Medicine and
the Ethiopian Red Cross Society for their collaboration. Our Appreciation also goes to Dr. Yemane
Teklai whose dedication made this work possible.
References
1. Trager W and Jensen JB. Human malaria parasites in continuous culture. Science 1976;193:674675.
2. Hynes JD, Diggs CL, Hines FA and Desjardins RE. Culture of human malaria parasites of
Plasmodium falciparum. Nature (Lond.) 1976;263:767-769.
3. Jensen JB and Trager W. Plasmodium falciparum in culture. Establishment of additional strains.
Am J Trop Med Hyg. 1978;27:743-746.
4. Desjardins RW, Canfield CJ, Haynes JD and Chulay JD. Quantitative assessment of antimalarial
activity in vitro by a semiautomated microdilution technique. Antimicrob. Agents
chemother.1979;16:710-718.
5. Nguyen-Dinh P and Trager W. Chloroquine resistance produced in vitro in an African strain of
human malaria. Science 1978;200:1397-1398.
6. Thaiothong S and Beale GH. Resistance of ten Thai isolates of Plasmodium falciparum to
chloroquine and pyrimethamine by in vitro tests. Trans R Soc Trop Med Hyg. 1981;75:271-274.
7. Thaithong S and Beale GH. malaria parasites. Chulalongkorn University Research Report series 1.
Bangkok, Thailand, 1992.
8. Jensen JB. Observations on gametogenesis in Plasmodium falciparum from continuous cultures.
Journal of Protozoology. 1979;26:129-132.
9. Ponnudurai T, Meuwissen JHE Th, Leeuwenberg ADEM, Verhave JP and Lensen AHW. The
production of mature gametocytes- of Plasmodium falciparum in continuous cultures of different
isolates infective to mosquitoes. Trans R Soc Trop Med Hyg. 1982;76:242-250.
10. NRMI/USAID/WHO. Workshop on Immunology of Malaria, Bethesda, MD, USA, 1979. Bull
Wld Hlth Org. 1979; 57: (Suppl.1): 1-288.
Original article
Morbid grief I: Are close relatives of the "redterror" victims of Addis Ababa still suffering from a
morbid grief and other complications of
bereavement?
Abdulreshid Abdullahi Bekry1 and Mohammed Haji Hyder Ali 2
Abstract: To assess whether close relatives of the "red-terror" victims of the Ethiopian revolution of
late 1970's are still suffering from a morbid grief and other complications of bereavement, a study was
carried out in Addis Ababa, Ethiopia, between February and May 1995. Ninety one close relatives
(sample I ) have rated themselves with the Texas Inventory of Grief 18 years after bereavement and
89(97.8%) of them were found to have a profound grief reaction. Compared to 87 non-bereaving
control group (sample II), they have scored significantly higher (P <0.001) on the General Health
Questionnaire-30 items version (GHQ-30), Beck Depression Inventory (BDI) and State Anxiety
Inventory (SAI). The correlation (r=+0.843, P<0.001) between the brief (BTIG) and the expanded
(ETIG) forms of Texas Inventory of Grief has indicated that the two forms are equally reliable and
useful in measuring grief reaction. All the rating scales are self-rating, Amharic translated with
acceptable face validity, but they are not yet concurrently validated with their corresponding English
versions. BTIG was recommended as a useful paper-and-pencil screening instrument with an arbitrary
cut-off point of 16.24 (i.e. - 1.96 Z- score) for this particular group and it was suggested that
respondents scoring 16.24 and above which amounts to 89(97.8%) be considered as probable cases of
morbid grief and as candidates for bereavement counselling. [Ethiop. J. Health Dev. 1997;11(3):241249]
Introduction
Grief is the usual response to the loss of a loved one. In an adult it follows uniform pattern that is
partially modified by the bereaved premorbid personality, the importance and abruptness of the loss,
and the presence of other recurrent stress (1). Freud (2) stated that mourning is a grave departure from
the normal attitude of life, but he neither regarded it as a pathological condition nor referred it for a
medical treatment. He felt that normal grief was self-limiting, was resolved by reality, and that
interference with it was useless and may even be harmful. Lindemann (3) considered uncomplicated
grief to be a definite syndrome with pathognomonic symptoms and a predictable course.
Clayton (4), in a study of normal bereavement had stated that there are only three symptoms:
depressed mood, sleep disturbance and crying that more than one-half of the subjects experience.
Three other symptoms: difficulty of concentrating, loss of interest in TV and news, and anorexia and/or
weight loss occur frequently, but still in less than half the subjects interviewed. He concluded that as
98% of those studied did not seek psychiatric assistance during the bereavement period, bereavement
is a psychological reaction rarely handled by the psychiatrist.
Normal grief commences within a 2 week time frame after a major loss of a significant attachment
figure. The bereaved individual demonstrates phasic response that involves a general, but not fixed
Morbid grief I 247
────────────────────────────────────────────────────────────
progression of phenomena. This is followed by a recovery progressively over the first year. The
______________________________________
1
From the Department of Psychiatry, Addis Ababa University, P.O.Box 9086, Addis Ababa, Ethiopia
and 2Amanuel Hospital, P.O.Box 1971, Addis Ababa, Ethiopia.
bereaved person can then recall the deceased person without subjective pain and feels that he or she
has grieved and has accepted the loss (5).
In the normal course of events the intensity, frequency and duration of grief reaction diminish with
time. This period of time of normal grief varies but rarely is more than six months according to the
western cultures. Grief beyond six months is arbitrarily considered to be prolonged (1). However, as
each relationship is different, each grief is also different and it is not possible to adopt a rigid time
frame for bereavement (6).
Anniversaries of the loss and significant holidays are critical points in time : there may be a transient
exacerbation of symptoms that last few days and sometimes this recurrent pattern is misconstrued as
cyclical bipolar disease (5). Double (multiple) loss seems to have an accumulated effect and this is
referred as ‘bereavement overload’ (7).
The complications of bereavement are (1,8):1. Pathological (morbid) grief which includes delayed, absent, severe and chronic grief,
2. Depression,
3. Anxiety disorders,
4. Alcohol or drug abuse, and
5. Mania- reported in persons with history of Mood Disorder.
The descriptive approach is generally not sufficient or satisfactory to differentiate normal from
pathologic bereavements as some of the characteristics are found in both forms of bereavement. The
relationship between the two forms is more of a continuous and the pathology is more related to the
intensity of a reaction or the duration of a reaction rather than to the simple presence or absence of a
specific behaviour (8).
Instruments have been developed to measure the intensity of symptoms of grief and bereavement
(9,10,11,12). These instruments help to identify cases of unresolved (morbid) grief and allow reliable
and valid description of the grief process. Concerning the duration of the grief process, pathological
grief as long as 10,15 and even 25 years was reported in some communities and it was further stated
that some aspects of ‘grief work’ may never end for a significant proportion of bereaved individuals
(10).
Numerous studies have identified risk factors for developing complications of bereavement
(1,4,11,13). We believe that our subjects who are close relatives1 of victims of the so-called ‘redterror’ of the Ethiopian revolution of late 1970's have encountered many of the risk-factors that could
complicate this bereavement process. Our day-to-day experience in the community and especially in
the psychiatric OPDs reminds us that the loss of loved ones in the ‘red-terror’ was the most commonly
mentioned event by the bereaved who were left with painful response on recalling the deceased. So far
no effort was made to assess the magnitude of complication of bereavement among the close relatives
of ‘red-terror’ victims. The purpose of this study was to find out whether close relatives of the ‘redterror’ victims suffer from complications of bereavement in general and in particular, to :
1. measure the intensity of their pathologic (morbid) grief;
2. find out whether they could be identified as probable cases (i.e. non-psychotic) with the GHQ-30
and score significantly higher than the control group on the GHQ-30, on the depression and the
anxiety inventories.
Based on this study a proposal will be submitted to set up a separate bereavement counselling
services in Addis Ababa and other cities for those who still suffer from the morbid grief. Close
relatives of the victims have already formed associations in Addis Ababa and other cities to look after
their psychosocial problems.
248 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
_____________________________________
A close relative includes here a first degree relative (i.e. father or mother, brother or sister, husban or
wife, son or daughter) or anybody with whom a respondent has lived for years and considered to be a
family member.
The hypothesis proposed was that our subjects still suffer from a morbid grief and would score
significantly higher than the control groups on GHQ-30, a depression inventory and an anxiety
inventory.
The design of this study also enabled us to elucidate two separate issues:
1. The phenomenology of pathologic grief processes among our subjects and the relationship of the
morbid grief scores, as measured by the BTIG, to other psychological constructs such as GHQ30, BDI and SAI scores. This is dealt with in paper II (21).
2. The influence of variables on the degree of grief reaction, depression and anxiety among our
subjects and this is dealt with in paper III (22).
Methods
Selection of Samples: Sample I (i.e. close relatives of ‘red-terror’ victims) was drawn randomly
from the list of families who suffered a loss or losses during the ‘red-terror’. The list was obtained
from the Addis Ababa branch of Human Rights League for the Victims of Red-Terror which has a
representative from all the 28 ‘Woredas’ (constituencies).
Sample II (i.e. control group) also was drawn randomly from among volunteers from each ‘woreda’.
To minimize the bias, which could be a source of weakness, each ‘Woreda’ representative had first
registered a number of volunteers (controls) who could match each of the candidates in sample I
(cases). Then, one control was selected for each case by drawing a name from a hat.
At least five subjects were expected to join each sample from each ‘woreda’ to make a total of 140
subjects in each sample. As much as possible, the control group from each ‘woreda’ was matched with
respect to variables like age, sex, occupation; etc. before the selection.
Excluded from the study, were those known to be psychiatric cases, drug addicts or suffering from a
severe physical illness or having current social, legal or economic crises. Among sample II, subjects
who lost a close relative in the last two years were also excluded.
All representatives of the ‘woredas’ have attended briefing sessions and panel discussions on the
study and given instructions about obtaining random samples specifically about drawing numbers from
a hat, about selecting matched control sample and about the time-table for distribution of the selfreport questionnaires. The authers are aware of the shortcomings of these and other self-report
questionnaires: they introduce reporting bias if the subject is too disturbed to complete the task; can
not read or understand the intent of the questions; or is motivated in some way to falsify or exaggerate
his responses. But they have clear advantage in saving professional time and expense. They are also
more sensitive in evaluation of subjective distress in subjects who are not very disturbed. For those
who can not read and write, a literate relative was permitted to read questions for the subject and fill
the questionnaires appropriately.
Both samples have filled first the demographic data sheet about themselves and sample I (cases)
have filled an additional data sheet about relatives lost in the ‘red-terror’.
Test materials and statistical analyses:
1. Expanded Texas Inventory of Grief (ETIG) and Brief Texas Inventory of Grief (BTIG)(9,10,11):
ETIG is a 34-item scale which also includes all the seven items of BTIG. Both are self-rated and each
item is scored on scale of 1-5, with five being assigned to a response seemingly most indicative of
unresolved grief. The ETIG generates a score between 34 and 170 and the BTIG between 7 and 35.
Both are valid and reliable, ETIG being suitable as a tool for assessing the outcome of the grief as
‘good’ or ‘bad’ and BTIG as a paper-and-pencil screening tool. Only sample I has completed ETIG.
Distribution of ETIG score of the group, mean score of the group on each of the 34 items of ETIG and
the percentage of positive endorsement of each item was calculated. To test the correlation between the
1
Morbid grief I 249
────────────────────────────────────────────────────────────
two forms of Texas Inventory of Grief scatter diagram was drawn, correlation coefficient (r) was
calculated and the regression line was drawn.
In similar manner, correlation between BTIG and GHQ-30, BTIG and BDI and BTIG and SAI was
tested. Also correlation between GHQ-30 and BDI, GHQ-30 and SAI and BDI and SAI was tested.
The results of these correlation tests and the phenomenological analyses of the positively endorsed
items of ETIG will be dealt with in paper II (21).
To know what variables of the bereaved and of the victims among sample I could influence the
degree of grief reaction, depression and anxiety, t-tests or Z-tests were applied to detect significant
differences in ETIG, BDI and SAI mean scores between the different variables. The result will be
discussed in paper III (22).
2. General Health Questionnaire (GHQ-30 item version) (14,15): The GHQ-30 is a self -rated
questionnaire which discriminates accurately between probable `cases' and 'non- cases', but it is not
intended to detect psychoses. It has a cut-off point of 4/5 and generates a score between 0 and 30 by
the standard GHQ scoring method. The mean GHQ-30 scores of sample I (cases) and Sample II
(controls) were tested with Z-test for a significant difference. The proportions above the cut-off point
were also tested for a significant difference by X2 test.
3. Beck Depression Inventory (BDI) (16): This is also a self-rated inventory with 21 items and each
item is rated (0-3). It generates a score between 0 and 63. It is a valid and a reliable inventory which
measures the depth of a depression.
The mean scores of sample I (cases) and Sample II (controls) on BDI were tested for a difference of
significance with Z-test.
4. The State Anxiety Inventory (SAI) (17):
This inventory too is self-rated and has 20 items and each item is rated (1-4). It generates a score
between 20 and 80. It is a valid, reliable and sensitive indicator of changes in the level of anxiety.
The mean scores of sample I and sample II on SAI were tested for a difference of significance with
Z-test.
All test materials mentioned above were carefully translated into Amharic: first, English-Amharic
dictionary (Oxford University Press) was used for a consistent translation of key words of each
question and statement that indicate the different degrees of severity of symptoms. Then, the over all
sense of each question and statement was examined to make sure that it has identical meaning with that
of the English version. Finally, after further consultations with colleagues, the authers were convinced
that the Amharic versions of the questionnaires have acceptable face validity and can measure
satisfactorily what they are supposed to measure. As these Amharic versions are not yet statistically
validated in our setting, the authers believe that separate researches have to be carried out to assess
their concurrent validity and this will be a step foreword towards standardization of these versions. All
statistical tests reported were two-tailed.
Results
Ninety eight sets of forms have been distributed to sample I (cases) and 91(92.86%) have returned
the completed forms. To sample II (controls), 93 sets of forms (ETIG excluded) have been distributed
and 87 (93.54%) have returned the completed forms. The rest were returned incomplete or empty.
250 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Figure 1: Scatter diagram of relation in 91 bereaved cases between ETIG and BTIG, Addis Ababa, 1996.
Table 1 shows the demographic variables of the two samples. It is evident from the table that there
is a preponderance of younger, male, single, educated and the employed among respondents of sample
II (controls). Therefore, the two samples can not be taken to be alike for the purpose of comparison.
However, a stratified sample is created in almost all variables. As a solution to this problem, the
authors decided to compare the whole sample at each stratum for a significant difference, at least on
BDI mean scores. The results of these analyses are incorporated for convenience into Table 1 and will
be dealt with in the discussion.
Table 1: Demographic variables of sample I (cases) and sample II (controls) (with their mean scores on BDI and significance levels incorporated for
convenience), Addis Ababa, 1996.
Variables
Sample I
(N=91)
(cases)
Mean score
Sample II
(N=87)
N(%)
(controls)
Mean score
16(17.58)
20(21.98)
9(9.89)
13(14.29)
20(21.98)
12(13.19)
1(1.1)
18.63
22.15
24.11
23.77
24.95
26.88
(34)
44(50.57)
22(25.25)
8(9.20)
7(8.05)
4(4.60)
1(1.15)
1(1.15)
11.95
16.68
14.38
10.57
20.5
(24)
(27)
49(53.85)
42(46.15)
23.78
22.52
61(70.11)
26(29.89)
13.34
14.35
<0.001
<0.02
Christ.
Muslim
Unspec.
Marital status:
80(87.91)
11(12.09)
0(0.0)
24.13
19.64
--
79(90.80)
5(5.75)
3(3.45)
13.84
18.6
13.0
<0.001
NS
--
Single
Married
Divorced
Separated
Widow/er
Unspec.
Educational level:
Illitr.
Primary
Second.
Tertiary
Vocational
Unspec.
Employment Status
23(25.27)
40(43.96)
3(3.3)
3(3.3)
21(23.08)
1(1.1)
19.30
24.28
41.33
31.00
21.00
(8)
46(52.87)
28(32.18)
3(3.45)
4(4.6)
6(6.9)
0(0.01)
13.22
14.18
16.00
10.00
23.33
--
<0.05
<0.01
<0.05
NS
NS
--
5(5.49)
40(43.96)
22(24.18)
3(3.3)
14(15.38)
7(7.69)
27.8
23.1
22.09
23.00
18.29
31.00
0(0.0)
16(18.39)
38(43.68)
11(12.64)
22(25.29)
0(0.0)
-17.94
14.39
12.25
11.23
--
-NS
<0.01
NS
<0.01
--
Employed
Pensioned
Dependent
S-employed
Unspec.
22(24.18)
31(34.07)
29(31.87)
5(5.49)
4(4.4)
17.59
21.94
27.10
40.00
20.75
46(52.87)
8(9.2)
27(31.03)
1(1.15)
5(5.75)
11.65
13.5
17.0
(9)
18.6
<0.05
<0.02
<0.01
-NS
N(%)
Age distribution:
20-29
30-39
40-49
50-59
60-69
70-79
unspec
Sex:
Male
Female
Religion:
Significance*
<0.02
NS
NS
<0.05
NS
---
P<
Morbid grief I 251
────────────────────────────────────────────────────────────
NS = Not significant
Unspec. = Unspecified
S.employed = self employed
* Z- test or t- test was performed depending on the number of respondents at each stratum.
Table 2 shows distribution of only ETIG scores of sample I (cases). Their mean ETIG score was
126.36 (range 74 to 158, SD=21.42) and on BTIG their mean was 26.88 (range 13 to 35, SD=5.43).
Table 3 shows the 34 items of ETIG and the mean scores of each item. The seven items of BTIG are
marked by asterisk (*). The overall mean score for all items of ETIG was 3.72 (range 1.71 to 4.81,
SD=0.88) and that of BTIG was 3.84 (range 2.44 to 4.64, SD= 0.77). No significant difference
between these two overall severity mean scores (P >0.5, t= 0.3349, DF=39).
When the results of ETIG and BTIG of individual respondents were plotted out (Fig 1), they
correlated fairly well with one another (r= +0.843, P<0.001) and indeed the correlation is strongly
positive and very highly significant.
Table 4 indicates the mean scores of the two samples on GHQ-30, BDI and SAI. On all scales the
mean scores of sample I are higher than the corresponding mean scores of sample II (P<0.001 in all)
and, indeed, these differences are very highly significant.
Further more, in GHQ-30 the proportion above the 4/5 cut-off point of probable `caseness' in
sample I (62 or 68.13%) is higher than in sample II (controls) (42 or 48.28%) and this difference too is
highly significant (P<0.01).
Discussion
The random selection of sample I respondents resulted in a reasonably fair distribution of different
age and sex groups and other variables. But among respondents of sample II, there is a preponderance
of younger, male, single, educated and employed as mentioned earlier. The solution to the bias thus
created by sample selection will be mentioned towards the end of the discussion.
Despite the long duration (18 years) between the loss of a close relative in the 'red-terror' and the
completion of the questionnaires, almost all respondents of sample I felt strongly affected by the loss.
The mean score of 126.36 on ETIG and 26.88 on BTIG and the overall mean score for all items on
ETIG (3.72) and on BTIG (3.84) indicate that the degree of their grief reaction is quite high.
Theoretically, it is expected that the magnitude of grief will diminish as a function of time from death
(9), but it remained high among those who lost a close relative in the ‘red-terror’. There are several
factors that could contribute to the ‘poor’ outcome of the individual grief reaction, but as a group the
circumstances surrounding the loss were painful, horrifying and mismanaged. The victims were young
(83 or
91.2% were below 30 years and 29 or 31.9% were below 20 years). In most cases there was lack of
funeral rituals which actually facilitates grief and helps to accept the reality and finality of death. It also
helps to express thoughts and feelings about the deceased and draws social support net work to the
bereaved.
Table 2: Distribution of ETIG* scores of sample I (cases), Addis Ababa, 1996.
Range of scores
No
%
<70
0
0.0
70-79
2
2.198
80-89
2
2.198
90-99
12
13.187
100-109
6
6.593
110-119
4
4.396
120-129
19
20.879
130-139
18
19.780
140-149
16
17.582
150-159
12
13.187
160-169
0
0.0
252 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
91
Mean =126.36, SD= 21.42
* Range of score that could be generated is (34 to 170)
100
The measure of correlation (r=+0.843, P<0.001) between the two grief inventories (ETIG and
BTIG) indicates that these are two forms of the same test and that the BTIG is equally useful and
reliable in measuring a grief reaction. Moreover, lack of significant difference between the overall
mean scores of all items of both inventories indicates that they measure a grief reaction with a similar
accuracy.
This implies that BTIG can serve as a quick paper-and-pencil screening tool for measuring the
magnitude of unresolved grief. Moreover, all the seven items of BTIG were positively endorsed by
respondents ranging from 96.2% to 97.8% (see paper II) (21).
All questions in the ETIG or BTIG refer specifically to the deceased person and the data obtained
consist of a constellation of symptoms relevant to the bereaved individuals. This means that these
inventories have high face validity in measuring grief. The issues of concurrent validity were already
mentioned under methodology.
It has to be clear that we are not trying to delineate a cut-off point between normal and abnormal
grief. Any such attempt has to take into consideration the severity of grief over time as grief will
diminish as a function of time from death. We are dealing with cases of abnormal grief (i.e. chronic
and possibly severe) and one of the purposes of this study is to propose counselling services for these
chronic mourners. Recommending all of them for counselling is quite unacceptable. Those who score
less than 1.96 standard deviation below the sample mean should be taken arbitrarily as suffering
significantly less.
Therefore, we would like to suggest an arbitrary cut-off point on the BTIG for identifying cases of
unresolved grief among close relatives of the ‘red-terror’ victims. This arbitrary cut-off point should be
(-1.96 Z-score), which is 16.24 on BTIG (18). Those who score 16.24 or above on BTIG should be
regarded arbitrarily as candidates for counselling for unresolved grief. This amounts to 89 (97.8%)
respondents of sample I.
Significantly higher mean score and higher proportion of respondents above the cut-off point of
probable ‘caseness’ on GHQ-30 indicate that close relatives of victims of the ‘red-terror’ (sample I) are
probably in higher distress than sample II. However, the GHQ-30 has failed to identify 29 (31.87%)
respondents as probable 'cases' from sample I. This subgroup of 29 has actually scored an average of
120.66 on ETIG, 15.034 on BDI and 40.103 on SAI, meaning that they too suffer from grief reaction,
depression and anxiety. It is known that the GHQ gives greater number of false-negatives when used in
chronic disorders as in chronic grief, or it is only efficient in detecting disorders of recent onset
(19,20).
Table 3: Response of 91 close relatives of the ‘Red-Terror’ victims (sample I or cases) to the Expanded Texas Inventory of Grief (ETIG), Addis
Ababa, 1996.
Item
Mean Score**
F 1
I cry inside for him/her
4.69
F 2
I still get angry when I think of him/her
4.66
F 3
since he/she died, I am more like him/her
2.89
F 4
I feel guilty when I think of him/her
1.71
*F 5
I am preoccupied with thoughts of him/her
4.64
F 6
I feel it is unfair
4.81
F 7
I feel he/she is stil with me at times
3.65
F 8
I have acquired the habits and interests of him/her
2.74
F 9
I have found someone to take his/her place
1.74
F 10
I would feel better if I could really cry
3.60
F 11
No one will ever take his/her place in my life
4.56
F 12
I hide my tears when I think about him/her
3.79
F 13
I have to laugh when I think about him/her
4.13
F 14
Now I can talk about the person without discomfort
1.81
F 15
Sometimes I dream about him/her
4.20
F 16
I cry when I think about him/her
4.37
F 17
A numbness comes over me when I think of him/her
4.22
Morbid grief I 253
────────────────────────────────────────────────────────────
F 18
I feel physically ill when I thimk of him/her
4.04
F 19
I feel I have adjusted well to the loss
3.55
F 20
I have never known a better person
3.81
*F 21
I canot accept his/her death
4.01
F 22
I am now functioning as well as before
3.03
*F 23
I get upset when I think about him/her
4.37
F 24
Things and people around me still remind me of him/h
4.41
F 25
I very much miss the person
4.40
F 26
It is painful to recall memories of him/her
4.35
F 27
I try to avoid thinking of him/her
3.29
*F 28
I feel just like the person who died
3.18
F 29
My health has not been good since he/she died
4.15
*F 30
I still fell the need to cry for him/her
4.23
*F31
I get upset each year about the time that he/she died
4.01
F 32
I can’t avoid thinking about him/her
4.29
F 33
I feel I have the same illness as him/her
2.62
*F 34
I have pain in the same area of my body as him/her
2.44
* The seven items of BTIG ** 1=completely false, 2=mostly false, 3=partly true and partly false, 4=mostly true and 5=completely true
Over all mean score for all items:
ETIG=3.72 (range 1.71 to 4.81, SD=0.88)
BTIG=3.84 (range 2.44 to 4.64, SD=0.77)
Significantly higher mean BDI score indicates that close relatives of the ‘red-terror’ victims are
more depressed than the control group. The difference in SAI scores also shows that close-relatives of
‘red-terror’ victims suffer from a significantly higher magnitude of anxiety than the control group.
At this point one might challenge that though sample I suffers from a severe form of unresolved
grief as mentioned earlier, significantly higher scores than sample II on GHQ-30, BDI and SAI could
not be attributed only to the effect of the unresolved grief as the two samples are not identical in all
respects. The authors would like to explain this issue as follows: 1. Both samples were selected
randomly by drawing names from a hat. Eventhough respondents in both samples were known to
‘woreda’ representatives, this has not improved the chance of getting ideally stratified matched
samples. 2. Sample I which has suffered the loss 18 years ago and the relatively newer generation from
which the different strata of sample II were drawn do not necessarily match and in such circumstances
bias in allocation can not be eliminated. 3. The solution to this problem was already mentioned under
result and here it will suffice if the reason why BDI alone was chosen for these analyses is elaborated.
The defect of using the GHQ-30 in chronic grief was already stated. Furthermore, if it is used to
compare respondents at each stratum, the numbers will be much smaller and this leads to larger
‘standard error of difference’ which leads to low t-values / Z-values and makes the difference in the
mean scores insignificant or less significant. The correlation between BDI and SAI (r=+0.647,
P<0.001) as shown on paper II indicates that these two inventories are closely linked to each other and
therefore the authors feel that it is unnecessary to repeat the analyses with SAI (21).
Now, referring back to Table 1 it is evident that respondents of sample I have higher mean scores
than sample II on BDI at all strata except that widows/ers have slightly less mean score which is not
significant. Most of the differences are significant except where the number of respondents is too few.
The disadvantage of having fewer respondents in statistical tests was already mentioned. The authors
believe that increasing the number of respondents at each stratum would help to bring the difference to
a significant level.
American studies (8) have shown that the risk of clinical complication one year after a
Table 4: Mean score of the two samples (cases and controls) on GHQ-30, BDI and SAI and their level of significance, Addis Ababa, 1996.
Test material
GHQ-30 score
GHQ-30 score(5+)
Sample
Mean
11.90
-
BDI Score
23.31
SAI
51.95
SD= Standard Deviation
I
(cases)
SD
9.68
-
(n=91
Range
0-30
13.07
15.19
0-55
22.80
N(%)
62(68.13)
Sample
II Mean
6.29
-
(controls)
SD
6.49
-
(n=87)
13.97
43.16
10.60
11.28
0-55
21-80
0-3
-
N(%)
Range
42(48.28)
Remarks
Z=4.59, P<0.01
X2=7.219, DF=1,
P<0.001
Z=5.247,P<0.001
Z=4.191,P<0.001
254 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
loss is (4-49%) for any disorder, (14-34%) for pathologic grief,(4-31%) for major depression and 39%
for panic disorders. Bereavement accounts for 15% of admissions to psychiatric wards and 20% of
consultations from medical and surgical wards. Such an estimation of risk of complication and load to
surgical or medical consultations due to bereavement is not available in our country and probably the
awareness of this problem is minimal in general.
In conclusion, since the Amharic versions of self-rating scales used in this study are not yet
concurrently validated with the corresponding English versions, their face validity have to be taken
into consideration. Also, this study can not be absolutely immune from other short- comings like
sampling bias or reporting bias. Keeping these in mind, the results of this study indicate that close
relatives of the ‘red-terror’ victims still suffer from a morbid/ pathological/ unresolved grief. The
results also indicate that they are more distressed, more depressed and more anxious than the control
group. All these differences are statistically very highly significant. The authors believe that setting up
bereavement counselling services for those who suffer would help in resolving the chronic grief and
other complications like depression and anxiety. Once such counselling services are established, they
can extend their services even to those whose bereavement is unrelated to the ‘red-terror’.
The Amharic version of BTIG is as reliable as ETIG and can serve as a short paper-and-pencil
screening instrument for measuring the extent of unresolved grief. For reasons already mentioned, we
would like to suggest a cut-off point of 16.24(i.e-1.96 Z-score) (18) and respondents scoring 16.24 and
above among close relatives of the `red-terror' victims should be identified as probable cases of morbid
grief and should be considered as candidates for bereavement counselling.
Acknowledgement
We gratefully acknowledge the Ethiopian Science and Technology Commission for it's financial
assistance. We are also thankful to the Addis Ababa branch of Human Rights League for the Victims
of red-terror for its help in distribution and collection of questionnaires.
References
1. De Vaul RD, Zisook S. Unresolved Grief: Clinical Considerations. Postgrad. Med. 1976;59:267-71.
2. Freud S. 'Mourning and Melancholia' (retracted by Clayton PJ, Desmarias L, and Winokur G. in
Amer. J. Psychiat. 1968;125(2):168-78) SE 1917;14:243-58.
3. Lindmann E. Symptomatology and management of acute grief (retracted by Clayton PJ, Desmarias
L, and Winokur G. in Amer. J. Psychiat. 1968;125(2):168-78) Amer.J. Phychiat. 1944;101:141-9.
4. Clayton PJ, Desmarias L, and Winokur G. A study of normal bereavement. Amer. J. Phychait.
1968;125(2):168-78.
5. Jacobs SC. Diagnostic criteria for normal grief. In : Jacobs SC, editor. Pathologic grief:
Maladaptation to loss. Washington DC: American Psychiatric press, 1993;363-5.
6. Wikan U. Bereavement and loss in two Muslim communities : Egypt and Bali compared.
Soc.Sci.Med,1988;27(5): 451-6.
7. Wardon JW. Abnormal Grief reactions: complicated mourning. In: Wardon JW, editor. Grief
counselling and grief Therapy : A handbook for the mental health practitioner. N.Y: Springer
Publishing Company, 1991;65-78.
8. Jacobs SC, The complication of bereavement. In: Jacobs SC, editor, Pathologic grief: Maladaptation
to loss. Washington DC: American Psychiatric press, 1993;39-54.
9. Faschingbauer TR, De vaul RD, Zisook S. Development of the Texas Inventory of Grief. Am.
J.Psychat. 1977; 134(6):696-8.
10. Zisook S, De Vaul RD Click, MA. Measuring symptoms of Grief and Bereavement. Am.J. Psychat.
1984; 139 (12): 1590-3.
11. Lundin T. Long-term outcome of bereavement. Brit. J. Psychait. 1984;145:424-8.
12. Liberman S. Nineteen cases of morbid grief Brit. J. Psyachiat. 1978;123:159-63.
13. Park CM. Bereavement. Brit. J. Psychiat. 1985;146:11-17.
14. Goldberg DP. The detection of psychiatric illness by questionnaire. Maudsley monograph 21.
London : Oxford University press, 1972.
15. Vachon MLS, etal. A controlled study of self-help intervention for Widows. Am. J. Psychait.
1980; 137(11):1380-84.
16. Beck AT, etal. An Inventory for measuring depression. Arch. Gen. Psychiat. 1961;4:561-71.
17. Spilburger CD, Gouruch RH, Lushene R. State- trait anxiety inventory manual. Palo Alto,
California: Consulting psychologist Press, 1970.
18. Cohes W, Holiday M. Statistics for social scientists. London : Harper and row Ltd, 1982;51-63.
19. Finlay-Jones RA, Murphey E. Severity of psychiatric Disorder and the 30- items General Health
Questionnaire. Brit. J. Psychiat. 1979;134:609-16.
20. Benjamin S, Decalmer P, Varan D. Community screening for mental illness: A validity study of
the General Health Questionnaire. Brit. J. Psyciat. 1982;140:174-180.
21. Abdullahi AB, Hyder MA. Morbid grief II: The phenomenology of pathologic grief process,
depression and anxiety among close relatives of “red-terror” victims. Ethiop J Health Dev.
1997;11(3):251-256.
22. Abdullahi AB, Hyder MA. Morbid grief III: The influence of variables on the degree of grief
reactions, depression and anxiety among close relatives of the “red-terror” victims. Ethiop J Health
Dev. 1997;11(3):257-261.
Original article
Morbid grief II: The phenomenology of pathologic
grief process, depression and anxiety among close
relatives of ‘red-terror’ victims.
Abdulreshid Abdullahi Bekry1, Mohammed Haji Hyder Ali 2
Abstract: To study the phenomenology of morbid grief and it's association to general distress,
depression and anxiety, 91 randomly selected close relatives of the ‘red-terror’ victims completed four
sets of questionnaires. All the questionnaires are self-rating, Amharic translated and with acceptable
face validity, but not yet concurrently validated with their corresponding English versions.
The percentage of positive endorsement and the mean score of each of the 34-items of ETIG
(Expanded Texas Inventory of Grief) has ranged from 85.7% to 100% and from 1.71 to 4.81
respectively, indicating high degree of morbid grief. Items indicating ‘good outcome’ were found to
show the opposite, i.e. ‘bad outcome’ even 18 years after the bereavement. The syndromes that belong
to the complications of grief reaction are vivid and circumscribed. The magnitudes of endorsement and
the mean scores of some of the items appear to be characteristic of the nature and circumstances of the
loss and appropriate interpretation is necessary. The correlations between BTIG mean scores and the
mean scores of GHQ-30 (General Health Questionnaire-30 item version), BDI (Beck Depression
Inventory), and SAI (State Anxiety Inventory) were found to be weak, but significant. The GHQ-30,
the BDI, and the SAI mean scores were found to have moderate to strong positive correlation
coefficient to one another indicating common linkage they have to the pathologic grief reaction.
[Ethiop. J. Health Dev. 1997;11(3):251-256]
Introduction
Sigmond Freud and Karl Abraham (1) tried to differentiate normal from pathologic grief early in the
century and their approach was descriptive. They have mentioned that certain characteristics as being
common in normal grief while other characteristics were common to pathologic grief. But subsequent
studies have indicated that some of the characteristics are found both in normal and pathologic forms
of grief and at present this is a fairly common experience. The relationship between normal and
abnormal grief reaction is seen today as a continuous one and what determines pathology is the
intensity of a reaction or the duration of a reaction rather than simple presence or absence of a specific
behaviour (2).
Chronic (morbid) grief commences within two weeks of the major loss of a significant attachment
figure. The bereaved individual fails to demonstrate a phasic response with progression into any one of
several of these phenomenological patterns without diminution of intensity of the response with time
for months or years. The bereaved individual's whole existence is dominated by the ongoing grief for,
and focus on, the lost individual, often to the extent that other relationships and functionings are
significantly impaired (3). Sometimes the grief could be exaggerated and clinical depression or anxiety
or alcohol or substance abuse may develop. Other rare complications include Post-Traumatic Stress
Disorder and, very rarely, Mania in persons with the history of affective disorder(3).
In paper I (11) we have tried to show the intensity of the morbid grief reaction, the degree of distress,
depression and anxiety among close relatives of ‘red-terror’ victims. Theoretically, it is
______________________________________
1
From the Department of psychiatry, Addis Ababa University, P.O.Box 9086, Addis Ababa, Ethiopia
and 2Amanuel Hospital, P.O.Box 1971, Addis Ababa, Ethiopia.
expected that grief will diminish as a function of time from the death and this was proved in a number
Morbid grief II 257
────────────────────────────────────────────────────────────
of studies (4,5), but in those bereaved due to ‘red-terror’, the scores remained relatively high.
This paper deals with the analysis of the order of positive endorsement of each item of ETIG and it's
mean score which will definitely cast more light of the phenomenology of morbid grief characteristic
to our respondents. It also assesses the outcome of the grief reaction and comments upon those items
that indicate ‘good’ outcome. This study will also clarify the relationship of the morbid grief scores, as
measured by the BTIG, to other existing psychological constructs such as GHQ-30, BDI and SAI
scores.
Methods
For the selection of samples, the test materials used and the statistical analyses, please refer back to
the methods of paper I (11).
Results
The demographic variables were given in paper I (11) (Table 1) and discussed appropriately. Table
1 shows the 34-items of the ETIG and the mean score of each item in a descending order. It also shows
the percentage of positive endorsement of each item by the 91 respondents who lost a close relative or
relatives 18 years ago in the ‘red-terror’.
The mean score has ranged from 1.71 for ‘Guilt feeling’ to 4.81 for ‘feeling it is unfair’ and the
percentage of positive endorsement has ranged from 85.7% for ‘acquiring the habits and interests of
the deceased’ (i.e. identification) to 100% for ‘feeling it is unfair’ and for ‘lack of substitution’. The
lowest positive endorsement of 85.7% mentioned above is very highly significant (P<0.001).
When the mean scores of items in Table 1 were compared with those reported by Lundin (8) 8 years
after the loss and by Zisook et al (5) 4.5 years (average) after the loss, a marked tendency of higher
scoring was observed in all except in 4 items (i.e. F4, F14, F19 and F22)
Table 1: Response of 91 close relatives of the ‘red-terror’ victims to the Expanded Texas Inventory of Grief (ETIG) in the descending order of mean scores (18 years
after bereavement), Addis Ababa, 1996.
No.
Item
Mean Score**
(+ve)
endorsement
%
1 (F6)
I feel it is unfair
4.81
100
2 (F1)
I cry inside for him/her
4.69
98.9
3 (F2)
I still get angry when I think of him/her
4.66
98.9
4*(F5)
I am preoccupied with thoughts of him/her
4.64
97.8
5 (F11)
No one will ever take his/her place in my life
4.56
100
6 (F24)
Things and people around me still remind me of him/her
4.41
98.9
7 (F25)
I very much miss the person
4.40
95.6
8 (F16)
I cry when I think about him/her
4.37
96.7
9*(F23)
I get upset when I think about him/her
4.37
97.8
10 (F26)
It is painfull to recall memories of him/her
4.35
95.6
11 (F32)
I can't avoid thinking about him/her
4.29
96.7
12*(F30)
I still feel the need to cry for him/her
4.23
97.8
13 (F17)
A numbness comes over me when I think of him/her
4.22
96.7
14 (F15)
Sometimes I dream about him/her
4.20
94.5
15 (F29)
My health has not been good since he/she died
4.15
97.8
16 (F13)
I have to laugh when I think about him/her
4.13
92.3
17 (F18)
I feel physically ill when I think of him/her
4.04
96.7
18*(F21)
I can not accept his/her death
4.01
91.2
19*(F31)
I get upset each year about the time that he/she died
4.01
96.7
20 (F20)
I have never known a better person
3.81
95.6
21 (F12)
I hide my tears when I think about him/her at time
3.79
91.2
22 (F7)
I feel he/she is still with me
3.65
96.7
23 (F10)
I would feel better if I could really cry
3.60
96.7
24 (F19)
I feel I have adjusted well to the loss
3.55
93.4
25 (F27)
I try to avoid thinking of him/her
3.29
93.4
26*(F28)
I feel just like the person who died
3.18
92.3
27 (F22)
I am now functioning as well as before
3.03
91.2
28 (F3)
Since he/she died, I am more like him/her
2.89
95.6
29 (F8)
I have acquired the habits and interests of him/her
2.74
85.7***
30 (F33)
I feel I have the same illness as him/her
2.62
92.3
31*(F34)
I have pain in the same area of my body as him/her
2.44
93.4
32 (F14)
Now I can talk about the person without discomfort
1.81
95.6
33 (F9)
I have found someone to talk his/her place
1.74
91.2
34 (F4)
I feel guilty when I think of him/her
1.71
90.1
258 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
* The seven items of BTIG
** 1= completely false, 2= mostly false, 3= Partly true and partly false, 4= mostly true, and 5= completely true. *** P<0.001
and one additional item, (F9), in case of Zisook's report. The significance of comparatively low
scoring in F4, F14, F19 and F22 will be discussed later together with other observations shown in
Table 2.
As shown in Table 3, there is a weak positive correlation between BTIG and GHQ-30 (r=+0.294,
P<0.01) and similar weak positive correlation between BTIG and BDI (r= +0.247, P<0.02). Between
BTIG and SAI the correlation is negative (r= -0.23, P<0.05) but similarly weak.
GHQ-30 has shown a moderate correlation with BDI (r= +0.631, P<0.001) and a strong correlation
with SAI (r= +0.736, P<0.001). BDI and SAI also has shown a moderate correlation (r= +0.647,
P<0.001) between themselves.
Table 2: Sub-sets of items for a good outcome, Addis Ababa, 1996.
Sub-set A
(Low scoring items)
F4
Guilt feelings
F 16
Sad when thinking about the lost person
*F21
Not accepting the loss
*F23
Feeling of being upset
F26
painful to recall memories
*F31
Upset at the anniversary of the loss
Sub-set B
(High scoring items)
F 14
Ability to talk about the lost person without discomfort
F 19
Feeling of good adjustment
F 22
Better functioning
* Three items of BTIG
Mean score
1.71
4.37
4.01
4.37
4.35
4.01
1.81
3.55
3.03
Discussion
One hundred percent positive endorsement and the highest mean score of 4.81 for ‘feeling it is
unfair’ deserves special comment. The inventories were completed between February and May 1995
which coincided with the first and the second court trials of the political officials who are held
responsible for the ‘red-terror’. There were also some coverages of the trials in the mass-media and
mass demonstrations were held in Addis Ababa asking the Government for a fair and quick trial. In
such circumstances 100% positive endorsement and the highest score for the ‘feeling it is unfair' is
expected. Probably respondents have simply expressed their accurate feelings.
‘Guilt feeling’, item F4, which has 90.1% positive endorsement and the lowest mean score, 1.71,
also deserves special comment. ‘Guilt feeling’, though a common experience of survivors in normal
circumstances, is irrational and yields itself to reality testing. Some survivors handle their own sense of
culpability by projecting their guilt on to others and blaming them for the death. Finding someone to
blame can be an attempt to affirm control and find a sense of meaning in ‘a difficult-to-understand
situation’ (1). Bowlby has mentioned the mental processes (i.e. defense) that mitigate the painfulness
of mourning and in particular he has shown that projection is a frequent and inevitable accompaniment
of object loss(6,7). In the ‘red-terror’ there are real persons or political organizations held responsible
for the act of killing and who are on trail at this moment. In such circumstances, though positive
endorsement of ‘guilt feeling’ on the inventory is expected, it is unlikely to get higher scores and our
finding could be a proof to that.
Another item, (F11), which shows ‘lack of substitution for the deceased’ was positively endorsed by
100% of respondents with 4.56 mean score. This indicates that they all still miss the deceased very
much and feel ‘no one will ever take his/her place’. Other two items, F20 and F25, indicating ‘lack of
substitution for the deceased’ or ‘feeling of missing of the deceased’ were both positively endorsed by
95.6% of the respondents with mean scores of 4.40 and 3.81 respectively. ‘Finding a substitution’, item
F9, though endorsed by 91.2%, has one of the lowest mean scores, 1.74, which is 31st in rank. The
overlap in the percentages of endorsement between items indicating ‘lack of substitution’ (F11, F20,
F25) on one hand and item (F9), an item indicating ‘finding a substitution, on the other hand
Morbid grief II 259
────────────────────────────────────────────────────────────
demonstrates a continuous effort of the bereaved to reorganize themselves- i.e. to accommodate to the
loss by facilitating their ability to live without the deceased and by facilitating emotional relocation of
the deceased to a new place in their life which allows the bereaved to move forward with life and form
new relationships. The discrepancy in scores between the two groups of items could indicate that our
respondents are rather disorganized due to morbid grief.
The ‘degree of mourning’ as manifested by ‘cry inside’ (F1), was endorsed positively by 98.9% of
the respondents with 4.69 mean score which is the second highest score. Other items related to crying
or tearfulness, (F10, F12, F16 and F30), were endorsed positively by respondents ranging from 91.2%
to 97.8% with mean scores ranging from 3.60 to 4.37.
‘Remembering the deceased by things and people around’ (F24) was endorsed by 98.9% with the mean
score of 4.41.
Items expressing ‘identification with the deceased’ (F3, F8, F28, F33 and F34) have relatively lower
mean scores ranging from 2.44 to 3.18 and their percentage of positive endorsement (85.7% to 95.6%)
is also relatively on the lower side except F3 which has 95.6% endorsement.
Memories or thoughts about the deceased (F18, F26, F23 and F2) as ‘causes of distress’ (pain and
anger/upset) was endorsed by respondents ranging from 95.6% to 98.6% and the mean scores too are
quite high, 4.04, 4.35, 4.37, and 4.66, respectively.
Items still indicating ‘perceptual set for the lost individual’ (F7, F15, F32 and F5) were endorsed by
respondents ranging from 94.5% to 97.8% and the mean scores range from 3.65 to 4.64. ‘Anniversary
mourning’ (F31) was endorsed by 96.7% with a mean score of 4.01. These two syndromes also are
quite prominent.
‘Not accepting the loss’ (F21), which has the same mean score of (4.01) as (F31), is known to be
manifested with a greater intensity during the ‘anniversary mourning’. Other specific grief reactions
indicating ‘not acceptance of the reality of the loss’ include:- a) yearning and pining for the lost
individual, b) need to talk about the lost individual, c) recurring of memories, which are usually
idealized, d) distress at reminders of the loss which could be intense, e) sadness and f) nostalgia (3).
‘Better functioning’ (F22), which has an equal percentage of endorsement (91.2%) as ‘not accepting
the loss’ (F21), has a lower mean score (3.03) than (F21) which is 4.01. Mean scores of items ‘better
functioning’ (F22), and ‘feelings of good adjustment’ (F19) must be taken cautiously when interpreting
the outcome of the grief reaction of our respondents.
Table 3: Pearson product-moment correlation coefficient (r) between different inventories and their level of significance(p),
Addis Ababa, 1996.
GHQ-30
BDI
SAI
r
p
r
p
r
p
BTIG
+0.294
<0.01
+0.247
<0.02
-0.23
-0.23
GHQ-30
--+0.631
0.001
+0.736
+0.736
BDI
----0.647
0.647
Examining the sub-sets of items for a ‘good outcome’ of grief reaction (see Table 2), the
respondents have scored high in sub-set A (i.e. low scoring items) except in F4 (i.e. guilt feeling)
which was already discussed in detail. High scores in sub-set A indicate a ‘bad outcome’. The mean
scores in sub-set B (i.e. high scoring items) are relatively lower indicating again a ‘bad outcome’. The
score for F14 (i.e. 1.81) is indisputably low, but one might argue that scores for F19, i.e. ‘feeling of
adjustment’ and for F22, i.e. ‘better functioning’ are not low enough in themselves. Our answer will be
that they are not comparatively high enough. In fact, when compared to Lundin's (8) and Zisook's (5)
studies, our respondents have lower mean score in all items of the sub-set B indicating comparatively
‘bad outcome’. The authors believe that in chronic grief reaction where the outcome is expected to be
‘poor’, items F19 and F22 may score moderately high as the bereaved have lived and functioned in the
same circumstances for a long time and have adapted or adjusted to it.
The magnitude of associations(r) between BTIG on one hand and GHQ-30, BDI, and SAI on the
other hand, though weak, are significant and these indicate that a change in BTIG score is rarely
260 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
associated with changes of similar degree in the other scales. In other words, higher degree of
mourning is rarely associated with higher degrees of distress, as measured by GHQ-30, or depression
or anxiety. Risk factors or susceptibility for developing complications like depression and anxiety
depends on (9): a/- age and gender of the bereaved, b/- the nature of attachment to the deceased and c/personality trait of the bereaved (i.e. constitutional factors). During acute bereavement there is
coincidence of pathologic grief, major depression and anxiety disorder, but the longitudinal data
suggests a relationship between pathologic grief and major depression only. Anxiety feelings in
chronic grief may be intense, but are not as frequent as depression and do not reach diagnostic criteria
for the disorder (3,10).
The week negative correlation between BTIG and SAI could be explained by the fact that SAI is a
sensitive indicator of change in the level of anxiety and is concerned with how respondents feel during
filling in of the forms. Possibly, some respondents who did not show sufficient degrees of grief on the
inventory might have been worried by their insufficient responses and thus leading to a relatively
higher scores on SAI.
Moderate to strong correlations between GHQ-30 and BDI, GHQ-30 and SAI, and BDI and SAI
which are very highly significant (or significant at 0.1% level) indicate that a change in one of these
variables is associated with similar, though not equal changes in others. As these variables are
complications of grief reaction as mentioned earlier, these correlations also indicate the common
linkage they have to the pathologic grief reaction.
In conclusion, this study shows that our respondents experience several signs and symptoms of
bereavement even 18 years after the loss of their loved ones in the ‘red-terror’. Percentage of positive
endorsement and in particular the mean score of each item of the ETIG, was quite high when compared
to other studies except in those items whose high scoring indicate ‘good outcome’. In other words the
results have indicated higher degrees of morbid grief with ‘bad outcome’. The rank of the
endorsement and the mean score of each item appear to be specific to the nature of the grief or to the
circumstances of the loss.
The syndromes that belong to the complications of grief are vivid and circumscribed. The ‘feeling
that it is unfair’ and ‘guilty feeling’ with their highest and lowest mean scores, items indicating
‘feelings of missing’ with a wide range in their mean scores and similarly items indicating ‘tearfulness’
with a wide range in their mean scores are endorsed by significantly high percentage of respondents.
‘Remembering the deceased’ was quite high in endorsement and mean score. The items manifesting
‘identification with the deceased’ have scored relatively low and the percentages of endorsement also
are generally low. The distress caused by memories or thoughts about the deceased is quite
remarkable and the percentage of endorsement of items and their mean scores too are relatively on the
higher side.
Items indicating the ‘presence of perceptual set for the lost individual ‘and the ‘anniversary
mourning’ are as prominent as items indicating ‘not accepting the reality of loss’. When examining the
sub-sets of items indicating ‘good outcome’, there is clear evidence that the outcome is ‘bad’.
Interpretation of all these items has to take into account the chronic nature of the grief and the
circumstances of the loss. Such an approach will help in the proper assessment and counselling those
who lost relatives in the ‘red-terror’.
This study also shows that a higher degree of bereavement does not necessarily mean, or rarely
mean, a higher degree of distress, depression or anxiety, but there is a strong evidence that the distress
as measured by GHQ-30, the depression and the anxiety all have a strong common linkage to the grief
reaction.
Acknowledgment
We gratefully acknowledge the Ethiopian Science and Technology Commission for its financial
assistance. We are also thankful to the Addis Ababa Branch of Human Rights League for the Victims
of "Red-Terror" for its help in the distribution and collection of the questionnaires. We also thank Wo.
Almaz Lemma for typing the manuscript.
References
1. Wardon JW. Abnormal Grief Reactions: Complicated Mourning. In Wardon JW, editor. Grief
Counselling and Grief Therapy: A Handbook for the Mental Health Practitioner. NY: Springer
Publication Company, 1991;65-78.
2. Horowitz MJ, Wilner N, Marmar C, Krupnick J. Pathological Grief and the activation of latent self
images. Amer J Psychiatry. 1980;137:1157-62.
3. Jacobs SC. Diagnostic criteria for Pathologic Grief. In: Jacobs SC, editor. Pathologic Grief:
maladaptation to loss. NY American Psychiatry Press, 1993;368-9.
4. Faschingbauer TR, De Vaul RD, Zisook S. Development of the Texas Inventory of Grief. Am J
Psychiat. 1977;134(6):696-8.
5. Zisook S, De Vaul RD, Click MA. Measuring Symptoms of Grief and Bereavement. Am J Psychiat.
1984; 139(12): 1590-3.
6. Bowlby J. Attachement and loss, Volume III. Loss: Sadness and Depression. Bucks (U.K). Hazell
Watson and Viney Ltd, 1980;44-74 and 137-141.
7. Bowlby J. Grief and mourning in infancy and early childhood. In Maddison D, and Walker WL.
Factors Affecting the Outcome of Conjugal Bereavement. Brit J Psychiat. 1967;113:1057-1067.
8. Lundin T. Long-term outcome of bereavement. Brit J Psychiat. 1984;145:424-8.
9. Jacobs SC. Personal Risk Factors of Complication. In: Jacobs SC, editor. Pathologic Grief:
Maladaptation to loss. NY. American Psychiatry Press, 1993;141-167.
10. Jacobs SC. Relationship Among Clinical Complication: In: Jacobs SC, editor. Pathologic Grief:
Maladaptation to loss NY. American Psychiatry Press, 1993;59-72.
11. Abdullahi AB, Hyder MA. Morbid grief I: Are close relatives of the “red-terror” victims of Addis
Ababa still suffering from a morbid grief and other complications of bereavement? Ethiop J Health
Dev. 1997;11(3):241-249.
Original article
Morbid grief III: The influence of variables on the
degree of grief reaction, depression and anxiety
among close relatives of the "red-terror" victims.
Abdulreshid Abdullahi Bekry1, Mohammed Haji Hyder Ali2
Abstract: To analyze some of the variables of the bereaved and of the victims that could significantly
influence the degree of grief reaction, depression and anxiety, 91 randomly selected close relatives of
victims of the ‘red-terror’ have completed the ETIG (Expanded Texas Inventory of Grief), BDI (Beck
Depression Inventory), and SAI (State Anxiety Inventory) 18 years after the loss. All these
questionnaires are self-rating, Amharic-translated and with acceptable face validity, but they are not yet
concurrently validated with their corresponding English versions. The results have shown that the
older age group (60-79 yrs) had a mean score of (131.63) on ETIG which is significantly higher
(P<0.05) than that (118.56) of the youngest age group (20-39 yrs). The widows/ers had the highest
mean score (139.95) on ETIG which is significantly higher (P<0.01 and P<0.001) than those of
married (124.83) and singles (117.61), respectively. They were also found to have a mean score of
56.62 on SAI which is significantly higher (P<0.05) than that of singles (47.70) only. Parents who lost
son(s) had a mean score of 132.26 on ETIG which is significantly higher (P<0.01) than those who lost
brother(s) (118.93). Among the variables of the victims, where the dead body was given to relatives
for funeral services, the mean score on the ETIG was 109.2 and this was found to be significantly
lower (P<0.02 and P<0.001) than where the dead body was left on the street and not given to relatives
(127.89) or where the dead body was neither left on the street nor given to relatives (129), respectively.
It was recommended that the older age group, widows/ers, parents who lost son(s) and those who
could not confirm the death of the victim by seeing the dead body, should get priority for counselling.
[Ethiop. J. Health Dev. 1997;11(3):257-261]
Introduction
In Paper I (first phase of this study which was reported in the same issue of EJHD) (4) it was shown
that close relatives of the ‘red-terror’ victims of the Ethiopian revolution in the late 1970's still suffer
from a morbid grief 18 years after the loss. Paper II (5) has dealt with the phenomenology of morbid
grief and it's association with the general distress, depression and anxiety and it has indicated that the
outcome of their grief reaction was ‘bad’. Variables that could affect the course of mourning were not
examined so far. Five categories of variables were identified by Bowlby (1):
1. The identity and role of the person lost,
2. The causes and circumstances of the loss,
3. The age and sex of the person bereaved,
4. The social and psychological circumstances affecting the bereaved about the time of and after the
loss, and
5. The personality of the bereaved with special reference to his/her capacities for making love
relationships and for responding to stressful situations.
______________________________________
1
From the Department of psychiatry, Addis Ababa University, P.O.Box 9086, Addis Ababa, Ethiopia
and 2Amanuel Hospital, P.O.Box 1971, Addis Ababa, Ethiopia.
The disordered mourning follows the loss of immediate family members with whom there has been,
until the loss, a close relationship, in which lives are deeply interwind (1). It was reported that loss of a
Morbid grief III 263
────────────────────────────────────────────────────────────
child is usually followed by severely disordered mourning whereas loss of a sibling during adult life is
not frequently followed by disordered mourning, but no adequate systematic data are available by
which that supposition can be checked (1). There are also difficulties in determining the differential
incidence of disordered mourning by the age and also by the sex of the bereaved as there are
difficulties in determining the differential incidence of disordered mourning following losses of
different kind (1).
The causes of loss and the circumstances in which it occurred in the ‘red-terror’ were more or less
similar, but social and psychological circumstances affecting the bereaved were quite different and the
influence of some of the relevant variables on the course or the degree of mourning have to be
understood so that effective help could be provided to the bereaved.
This paper deals with the analyses of some of the variables of the bereaved and of the victims that
could significantly influence the degree of the grief reaction, the depression and the anxiety of the
close relatives of victims of the ‘red-terror’.
Methods
For the selection of samples, the test materials used and the statistical analyses, please refer back to
the methods of paper I (4).
Results
Table 1 shows some of the variables of the bereaved thought to be relevant for the analyses. Among
the three arbitrary age groups, the mean ETIG score was found to be increasing with age. However, the
difference is significant (P<0.05) only between group A (age range 20-39) and group C (age range 6079). The difference between groups A and B or between B and C is not significant. With regard to
depression and anxiety there are no significant differences between any of the age groups.
The difference between male and female mean scores on ETIG, BDI, and SAI are not significant.
Similarly, no significant differences between the two religion groups on the above mentioned
inventories.
With regard to marital status, the mean score on the ETIG of the widows/ers (139.95) is the highest
which is significantly higher than the mean scores of married groups (124.83) and singles (117.61).
But no significant difference was found between the married group and the singles. The mean score on
SAI of the widows/ers (56.52) is also the highest, but this mean score is significantly higher (P<0.05)
than that of singles (47.70) only. No significant difference on BDI was found between the different
marital groups.
Examining the employment status, those who are pensioned or dependent on others have higher
mean scores than the employed group on all the inventories, but none of these mean scores is
significantly higher than the corresponding mean score.
Now, coming to traditional mourning, there is no significant difference on ETIG, BDI and SAI
mean scores between the group permitted to carry on traditional mourning and the group prevented.
Those who lost son(s) have higher mean scores on ETIG, BDI, and SAI (132.26, 26.37, and 53.49,
respectively) than those who lost brother(s) (118.93, 20.9, and 47.8, respectively), but the difference is
significant (P<0.01) only on ETIG.
Table 2 shows the different variables of the deceased which are considered to be appropriate for the
analyses. Examining the different age groups for significant differences on ETIG, BDI, and SAI, mean
scores have shown no such significant differences. The same thing is true when the ‘circumstances
before bereavement’ were examined : imprisonment, torturing the victims before actual killing or
dumping the mutilated body on the street have not shown significant differences on the mean scores of
ETIG, BDI or SAI.
Other variables of the deceased, such as leaving children behind, neighbourhood participation in the
traditional mourning, experience of mourning before the ‘red-terror’, and employment status of the
deceased have not manifested significant differences on the mean scores of ETIG, BDI, and SAI.
Table 1: The different variables of the bereaved and their mean scores, SDs and levels of significance on ETIG, BDI, and SAI, Addis Ababa, 1996.
264 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Variables
No*
mean
ETIG
SD
Significance
P< **
Mean
BDI
SD
A(20-39)
36
118.56
22.38
P<0.05
20.58
11.38
B(40-59)
22
127.18
2.61
(A&C)
24.14
13.45
C(60-79)
32
131.63
23.65
25.47
14.47
A male
49
122.53
22.92
25.35
13.7
B female
42
130.26
18.95
22.19
13.62
A Christ.
80
126.3
20.95
B Muslim
11
124.64
25.71
A Single
23
117.61
20.91
B Married
40
124.83
Marital Status:
A Single
23
B Married
40
Significance
P<**
Mean
SAI
SD
48.81
14.1
52.68
13.84
53.94
15.15
49.82
14.64
54.29
14.36
3.35
4.69
Significance
P<**
Present Age (yrs):
NS
Sex:
NS
NS
Religion:
NS
NS
23.8
13.0
18.82
11.68
4.82
4.59
NS(A&B)
19.35
12.43
7.70
3.66
23.43
P<0.01
(B&C)
24.28
12.25
0.23
4.06
117.61
20.91
NS(A&B)
19.35
12.43
124.83
23.43
P<0.01
(B&C)
24.28
12.25
139.95
10.04
P<0.001
(A&C)
21.90
12.33
27
121.04
22.74
NS
21.67
14.27
dependent
60
128.75
20.74
24.90
12.25
Traditional
mourning:
Permitted
33
25.03
21.25
21.52
13.21
Prevented
44
126.39
23.86
23.14
12.87
43
32.26
17.55
26.37
14.18
29
118.93
21.10
20.90
12.13
Marital Status:
C Widow/er
21
Employment
Status:
A Self/
others
NS
NS
56.52
50.93
47.70
NS(A&B)
50.23
NS(B&C)
14.50
P<0.05
(A&C)
12.56
NS
B Pension/
Relationship
victims
bereaved:
Son(s)
Brother(s)
NS
NS
2.90
15.68
53.36
11.31
48.32
15.75
53.49
5.33
47.80
12.88
NS
of
to
P<0.01
NS
* Number of respondents was 91. Only major groups are included for analyses. Few have not responded to certain questions.
** Appropriate tests, Z or T, were used depending on the number in the groups
NS= Not significant
SD= Standard Deviation
NS
Morbid grief III 265
────────────────────────────────────────────────────────────
The only variable that has manifested significant differences on the mean scores of grief inventory
(ETIG) is the ‘Circumstances soon after death’. Those who were given the dead bodies to arrange their
own burial services (i.e. group C) have significantly less mean score on ETIG than both group A
(P<0.02) and group B (P<0.01) but no such significant difference was observed between groups A and
B.
Discussion
In papers I and II (4, 5), it was shown that 91 respondents who lost close relatives in the ‘red-terror’
18 years ago are still suffering from a severe form of morbid grief, depression and anxiety. The effects
of some of the variables on the mean scores of grief, depression and anxiety are shown in Table 1 and
2.
Results on Table 1 indicate that the older age group (60-79 yrs) are more severely (P<0.05)
affected by the chronic grief reaction than the youngest age group (20-39 yrs). It is also true that they
suffer more from depression and anxiety than the youngest age group, though not significantly so.
Widows/ers suffer significantly more from chronic grief reaction than married (P<0.01) and singles
(P<0.001). Though they are also more anxious, they are significantly so (P<0.05) when compared to
singles only.
Those who lost son(s) in the ‘red-terror’ suffer more than those who lost brother(s) from chronic
grief reaction, depression and anxiety, but significantly (P<0.01) so in grief reaction only. Earlier in the
introduction, it was stated that there are no adequate systematic data to check the supposition that loss
of a child is followed by a severely disordered mourning than the loss of sibling. But the above
finding (see also Table 1),
266 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 2: The different variables of the victims and their mean scores, SDs, and the level of significance on ETIG, BDI and SAI, Addis Ababa, 1996.
Variables
No*
mean
ETIG
Significance
Mean
BDI SD
Significance
Mean
SAI SD
Significance.
SD
P< **
P<**
P<**
AGE (yrs)(at time
of deaths):
A(10-19)
29
126
22.14
23.21
12.72
50.83
16.81
B(20-29)
54
128.07
20.44
NS
23.43
14.02
51.63
14.47
NS
C(30-39)
16
130.06
22.54
23.25
10.42
52.25
14.06
D(40+)
11
129.18
24.94
22.00
11.19
54.82
14.64
Circumstances
before deaths:
A.Imprisoned
17
118.71
18.57
22.35
10.68
53.882
11.40
B.Imprisoned
26
127.19
22.31
NS
23.12
14.57
NS
50.70
17.16
NS
+Tortured
C.Imprisoned
34
126.71
21.87
21.97
13.41
49.41
15.04
+Tortured
+dead body
on the street
Circumstances
after death:
A. Body left on
36
127.89
20.7
NS(A&B)
23.08
13.86
50.19
14.79
street not given
B. Body not left on
38
129
21.87
P<0.01 (B&C)
22.92
14.14
NS
52.76
16.08
NS
street, not given
C. Body given to
9
109.2
18.52
P<0.02 (A&C)
27.56
9.50
56.44
19.51
relatives
Children left:
A. Yes
15
127.8
21.91
NS
25.3
14.42
NS
53.0
15.69
NS
B. No
76
125.76
20.67
21.40
11.12
51.07
13.75
Neighbourhood
participation in
mourning:
A. Yes
46
129.2
21.91
NS
25.3
14.42
NS
53.0
15.69
NS
B. No
45
122.91
20.67
21.40
11.12
51.07
13.75
Mourning before
A. Yes
51
123.9
22.34
NS
24.49
13.79
NS
52.98
15.55
NS
B. No
38
128.3
20.45
21.1
11.90
50.39
13.86
Employment
state:
A.employed
39
126.46
21.69
NS
23.26
12.52
NS
51.1
15.27
NS
B. Student
51
126.02
23.45
23.16
13.04
52.38
14.24
_________________________________________________________________________________________________________________________________
* The number of victims were over 91 as there were some multiple losses. Only major groups were included for analyses. Few have not responded to
certain questions.
** Appropriate tests, Z or t, were applied depending on the number in the group.
NS= not significant
SD= Standard Deviation
whereby 43 respondents who lost son(s) are compared to 29 respondents who lost brother(s), is
adequate enough to support that supposition. These above findings imply that the older age group, the
widows/ers and parents who lost son(s) suffer more and therefore deserve priority for counselling.
Here, it has to be remembered that there is a possibility of an overlap of these three variables, i.e. old
age, widowhood and being a parent who lost a son(s) 18 years ago in the ‘red-terror’. Lundin (2), in his
assessment of bereavement outcome 8 years after the loss, found that parents as a group had a more
pronounced grief reaction than widows and widowers which is contrary to our finding though the same
instrument (i.e. ETIG) is used in both studies. He compared the mean scores of each item whereas we
compared the mean scores of all items. The most important difference is that our respondents have
recorded their present marital status and not their marital status at the time of the loss. In fact there
were only five respondents in our study who lost their spouses in the ‘red-terror’. The rest (i.e. 16) lost
their spouses after the ‘red-terror’ was over. Therefore, these could be considered as having
‘bereavement overload’.
Morbid grief III 267
────────────────────────────────────────────────────────────
Table 2 has shown one very important variable which has a considerable effect on the degree of
mourning. That is ‘giving the dead body for funeral services/ rituals’. The mean score of this group
(i.e. group C) on ETIG is 109.2 which is comparatively the lowest recorded mean score in both Tables
1 and 2. It is known that the funeral rituals help in aiding the healthy resolution of grief as follows (3):
1. Seeing the body of the deceased loved one helps to bring home the reality and finality of death
which is the first task of mourning.
2. It helps the grief process as it allows to express thoughts and feelings about the deceased.
3. The ritual can also be a reflection of the life of the person who is gone, and
4. It helps to draw a social support network close to the bereaved family shortly after the loss has
occurred and this can facilitate the grief process.
Here, it has to be remembered that ‘giving the dead body for the funeral rituals’ is not the same as
‘permitting traditional mourning’ which can be performed without seeing the dead body. It was already
stated (see result) that there is no significant difference whether traditional mourning is permitted or
prevented. It is concluded from the findings of this study that old age group, widows/ers, and parents
who lost sons suffer more from pathologic grief reaction. It is also concluded that carrying out funeral
services is associated with comparatively lowest degree of pathological grief reaction. We have
already recommended setting up a bereavement counselling services for those who suffer from a
pathologic grief and we believe that the findings of this paper will help in giving priorities for
treatment to certain groups and in monitoring the outcome.
Acknowledgement
We gratefully acknowledge the Ethiopian Science and Technology Commission for its financial
assistance. We are also thankful to the Addis Ababa branch of Human Rights League for the Victims
of ‘Red-Terror’ for its help in the distribution and collection of questionnaires.
Reference
1. Bowlby J. attachment and loss, volume III. Loss: Sadness and depression. Bucks (UK). Harzell
Watson and Vinety Ltd, 1980,172-201.
2. Lundin T. Long-term outcome of bereavement. Brit J Psychiat. 1984;145:424-8.
3. Wardon JW. Grief counselling: Facilitator uncomplicated grief. In Wardon JW, editor. Grief
counselling and grief therapy: A handbook for the mental health practitioner. NY springer Publishing
Company, 1981;65-78.
4. Abdullahi AB, Hyder MA. Morbid grief I: Are close relatives of the “red-terror” victims of Addis
Ababa still suffering from a morbid grief and other complications of bereavement?. Etiop J Health
Dev. 1997;11(3):241-249.
5. Abdullahi AB, Hyder MA. Morbid grief II: The phenomenology of pathologic grief process,
depression and anxiety among close relatives of “red-terror” victims. Ethiop J Health Dev.
1997;11(3):251-256.
268 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Original article
Patterns of prescription in Jimma Hospital
Mohammed Abdulahi1, Tesfaye Shiferaw1
Abstract: A total of 2170 prescriptions were recorded over a period of one year using structured
questionnaire to systematically study drug-prescription patterns in Jimma Hospital. The prescription
patterns demonstrated 94.8% for essential drugs and 75.2% for generic prescribing. The average
number of drugs per prescription was 1.59 and the cost per prescription indicated 3.28 Birr. The
proportion of prescriptions with antimicrobial combination accounted for 2.72%. The most frequently
prescribed agents were antimicrobials 33.1%. The proportion of prescriptions with injection was
20.2%. Of the major diseases, Respiratory Tract Infections accounted for 13.6% of the prescriptions.
The findings indicated that there were good signs of rational drug prescribing as noted by low average
number of drugs per prescription and high proportion of generic prescribing that matched with
prevalent disease patterns, although, there seemed to be problems with antimicrobial combinations and
high frequency of injections. [Ethiop. J. Health Dev. 1997;11(3):263-267]
Introduction
Drugs are one of the major components of the health care system and play important role in saving
lives when rationally used. There is no enough information on the patterns of drug use in different
parts of the world including Ethiopia.
The limited data on prescription patterns and drug use indicate that, drug utilization in both
developed and developing countries is generally not rational (6). The major problems on prescription
practices are excessive prescribing, inadequate prescribing and incorrect prescribing. There is
considerable evidence that antibiotic prescribing is excessive and inappropriate even in developed
countries. In developing countries, there is inappropriate use of antibiotics and other drugs.
Antibiotics account for the largest single group of drugs purchased though their consumption varies
widely among countries. The volume of use of other drugs also far exceeds the prevalence of the
diseases supposed to be treated with (1). The consequences of this include, subjecting patients to
unnecessary adverse effects, reproduction of drug-resistant microorganisms
It was also shown (2) that there is over use and misuse of ineffective or obsolete products, creating
unnecessary risk to the patients and additional cost for both individuals and the health care system.
The reasons for such practices are multiple and include inadequate training, inadequate information on
the drugs, the promotional activities of drug companies, pressure form patients and a false perception
of “a pill for every illness” (4).
The greatest irrationality in drug use is using when they are not needed. Many clinical symptoms
are caused by self-limiting sicknesses may not require treatment. However, many patients are given
prescription from which they don’t benefit. Even when drug is required, some patients expect more
than one drug on each prescription. In countries where drug shortage occurs or distribution is uneven,
this may mean that while one patient gets too many drugs others in need are deprived of it (4).
The rational use of drugs depends on the knowledge and attitude of the public. Raising public
awareness by educating them about the basic concepts that users of medicines need to understand
______________________________________
1
From the Jimma Institute of Health Sciences P.O. Box 378 Jimma, Ethiopia.
and waste of limited resources particularly in developing countries.
Patterns of prescription in Jimma Hospital 269
────────────────────────────────────────────────────────────
may minimize self medication. The use of generic drugs, when encouraged among the population is
also a means of rationalising drug use and decreasing family expenditure.
Today there is an increased trend in drug consumption all over the world, but this does not mean
that people are in better health. The increased consumptions could be due to patterns of prescribing
and the attraction that drugs exert (5). One of the objectives of the national drug policy of Ethiopia is
to limit the proliferation of unnecessary products by using the developed national drug list (6).
The objective of this study was to determine the patterns of prescription in Jimma Hospital through
inventory of records over a period of one year.
Methods
Drug prescription patterns were assessed in Jimma Hospital during the period December 1990 to
November 1991. A total of 2170 prescriptions which included both paying and free patients were
recorded over a year with emphasis on patient identification number, cost of drugs on a prescription,
the extent of generic prescribing and diagnosis from patient’s records.
Prescriptions written by physicians for only one diagnosis at the Out Patient Department of Jimma
Hospital which includes Medical, Surgical, Paediatrics, Gyn., Ophthalmology and Dental units were
abstracted during the study period to avoid the ambiguity that may be caused by multiple diagnosis.
Except for new cases, prescriptions for anti-TB drugs were excluded to avoid the influence of repeat
patients prescriptions during the analysis. Data was collected for one week during each month of the
year and data processing and analysis were done using EPI-INFO & SPSS computer programs.
The average number of drugs per prescription, average cost of drugs per prescription, percentage of
prescriptions with antimicrobial combination, percentage of drugs prescribed by generic names,
percentage of drugs prescribed out of Essential Drug List for Ethiopia and percentage of prescriptions
with
injections were indicators used for analysis and calculated over a period of 12 weeks.
Definitions
1. Antimicrobials in this particular study include: All antibiotics and other antibacterials, antiviral,
systemic antifungals and antimalarials prescribed.
2. Essential drug List: list of those drugs that satisfy the health care needs of the largest segment of a
given population and should be available in adequate quantity and proper dosage form all the
times.
3. Polypharmacy: a trend of prescribing three or more drugs on a prescription paper.
Results
Two thousand one hundred and seventy patient records were collected with the diagnosis of which
prescriptions were written in Jimma Hospital. The records revealed that Respiratory Tract Infections
(305 or 16.3%), Helminthiasis (258 or 13.8%) followed by Skin Problems including accidental injury
and soft tissue laceration (256 or 13.7%) were the major causes of morbidity (Table 1).
270 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 1: Major diseases for which patients visited Jimma Hospital during the study period (Dec. 1990 to Nov. 1991) (Total Number of prescriptions =2,170)
S. No
Type of Disease
Frequency
Percent
1
Respiratory Tract Infections
305
16.3
2
Helminthiasis
258
13.8
3
Skin problems (including soft tissue
256
13.7
laceration and accidental wound)
4
Gastroenteritis including diarrhoea
132
7.0
5
Urinary Tract Infections
131
7.0
6
C.N.S problems
102
5.4
7
Gastritis and PUD
95
5.1
8
ENT problems
73
3.9
9
Sexually Transmitted Diseases
72
3.8
10
Acute Febrile Illnesses
63
3.4
11
Maternal problems
62
3.3
12
Rheumatoid Arthritis
60
3.2
13
Bronchial Asthma
57
3.0
14
Pelvic Inflammatory Diseases
48
2.6
15
Tuberculosis (only new cases)
31
1.7
16
Malaria
29
1.5
17
Cardiovascular Diseases
28
1.5
18
Dental problems
27
1.4
19
Anemia
24
1.3
20
Haemorrhoids
20
1.1
Total
1873
100.0
Further analysis of the patients record indicated that antimicrobials, 1141(33.1%), analgesics,
706(20.5%) vitamins and minerals, 301(8.7%) anthelimintics 287(8.3%), followed by respiratory
system drugs, 217(6.3%), were the most frequently prescribed drugs (Table 2)
Table 2: Prescribed Drug characteristics by Pharmacological classification, Jimma Hospital, December 1990-November 1991.
(Total Number of drugs = 3445)
S.No.
Pharmacological Classification
Frequency
Percent
1
Antimicrobials
1141
33.1
2
Analgesics
706
20.5
3
Vitamins nd Minerals
301
8.7
4
Anthelmintic and Antifilarials
287
8.3
5
espiratory system Drugs
217
6.3
6
Antiprotozoals
175
5.1
7
Antacids and PUD Drugs
148
4.3
8
Topical Antifungal and Other Topicals
121
3.5
9
C.N.S Drugs
112
3.3
10
Oral Rehydration Therapy
59
1.7
11
Cardiovascular System Drugs
41
1.2
12
Steroids and Hormonal preparations Including
36
1.0
contraceptives
13
Antihistamines
35
1.0
14
Drugs for Bronchial Asthma
24
0.7
15
Antihaemorroidals
21
0.6
16
Cathartic/Laxatives
16
0.5
17
Hypoglycaemic agents
5
0.15
Total
3445
100.00
A total of 3445 drugs were prescribed on 2170 prescriptions. Of these 1178(54.3%) prescriptions
contained one drug, 797(36.7) contained two drugs, 172(7.9%) contained three drugs and only
19(0.9%) of them contained four drugs (Table 3). This gave, on average, 1.59 drugs per prescription
with 3,266(94.8%) of the drugs from the Essential Drug List for Ethiopia.
Table 3: Number of Drugs per prescription, Jimma Hospital, December 1990-November 1991
S. No.
Number of Drugs
Frequency
Percent
Per-prescription
1
1
1178
54.3
2
2
797
36.7
Patterns of prescription in Jimma Hospital 271
────────────────────────────────────────────────────────────
3
4
5
3
4
5
Total
172
19
4
2,170
7.9
0.9
0.2
100.00
Analysis of the data revealed that 2590(75.2%) of the drugs were prescribed by generic names
(Table 4)
Table 4: Generic versus Brand Prescribing Jimma Hospital December 1990-November 1991
S. No
Type of prescribing
Frequency
Percent
1
Generic prescribing
2,590
75.2
2
Brand prescribing
773
22.4
3
Prescribing not specific
82
2.4
Total
3,445
100.00
A number of antimicrobial agents were prescribed for treating infectious diseases of which
ampicillin, 379(33.2%), was the leading, followed by procaine penicillin, 291(25.5%). Of the
analgesics, paracetamol 266(37.7%), was most frequently prescribed followed by aspirin, 116(16.4%),
and dipyrone, 90(12.7%). Among the anthelminthic drugs, mebendazole 133(46.3%), piperazine,
49(17.1%), and pyrantel pamoate, 35(12.2%), were highly prescribed.
From the antimicrobials prescribed, some were prescribed in combination. Ampicillin with
chloramphenicol capsules 17(28.8%) and procaine penicillin with Chloramphenicol Capsules
14(23.7%) were the leading antimicrobial combinations (Table 5).
Table 5: Antimicrobial combinations, Jimma Hospital, December 1990-November 1991
S. No.
Antimicrobial combined
Frequency
Percent
1
Ampicillin caps, and chloramphenicol caps
17
28.8
2
Procaine pen fort. and Chloramphenicol
14
23.7
3
Procaine pen. fort. and Tetracyclines
6
10.2
4
Procaine pen. fort and Ampicllin caps
4
6.8
5
Ampicillin caps and Tetracyclines caps
4
6.8
6
Ampicillin inj. and Ampicillin capsule
2
3.4
7
Others
12
20.3
Total
59
100.00
Discussion
The drug pattern indicated that infectious, malnutrition and parasitic diseases were the major health
problems of the patients visiting the hospital (Table 2) and this was in agreement with the overall heath
problems of the country.
Almost all drugs prescribed for the health problems in the hospital were in the Essential Drug List
for the country. Few drugs prescribed out of the list were those that were in the National Drug List of
Ethiopia.
Among the analgesics, the frequent prescribing of paracetamol, 266(37.7%), was a practice to be
encouraged due to the fact that it is cheap and a relatively safer analgesic at therapeutic dose.
However, considering its reported (7) adverse effects of agranulocytosis and anaphylactic shock, the
90(12.7%) prescriptions for dipyrone needs attention. The high prescribing rate of Mebendazole
133(46.3%) as anthelminthic agent was encouraging practice because it is cheap broad spectrum
anthelminthic agent which is useful for mixed worm infections.
The other encouraging practice was that Tetraycline was not prescribed for children below the age
of 7 and polypharmacy was also not frequent in that the number of prescriptions calling for more than
two drugs were only 205 or 9% (Table 3). Analysis of the drugs prescribed gave an average of 1.59
drugs per-rescription. This was encouraging compared to the values reported from similar studies
272 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
conducted in some African countries like Kenya, Cameroon and Tunisia, where the minimum number
of drugs per prescription were 3, 5 and 3 respectively (8).
The world-wide overuse and misuse of antimicrobial agents is of concern (1). There is also frequent
use of antimicrobial combinations. The present study also revealed the situation in that there was
frequent use of antimicrobial combinations all of which could not be justified. Such prescribing
practice and frequent use of injections in the hospital calls for attention and systematic review.
The present study on prescription patterns in Jimma Hospital has given the opportunity to
know about the presenting diseases and prescription patterns in the hospital.
The prescribing pattern of most of the drugs, number of drugs per-prescription and the extent of
generic prescribing demonstrated favourable indications of rational prescribing.
The initiation of implementation of the Essential Drugs List at all levels in the country and the status
of the hospital (referral teaching hospital) may have contributed to the favourable situation in Jimma
Hospital.
Though this study revealed that there were good indications of rational prescribing in Jimma
Hospital, one cannot conclude about the real situation of drug use in Jimma, because the hospital was
referral-teaching hospital and data collected includes only records of ambulatory patients with single
diagnosis. For the latter, a comprehensive practice survey that involves all patients records should be
carried out.
In Ethiopia, like in many other developing countries, infectious diseases, malnutrition and
helminthiasis are the major health problems. The findings of the present study revealed the same.
The prescription patterns also matched the disease patterns in the hospital. However, the frequent use
of antimicrobial combinations and injections has to be critically looked into and corrected unless
indicated otherwise.
To correct this prescribing practice on antimicrobial combinations and injections, provision of
refresher course on rational drug use and intensification of information, education and communication
on the National Drug Policy are recommended. To further improve the positive prescribing
behaviours, in corporation of the concept of rational drug use and cost awareness in the curricula of
health workers might be necessary.
Acknowledgement
We would like to thank Ato Fasil Tessema Community Health Department, Jimma Institute of
Health Sciences, for his assistance in the statistical analysis of the data. We are also grateful to W/ro
Senait Kassa and W/t Tsigereda Fisseha for their secretarial help.
References
1. WHO., The World Drug Situation, Geneva, 1988.
2. WHO., Action Programme on Essential Drugs. Mexican conference on Essential Drugs. Essential
Drugs Monitor, 1989;8:11.
3. Ministry of Health, Ethiopia. Comprehensive Health Service Directory, August 1988.
4. WHO Action Programme on Essential Drugs. Rational Use. A Global Priority Editorial, Essential
Drugs Monitor 1988;7:1.
5. WHO Action Programme on Essential Drugs. Rational Use, Essential Drug Monitor
1988;6:11-12.
6. Leka T, Abadir M, Pinto A. Drug Prescription Patterns in Rural and Regional Hospitals in
Ethiopia. Ethiopian Pharmaceutical Journal, 1990;8:36-45.
7. Chetley A. Problem Drugs, Dipyrone a drug no one needs Health Action International, 1993;81.
8. WHO Action Programme on Essential Drugs. More Drugs Better Health, Essential Drugs Monitor,
1988;7:14.
Original article
Reinfection of School children with Schistosoma
mansoni in the Finchaa Valley, Western Ethiopia
Berhanu Erko1, Teferi Gemetchu1, Girmay Medhin 1, Hailu Birrie 1
Abstract: Reinfection of school children with Schistosoma mansoni and factors thought to be
responsible were studied in Finchaa Valley during 1993-95. A cohort of school children treated and
cured of schistosome infection were followed for 12 months for reinfection studies. The annual
reinfection rate and intensity were 26% and 374 EPG (eggs per gram of stool), respectively. Pretreatment intensity level was regained faster than pre-treatment prevalence 12 months post-treatment.
Reinfection rate was associated with age, sex and seasonality of infection in snail hosts. Peak in snail
population density and associated cercarial infection was observed in the dry season when high
reinfection rate was found among school children. Rainfall appeared to have a pronounced effect in
governing snail population density and cercarial infection. Detailed investigations involving all age
groups on immunological response and water contact activities are necessary to assess the role of
immunity and exposure on schistosome reinfection patterns. The implication of the findings to the
control strategies are discussed. [Ethiop. J. Health Dev. 1997;11(3):269-00]
Introduction
The commonest observation in schistosomiasis reinfection studies is that reinfection rate and
intensity decline with increasing age. This observation has been an issue of inconclusive debate for
more than a decade. Two suggestions put forward to explain this phenomenon are diminishing water
contact activities and gradual acquisition of resistance to reinfection with increasing age. According to
Warren (1), there has been no definitive evidence that protective immunity plays any role in
determining the prevalence and intensity of schistosome infection. In studies made in the Volta Lake
area in Ghana Dalton and Pole (2) concluded that exposure to infected water alone could explain a
pattern of infection in a community. On the basis of the relationship of eosinophilia and resistance to
reinfection Sturrock et al (3) suggested that immune system was involved in resistance to reinfection.
Other studies that support the involvement of immune system in the
resistance to reinfection were those of Butterworth et al (4), Hagan et al (5), and Wilkins et al (6).
In the schistosomiasis control program, chemotherapy has proved efficient in the control of
morbidity in individual patients but has not proved successful in the management of transmission when
used alone (7). The effect of chemotherapy on prevalences and intensities is rapidly reversed by
reinfection (8). This has necessitated the backing up of chemotherapy with snail control measures and
this presupposes the provision of baseline data on snail ecology.
Although there have been a few studies on the transmission dynamics of S. mansoni in Ethiopia
(9,10), information on reinfection patterns of schistosomes following chemotherapy is lacking. The
present studies, therefore, attempt to investigate snail population dynamics and associated cercarial
infection in snail hosts relative to schistosome reinfection patterns in school children.
______________________________________
1
From the Institute of Pathobiology, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia.
274 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Methods
Study Area and Population: The study was carried out in the Finchaa Sugar Project Area, Finchaa
Valley, Wellega, Western Ethiopia (Fig.1). The study area is about 385 km west of Addis Ababa and
is situated at an altitude of 1280 m above sea level. Here large acreage is being developed for sugar
cane plantation using irrigation. At present there are about 10,000 people living in the project area.
The Sugar Project has a polyclinic that is staffed with medical doctors and auxiliary staff. The project
has about six camps in the plantation area and each camp has one community health agent. The study
subjects were school children attending Finchaa Valley Elementary School.
Figure 1: Skech-map of the study area
Reinfection Studies: Out of 1000 school children attending Finchaa Valley Elementary School 126
children aged 13 years and below were stool-examined in October 1993 using Kato smear to select a
cohort of school children for Schistosoma mansoni reinfection studies. The Kato template used
delivered 41.7 mg of stool plug and a single Kato smear per stool specimen was used throughout. All
Reinfection of Schistosoma mansoni in Ethiopia 275
────────────────────────────────────────────────────────────
children found positive for S. mansoni were treated with praziquantel at a single dose of 40 mg/kg
body weight to remove pre-existing worms. The treated children were re-examined two months later in
December 1993 using both Kato smear and Formol-ether concentration techniques. The use of
Formol-ether concentration technique was to rule out false negatives as much as possible so that the
cohort constitutes only cured children. The cohort was followed up for 12 months.
Snail Survey and Schistosome Infection: At the start of the study, general malacological survey was
made in all water bodies to identify and map human water contact sites. Out of the identified five sites
harboring Biomphalaria pfeifferi, only two with intense human water contact activities were selected
for snail collection sites. Nevertheless, the other three were also inspected for schistosome
transmission during snail collection period. Snails were collected with a standard scoop from
vegetation in marked stretch of the streams for 10 minutes. B. pfeifferi collected in this way were
counted and later transported to laboratories at the Finchaa Sugar Project Clinic or the Institute of
Pathobiology and checked for shedding schistosome cercariae. In the event of shedding, the cercariae
were identified to the genus level using tail morphology (11). Observation was made on water
turbidity and water velocity was also measured. Meteorological data, especially of rainfall and
temperature, were received from the Meteorology Station at the Finchaa Sugar Project.
Results
Prevalence and intensity of infection before treatment and during reinfection are presented in Table
1. Pre-treatment prevalence of schistosomiasis mansoni in the school children was 78%. Sixteen
percent of the reinfection rate occurred in the first six-months showing that higher reinfection takes
place during the
Table 1: Prevalence and intensity of infection before treatment and during reinfection among schoolchildren, in Finchaa Valley, Ethiopia, 1994.
Male
Female
Total
Pre-treatment prevalence
78(58/74) 7
77 (40/52)
78 (98/126)
Intensity (EPG)
262
317
283
6-months post-treatment prevalence
17 (10/58)
13 (5/40)
16 (16/98)
Intensity (EPG)
85
68
78
12-months post-treatment prevalence
33 (19/57)
16 (6/38)
26 (25/95)
Intensity (EPG)
478
102
374
276 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Figure 2: Seasonal fluctuation in Biomphalari pfeifferi population and schistosomal infection in relation to rainfall and
temperature in 2 streams in Finchaa Valley, Wellega, Ethiopia.
early post-treatment period. Males showed higher reinfection rate and intensity. Pre-treatment
intensity level was regained faster than pre-treatment prevalence level. Children re-acquired pretreatment egg load by 12 months post-treatment. Reinfection rate appeared to rise with increasing age
in the age groups studied. The annual reinfection rate for the age group of 6 - 9 years was 18% while it
was 31% for age group of 10 - 13 years (Table 2).
Fig. 2 shows results of snail survey and meteorological data. Snail population density showed
seasonal fluctuation with peak density from December to March when there was little or no rain. Few
or no B. pfeifferi were recovered during the big rains from May to September. There was a definite
positive association between snail population and cercarial infection in the dry season. In the dry
season the water was clear but during the rainy season it became highly turbid. The velocity of the
Reinfection of Schistosoma mansoni in Ethiopia 277
────────────────────────────────────────────────────────────
water during the dry season fell to below 10 cm/second while it went well over 25 cm/second during
the rainy season.
Interviews made with schoolchildren showed that 43 out of 57 males (75%) and 20 out of 38
females (53%) were engaged in bathing and playing in water. Such water contact activities as water
collection and laundering were equally performed by children of both sexes. It was not possible to
estimate duration of exposure for each activity for specific ages on the basis of the interview.
Discussion
Age-specific infection rates tended to increase with increasing age, i.e., from age six to 13 years.
On the contrary, in age-related schistosome reinfection studies, Wilkins et al (6) observed heavier
reinfection levels in children under ten years of age than in 10 - 14-year-olds. Explanation to the
controversy between these observations is deferred until detailed exposure and immunological studies
involving a cohort of subjects of all age groups be made in Finchaa Valley ecological setting. In such
a study, allowances have to be made for exposure to assess the role immunity plays in reinfection after
removing pre-existing worms with drug therapy.
Despite equal pre-treatment prevalence of infection in male and female children, six-months and 12months post-treatment prevalence of infections were higher among males than among females. It is
difficult to explain why pre-treatment prevalence of infection was the same for both sexes and why
post-treatment prevalence of infection was not so. Higher post-treatment prevalence of infection
among males could be explained in terms of differences in exposures to infective water between sexes.
In agreement with Kloos and Lemma (12), interview made with school children showed that more
males than females were engaged in playing in water and bathing. Although duration of exposure was
not determined for specific activity, the two activities involve long duration and maximal bodily
exposure to infective water.
Table 2: Twelve-months post-treatment prevalence of Schistosoma mansoni reinfection among school children by age and sex in Finchaa Valley,
1994.
Age
Male
Female
Total
Group
Number
Numbe
Number
Number
Number
Numbe
(Yrs)
Exam
r(%)+ ve
Exam
(%) + ve
Exam
r(%) + ve
6
2
0(0)
2
0(0)
7
2
0(0)
1
0(0)
3
0(0)
8
9
3(33)
4
0(0)
13
3(23)
9
10
2(20)
6
1(17)
16
3(19)
10
10
4(36)
8
0(0)
18
4(22)
11
7
4(37)
10
2(20)
17
6(35)
12
13
5(39)
5
2(40)
18
7(39)
13
4
1(25)
4
1(25)
8
2(25)
Total
57
19(33)
38
6(16)
95
25(26)
The first six-months reinfection rate was higher than the second six-months of reinfection as the
former period coincided with the dry season when more snails and higher cercarial infection were
observed. This suggests that higher transmission takes place during the dry season. Nevertheless, the
observation of cercarial infection during the dry season by no means rule out the possibility of
intermittent transmission in other seasons of the year. In a lake ecology, in northwestern part of
Ethiopia, heavy transmission in the dry season with little intermittent transmission in the rest of the
year has been observed (9).
Seasonality in snail population density and cercarial infection appear to be mainly influenced by
rainfall cycles. Decline in snail population during the rainy season could result from direct and
indirect effects of rainfall. High turbidity or siltation caused by flooding and increased water velocity
observed during the rainy season could be categorized as an indirect effect of rainfall while splashing
out of snails in a flood is a direct effect of rainfall. Both of these factors appeared to affect the snail
population density and cercarial infection negatively
In reinfection of schoolchildren with Schistosoma mansoni, age and sex of the subjects and
278 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
seasonality of infection in snails appeared to play important role in influencing the rate at which
individuals become infected. Detailed exposure and immunological studies involving all age- groups
are essential to assess the effect of exposure and immunity on reinfection patterns of schistosomes.
The observation that prevalence and intensity level rose to the pre-treatment level in a year after posttreatment implies that chemotherapy alone is not sufficient for the control of schistosomiasis.
Acknowledgement
This study received financial assistance from the Research and Publications Office of Addis Ababa
University. The authors wish to express their gratitude to Messrs Abrham Redda, Negash Gemeda,
and Sisay Dessie for their technical assistance.
References
1. Warren KS. Selective primary health care: strategies for control of disease in the developing world.
I. Schistosomiasis. Rev Infect Dis 1982;4:715-726.
2. Dalton PR, Pole D. Water contact patterns in relation to Schistosoma haematobium infection. Bull
wld Hlth Org 1978;56:417-426.
3. Sturrock RF, Kimani R, Cottrell BJ, Butterworth AE, Seitz HM, Siongok ATK, Houba V.
Observations on possible immunity to reinfection among Kenyan school children after treatment for
Schistosoma mansoni. Trans Roy Soc Trop Med Hyg 1983;77:366-371.
4. Butterworth A, Capron M, Cordingley S, Dalton PR, Dunne DW, Kariuki HC et al. Immunity after
treatment of human schistosomiasis mansoni. II. Identification of resistant individuals, and analysis of
their immune responses. Trans Roy Soc Trop Med Hyg 1985;79:393-408.
5. Hagan P, Wilkins HA, Blumenthal UJ, Hayes RJ,and Greenwood BM. Eosinophilia and resistance
to Schistosoma haematobium in man. Parasite Immunology 1985;7:625-632.
6. Wilkins HA, Blumenthal UJ, Hagan P, Hayes RJ, Tulloch S. Resistance to reinfection after
treatment of urinary shistosomiasis. Trans Roy Soc Trop Med Hyg 1987;81:29-35.
7. Kloetzel K. Reinfection after treatment of schistosomiasis: Environment or “Predisposition”?
Revista Do Instituto De Medicina Tropical De Sao Paulo 1990;32:138-139.
8. Engels D, Ndoricimpa J, Gryseels B. Schistosomiasis mansoni in Burundi: Progress in its control
since 1985. Bull Wld Hlth Org 1993;71:207-214.
9. Erko B, Tedla S, and Petros B. Transmission of intestinal schistosomiasis in Bahir Dar Northwest
Ethiopia. Ethiop Med J 1991;29:199-211.
10. Abebe F, Tedla S, Birrie H, and Medhin G. Transmission dynamics of Schistosoma mansoni in an
irrigation setting in Ethiopia. Ethiop J Hlth Dev 1995; 9:147-156.
11. Frandsen F, and Christensen NØ. An introductory guide to the identification of cercariae from
African freshwater snails with special reference to cercariae of trematode species of medical and
veterinary importance. Acta Tropica 1984;41:181-202.
12. Kloos H, and Lemma A. The epidemiology of Schistosoma mansoni infection in Tensae Berhan:
Human water contact patterns. Ethiop Med J 1980;18:91-98.
Neonatal mortality in Addis Ababa 279
────────────────────────────────────────────────────────────
Original article
Neonatal mortality among hospital delivered babies
in Addis Ababa, Ethiopia
Yodit Sahle-Mariam1 and Yemane Berhane2
Abstract: A follow up study to determine the magnitude of neonatal mortality and identify its
determinants was done on a cohort of babies born at the health institutions in Addis Ababa. Baseline
information on risk behaviour of the mother and child characteristics were recorded just after delivery.
Then, follow up interviews were done on the 7th and 28th day by visiting each neonate/mother at
home. A total of 1334 singleton newborns were included in the study. The neonatal mortality rate was
found to be 71.9 per 1000 live birth with early and late neonatal mortality rates of 50.9 and 20.9 per
1000 live berth, respectively. Low birth weight and prematurity were associated with a higher risk of
neonatal death. Establishing essential neonatal care facilities at all levels and further study among
home deliveries are recommended. [Ethiop. J. Health Dev. 1997;11(3):275-281]
Introduction
Developing countries are known to carry a heavy burden of diseases and death, many folds higher
than the developed countries. This is seen mainly in vulnerable groups such as children and women in
the reproductive phase of life; in these groups inadequate nutrition, physiologic demands and lack of
resistance make the effects of diseases more serious(1,2).
As health services coverage are low and many births are occurring at home in most developing
countries, there is no reliable estimate of neonatal mortality. Published official statistics usually
present a biased image of the public health and in almost every case underestimate the health
problems(3,4). However, hospital-based studies from the developing world have indicated that about
half of the infant deaths occur during the neonatal period and estimate neonatal mortality from 20 to
106 per 1000 live birth(5-9). Generally, hospital-based data tend to over estimate mortality because
of the selection factor of over representation of the severely sick children.
Like many other priority health problems in the developing countries, factors related to the
increased risk of neonatal mortality are generally preventable. Immature maternal reproductive
development because of early marriage, high family size, short spacing, single parent and low
educational status of the mothers are some of the identified risk factors for low survival of the
newborns, all of which have their roots in poverty and inequity( 10-14).
In addition to the above demographic factors poor nutrition, cigarette smoking, substance use, poor
health services during pregnancy and during postpartum, and harmful customs, beliefs and taboos
influence neonatal mortality(15-19). Child factors associated with low survival include low birth
weight, birth asphyxia, trauma, and infections such as syphilis, malaria and neonatal tetanus(2023).Developing countries need to develop preventive strategies to alter this situation. Useful mesures
include placing greater emphasis on identifying and reducing risks before pregnancy, and expansion of
basic obstetric services to the majority of the population. As neonatal morbidity
______________________________________
1
From the Family Health Team, Addis Ababa Health Bureau, Addis Ababa, Ethiopia, and
2
Department of Community Health, Faculty of Medicine, Addis Ababa University P.O.Box 9086,
Addis Ababa, Ethiopia
and mortality in developing countries could be largely uninfluenced by technological advances only in
280 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
the field of neonatology, it is very important to put more emphasis on improving nutritional and other
basic health requirements of the children(24,25).
Methods
Study Area: Addis Ababa is the capital city of Ethiopia with an estimated population of 2.4 million.
According to a census conducted in 1984, its population growth rate is 5.1% and the fertility rate was
3.2%(26). High migration rates from the countryside by people looking for employment opportunities
and a better life have contributed to the tremendous increase of the capital's population. Addis Ababa
is administratively divided into six zones.
The city has a relatively higher number of health institutions compared to the other parts of the
country. There are 12 hospitals, 14 health centres, about 74 clinics and 24 health posts. Out of these
health institutions 5 of the hospitals, 12 of the health centres and 5 of the clinics render delivery
services. In the year preceding the study a total of 32,000 deliveries were reported by these health
institutions, which is 50-60% of the expected annual deliveries in the city.
Study design and population: The study was conducted on a cohort of newborns delivered in the
health institutions of zone 1 and zone 2 in the city of Addis Ababa, Ethiopia. The two zones were
selected because they are the central old parts of the city where population density is high and health
institutions exist in adequate number to generate sufficient sample within the time frame of the study.
Recruitment for the study was done during November and December, 1994 in all health institutions
which were providing delivery service in the two zones. To avoid the complex interaction of factors
and for logistic reasons only single live birth babies whose parents were residents of Addis Ababa were
enrolled into the follow-up study. Stillbirth, multiple deliveries and babies whose parents were not
residents of Addis Ababa were excluded from the study. Assuming a neonatal mortality rate of 10%,
with 90% power and 95% certainty, the sample size required for the study was calculated to be 1365.
Data collection and analysis: Baseline data and follow-up information were collected using
pretested questionnaire. The questionnaire was prepared in English and later on translated into
Amharic for feild use. Midwives and high school graduates collected the data. The midwives
completed the baseline questionnaire and the high school graduates filled in the follow-up
questionnaire at home on the 7th and the 28th day after birth. A 3-days training was given to
everybody involved in the project. The questionnaire was pretested on mothers who delivered during
the training period in the health institutions that participated in the study. Based on the pretest few
adjustments were made in the final version of the questionnaire. Regular checking of data quality was
conducted by the research coordinator. The principal investigator monitored the overall quality and
conduct of the study. Data were entered, stored and analyzed using Epi Info version 5 statistical
package. Multivariate analyses were done using SAS statistical software.
Ethical considerations: Informed consent was obtained from all study participants. Mothers who
volunteered to participate kindly provided a description of their residential area and their house
number as well as telephone numbers(either own or neighbour) to facilitate the home visits.
Informations were kept confidentially by the principal investigator. Referral was arranged for those
neonates who were reported to have been sick during home visits.
Results
A total of 1606 deliveries took place during the enrolment period, from November 8 to December
22, 1994. Out of these 180 (11%) were from outside Addis Ababa, 38 were still- births, two were
triple deliveries (6 babies) and 25 were twin deliveries (50 babies), and the remaining were 1334
singleton live births, who were residents of Addis Ababa. Hence, the analysis was focused on the,
1334 singleton live births. Of these 693 (51.9%) were male, and 641(48.1%) were female babies.
Of the 1334 mothers 1160 (87%) were married and living in union. Also the majority (87.4%) of
them reported to had at least one ANC consultation during their last pregnancy and 42.1% of mothers
Neonatal mortality in Addis Ababa 281
────────────────────────────────────────────────────────────
were family planning service users before the last pregnancy. Cigarette smoking, chat chewing and
alcohol consumption, were reported by 21(1.6%), 93(6.9%) and 217(16.2%) of the mothers,
respectively, during their last pregnancy.
On the 7th day home visit, of 1334 live singleton babies, 61 were lost to follow-up and 68 neonate
were reported to have died. On the 28th day visit, 10 were lost to follow-up and 28 neonate were
reported to have died. Hence, neonatal mortality rate was 71.9 per 1000 live birth(95% confidence
interval: 58.0 to 85.8 per 1000 live births with early and late neonatal mortality rates of 50.9(68/1334)
and 23.2(28/1205) per 1000 LB, respectively.
These rates showed that mortality was higher during the very early life of the babies. The perinatal
mortality was 83 per 1000 total births(134/1606).
The distribution of maternal characteristics and neonatal death, and the crude and adjusted odds
ratios for the risk factors included in the logistic regression model are shown in the Table 1. Those
babies born from mothers who were non-users of family planning methods were found to be at a higher
risk of neonatal death in the multi-variate analysis. Though not statistically significant in the logistic
model, babies born from single mothers and low income family were also found to be at a higher risk
of neonatal death than babies born from married mothers and high income families. Also babies born
from highly educated mothers and who use ANC services were at a lower risk of dying in the crude
analysis compared to the other categories(Table 1).
Table 1: Neonatal mortality and related maternal characteristics in Addis Ababa.
Maternal age
15-19
20-24
25-29
30-34
35+
Marital status
Married
Others
Income
0-100
01-250
251-600
601+
Maternal education
Illiterate
1-8 grade
9-12 grade
12+
ANC during last pregnancy
Used
Not used
Family planning service
Non-user
User
Population
Death number(%)
Crude OR(95%CI)
Adjusted OR(95%CI)
125
362
382
236
158
9 (7.2)
29 (8.0)
21 (5.5)
22 (9.3)
15 (9.5)
1.00
1.13(0.54,2.38)
0.76(0.34,1.70)
2.11(0.72,6.20)
1.37(0.58,3.23)
1.00
1.18(0.48,2.95)
1.12(0.41,3.07)
2.11(0.72,6.20)
1.60(0.47,5.50)
1104
159
70 ( 6.3) 1.00
26 (16.4) 2.92(1.80,4.74
1.00
1.30(0.61,2.80)
279
277
453
254
37 (13.2)1.00
25 (10.5)0.65(0.38,1.11)
22 ( 4.9)0.33(0.19,0.57)
12 ( 4.7)0.32(0.17,0.64)
1.00
0.99(0.49,1.99)
0.57(0.27,1,19)
0.48(0.18,1.27)
202
476
499
86
24 (11.9)1.00
33 ( 6.9)0.56(0.32,0.97)
35 ( 7.0)0.56(0.32,0.97)
4 ( 4.7)
1.00
0.80(0.40,1.59)
0.85(0.40,1.79)
0.36(0.12,1.07)
1105
158
70 ( 6.3)1.00
26 (16.5)2.90(1.78,4.69)
1.00
1.66(0.87,3.17)
736
527
66 (9.0)
30(5.7)
1.00
0.61(0.39,0.96)
Table 2: Selected Maternal Behaviour and Neonatal Mortality in Addis Ababa.
Maternal Behaviour
Cigarette smoking
Yes
No
Chat chewing
Yes
No
Alcohol consumption
Yes
Population
Death No.(%)
Crude OR(95%CI)
Adjusted OR(95%CI)
19
1244
6 (31.5)
90 (7.2)
5.92(1.95,17.20)
1.00
1.22(0.33,4.39)
1.00
85
117
15 (17.6)
81 ( 7.4)
2.90(1.52,5.48)
1.00
0.87(0.39,1.95)
1.00
209
21 (10.0)
1.46(0.05,2.49)
0.81(0.80,1.49)
0.62(0.15,2.58)
1.00
0.50(0.29,0.87)
282 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
No
1054
75 (7.1)
1.00
In this study the logistic regrassion analysis did not show significant association between cigarette
smoking, chat chewing and alcohol consumption, and neonatal mortality(Table 2). Among the child
characteristics studied gestational age and birth weight were shown to be good predictors of neonatal
survival. Term and normal weight babies had a significantly lower risk of dying: adjusted odds ratios
were 0.33(95% CI, 0.19-0.59) and 0.06(95% CI, 0.04-0.12) for term and normal weight babies
respectively(Table 3).
Discussion
The neonatal mortality rate observed in this study is consistent with previous estimates (5,9).
However, a slight overestimation may have occurred because of the selection of hospital delivered
babies in which high risk mothers could be over-represented. According to the findings of this study,
neonatal mortality may account for 65% of the infant mortality in this country, if one considers the
official national infant mortality rate of 110/1000 live births(26). The observation of a significantly
higher risk of neonatal mortality among the low birth weight and preterm babies is also in
conformity with previous reports from the developing countries(4,6,12).
Though not statistically significant in the logistic analysis, the difference observed in the crude
analysis between the illiterate and the highly educated mothers was remarkable, that is, babies born
from highly educated mothers were at a lower risk of mortality than babies born from illiterate
mothers. Maternal education is believed to be one way of reducing mortality during infancy, since it
provides the mother with the necessary skill for child care(15).
The effect of marital status on newborn survival is believed to have been over- shadowed in this
study since the majority of the mothers were married. The effect of family income and maternal age on
neonatal survival noted with other studies were not also obvious in this study. This might be due to the
imprecise reporting of these variables either because of genuine ignorance or taboos related with these
variables. Both variables are socially sensitive in the Ethiopian context. This observation was also
made by Kloos(27).
In this study babies born from mothers who were family planning users were found to be at a lower
risk of neonatal mortality than those born from non-users, which is consistent with other studies
too(27). Unlike other studies antenatal care (ANC) attendance showed no statistically significant
association. This may be due to the fact that no consideration was given to the number of visits made.
Obviously, only one visit may not make any difference.
Although cigarette smoking, chat chewing and alcohol consumption by pregnant women were
indicated to influence neonatal mortality in other studies(18,19), it was not obvious in this study. The
reason for that could be the lack of information on the amount and duration of their use as well as the
small number of mothers who reported using these substances.
Losses to
follow-up were small, only 71(5.3%) of cohort member were lost on the follow-up visits. When
considering the complexity of doing a follow-up study in a
Neonatal mortality in Addis Ababa 283
────────────────────────────────────────────────────────────
Table 3: Newborn characteristics and neonatal mortality in Addis Ababa.
Sex
Male
Female
Gestational age
(in weeks)
28-37
>37
Birth weight
(in grams)
<2500
2500
Birth order
1st
2nd-4th
5th +
Population
Death number(%)
Crude OR(95%CI)
Adjusted OR(95%CI)
656
607
51(7.8)
45(7.4)
1.00
0.95(0.63,1.45)
1.00
0.74(0.46,1.22)
223
35(15.6)1.00
1.00
1040
61( 5.9)
0.34(0.22,0.52)
0.33(0.19,0.59)
112
1151
49(43.8)1.00
47(4.1)
1.00
0.06(0.03,0.09)
0.06(0.04,0.12)
527
514
222
42(8.0)
39(7.9)
15(6.8)
1.00
0.95(0.60,1.50)
0.84(0.46,1.55)
1.00
0.54(0.81,2.88)
0.71(0.28,1.82)
large urban setting like Addis Ababa, where locating houses is extremely difficult to a stranger, the loss
was negligible and it could not have any significant effect on the findings of the study. However, if we
take all the lost as alive or as dead, the neonatal mortality ranges from 7.1 to 12.4/1000 live births.
This provides the worst and the best scenarios.
Though the use of new borns delivered in health institutions could be regarded as the weakness of
this study, the other alternative of including home deliveries was practically impossible for obvious
logistic reasons. However, the study gained its strength for successfully following the babies at home,
which was very important since in developing countries a significant proportion of deaths occur
outside health care facilities. Hence, the selection procedure may limit the generalizability of the
findings but not the validity. To make further generalizability of the findings we recommend a similar
study among home delivered babies. A fairly high sample size was used to minimize the role of
chance. Bias was also minimized by using uniform inclusion criteria and achieving high follow-up
rates. Reliability was maintained by prior training of supervisors and interviewers, by using pretested
questionnaire, and by regular supervision.
Multivariate analysis was utilized for control of
confounding.
In conclusion, the major risk factors identified for increased risk of neonatal mortality in this study,
low birth weight and prematurity, seem universal in the developing countries and these factors are
deeply rooted in poverty and social inequalities. Therefore, a logical strategy to reduce neonatal
mortality at the moment could be to improve and/or expand neonatal care facilities and access to such
services while exploring for long-lasting solutions through economic and social development. In
Addis Ababa, equipping all health institutions to provide the essential neonatal care and education of
mothers to use such services deserve due consideration(28). Currently the Ethio-Swedish Children's
Hospital is the only centre capable of providing such service in Addis Ababa. It has to be noted that
unless we are able to bring down such very high risk of child mortality starting from the early time of
life, we can not be successful in our global efforts to regulate fertility and to promote safe motherhood.
Acknowledgment
Financial support for the study was obtained from the International Development Research Centre
(IDRC-Canada). Authors are grateful to the Department of Community Health, Faculty of Medicine,
Addis Ababa University for providing material support. Special thanks
goes to the health workers at the delivery service units of Tikur Anbessa, Zewditu and Gandhi
Hospitals, and Lideta, Higher 23, Addis Ketema and Tekle-Hiamanot health centres who were very
helpful in the data collection.
284 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
References
1. Hank M, Horbar JD, Malloy IH. Very Low Birth Weight Outcomes of the National Institute of
Child Health and Human Development, Neonatal Care Practices, May 1991;87(5):587-596.
2. Velema JP, Alihonou EM, Gandaho T, Hounye FH. Childhood Mortality among Users and Nonusers of primary health care in a Rural West African Community. International Journal Of
Epidemiology. 1991;20(5):474-479.
3. The Cebu Study Team. Underlying and Proximate Determinants of Child Health: The Cebu
Longitudinal Health and Nutrition Study. American Journal of Epidemiology. 1991;133(2):185-200. 4.
Lumbiganon P, Pothinam S, Patithat N . Why are Thai Official Perinatal and Infant Mortality Rates so
Low? International Journal Of Epidemiology. 1990;19(4):997-1000.
5. Leroy O, Garenne M. Risk Factors of Neonatal Tetanus in Senegal. International Journal of
Epidemiology. 1991;20(2):521-525.
6. UNICEF. Children and Women in Ethiopia: a situation analysis. The Children, Family and Youth
Service Organization, Addis Ababa, Ethiopia, August 1989.
7. Linge K, Rogo KO. Factors Influencing Early Perinatal Mortality In A Rural District Hospital. East
African Medical Journal. 1992;69(4):184-187.
8. Taha TE, Gray RH, Abdelwahab MM, Abdelhafeez AR, Abdelsalam AB. Levels and Determinants
of Perinatal mortality in Central Sudan. International Journal of Gynaecology and Obstetrics.
1994;45:109-115.
9. Bartlett AV, Elizabeth M, Antonio M. Neonatal and Early Post neonatal Morbidity and Mortality
in a Rural Guatemalan Community: The Importance of Infectious Diseases and Their Management.
Paediatrics Infectious Disease J. 1991;10:752-7.
10. Singh M. Hospital-based data on Perinatal and Neonatal Mortality in India. Indian Paediatrics.
1986;23:579-584.
11. Ghide E, Assefa M, Mohammed S, Tessema F . Analysis of Deliveries in Jimma Hospital: A Four
Year Retrospective Study. Ethiopian Journal of Health Development. 1991;5(1):3-6.
12. Sukanich AC, Rogers KD, Mcdonald HM. Physical Maturity and Outcome of Pregnancy in
Primiparas Younger Than 16 Years of Age. Paediatrics 1986;78(1):31-36.
13. Pritchard CW, Mfphm T. Preterm Birth, Low Birth Weight and The Stressfulness of the
Household Role For Pregnant Women. Social Science and Medicine. 1994;38(1):89-96.
14. Naeye RL, Tafari N. Risk factors in pregnancy and diseases of the fetus and newborns. Williams
and Wilkins. Baltimore. 1983:213-26.
15. Kinfu Y. Maternal Education and Child Survival in Addis Ababa. Ethiopian Journal of Health
Development. 1992;6(1):11-18.
16. Huttly SR, Victora CG, Barros FC, Vaughan JP. Birth Spacing and Child Health in Urban
Brazilian Children. Paediatrics. 1992;89(6):1049-1053.
17. Mavalankar DV, Trivedi CR, Gray RH. Levels and Factors or Perinatal Mortality in Ahmedabad,
India. Bulletin World Health Organization. 1991;4:435-42.
18. Fox SH, Koepsell TD, Janet RD. Birth Weight and Smoking During Pregnancy - Effect
Modification by Maternal Age. American Journal of Epidemiology. 1994;139(10):1008-1011.
19. Ghani NA, Eriksson M , Kristiansson B, Qirbi A. The Influence of Khat-Chewing on Birth-Weight
in Full-Term Infants. Social Science and Medicine. 1987;24(7):625-627.
20. Dale LP, David RB, Betty T. 28-Day Survival Rates of 6676 Neonates with Birth Weight of 1250
Grams or Less. Paediatrics. 1991;87(1):7-15.
21. Zein AZ. The distribution of low birth weight in the Gondar administrative region North Western
Ethiopia. Ethiopian Journal of Health Development. 1991;5(2):71-74.
22. Victora CG , Smith PG, Vaughan JP. Influence of Birth Weight on Mortality From Infectious
Disease: A Case-control Study. Paediatrics. 1988;81(6):807-811.
23. Edith C. Kieffer G. Geographic patterns of low birth weight in Hawaii. Social Science and
Medicine. 1993;36(4):557-564.
24. Gkay RH, Ferraz EM. Levels of Determinants of Early Neonatal Mortality in North and North
eastern of Brazil: Results of a surveillance and case- control study. International Journal of
Epidemiology. 1991;20(2):467-472.
25. Grant JP. The state of the World Children, Oxford University Press, New York, UNICEF. 1991.
26. Central Statistic Authority. Population and housing census of Ethiopia, Analytical report on
results for Addis Ababa. Central Statistics Authority. Addis Ababa 1987.
27. Kloos H, Degefa A, Hundessa A, et al. Illness and health behaviour in Addis Ababa and rural
central Ethiopia. Social Science and Medicine. 1987;25:1003-19.
28. Majinge CR and Lema VM. Pregnancy intervals: Their determinants and fetal outcome at the
KCMC, Moshi, Tanzania. East African Medical Journal. 1993;70(9):544-550.
Original article
Infant survivorship and occurrence of multiplebirths: A longitudinal community-based study, south
west Ethiopia
Makonnen Asefa1, Fasil Tessema1
Abstract: A one year live-birth cohort was studied in 46 urban and 64 rural ‘kebeles’ in south-west
Ethiopia, in 1992-94. In order to recruit all live-births in each of the study kebeles, pregnant women
were identified in their second trimester and monitored by trained TBAs and enumerators using house
to house visit. Each infant-mother pair of the cohort was visited soon after birth followed by regular
bimonthly visit to the end of their first year or to an earlier death. Infant mortality was calculated as a
life tables estimate using the ‘survival’ programme in SPSS. There were a total of 8162 births (8050
singletons, 111 twins and one set of triplets) of which 856 died, indicating an infant mortality of
104.8/1000 (estimated probability of surviving to 1 year 0.8952, with s.e. 0.004). The occurrence of
multiple-births was 13.6/1000 live births. The infant mortality among them was 446.8/1000 (estimated
probability of surviving to 364 days .5532, with s.e. 0.0034). This is a matter of serious concern. This
study is the first of its kind based on rigorous study design, bigger sample size comprising different
population groups and wider areas. The findings could help to formulate policy and health care
programmes. [Ethiop. J. Health Dev. 1997;11(3):283-288]
Introduction
Vital events registration is non-existent in Ethiopia (1, 2). In such settings, a longitudinal
community-based study is needed to generate sound data for formulation of policy and health care
programmes. The scanty data available on infant survivorship in the country are mainly based on the
preceding 12 months recall period which could lead to bias due to omissions and vaguely defined
recall period (1, 3-6). New-born deaths are not reported due to local taboos (1, 7). Reports on the
occurrence of multiple-births and survivorship in the country are based on hospital data (8, 9).
In order to generate a reliable data, a community-based prospective follow-up study of live-birth
cohorts which incorporates Jimma, Illubabor, and Keffecho administrative zones in South-West
Ethiopia was undertaken.
Methods
This study was conducted in 46 urban and 65 rural ‘kebeles’, with an estimated population of 300,200
in the administrative zones of Jimma, Illubabor, and Keffecho, South-West Ethiopia in 1992-94.
Figure 1 shows the sketch map of the study area. The altitude of the study kebeles ranges from 1500 m
to just below 2000m. The area’s main crops include maize, sorghum and coffee. Trained enumerators
and TBAs identify expectant mothers in their second trimester by house to house visits in their
respective catchment kebeles. In order to identify all live-births, all the traditional birth attendants
(TBAs) in the above mentioned kebeles were involved in the field-work. TBAs are women residents
of the kebele they serve and by tradition they visit and assist women during pregnancy and delivery.
The
______________________________________
1
From the Department of Community Health, Jimma Institute of Health Sciences, P.O.Box, 378,
Jimma, Ethiopia.
TBAs had easy access to women in the fertile age group, and were able to assess their pregnancy
status. Each TBA was given responsibility for about 300 houses, and went house to house regularly to
Infant survivorship and occurrence of multiple-births 287
────────────────────────────────────────────────────────────
locate pregnant women in their second trimester. The TBA reported daily in person to the enumerator
responsible for her kebele. For each three kebeles one high school completed girl enumerator was
assigned. The enumerator registered the address of the expectant mother. After registration both the
TBA and the enumerator monitored the expectant mother so as to reach her on time soon after
delivery. For each three or more enumerators one supervisor (mainly nurses) was assigned who
checked and supported data collection. In each kebele all one year live-birth cohorts were recruited for
this study and followed for one year from 1992 to 1994. Data collection was made at birth, on infants
anthropometric and mothers socio-economic variables. Then regular follow-up was made bimonthly
until their first birthday or to an earlier death. Information was also gathered from mothers, key
informants on their views about multiple births. Two high school completed students were trained for
computer data entry and two statisticians were responsible for data processing and analysis. Data were
entered daily, so as to allow fast feedback for quality control procedures. Details of the study
methodology have been given elsewhere (10).
Results
In this study a total of 8162 deliveries (8050 singletons and 111 twins, one set of triplets) were
recruited in one year period in the study areas. The overall incidence of multiple-births was 13.6 per
thousand deliveries. The incidence for twins and triplets was 13.5 and 0.12 per thousand deliveries,
respectively. By the end of the one year follow-up 450 infants were known to have moved out of the
study area, 40 infants were withdrawn from the study by their mothers, 106 infants were lost for
follow-up and 856 infants have died. Of 8275 infants about 92.8% were successfully followed up to
their first birthday or to an earlier death. Infant mortality was calculated as life tables estimates using
the survival program in SPSS/PC+ (11) and presented in Tables 1 and 2. For singletons it was
Table 1: Infant Mortality Rate per 1000 live - births by Age (in days), Birth Type and Sex
Age
7
28
90
180
364
Male
14.0
29.6
53.3
74.7
103.7
Single
Female
12.2
22.7
43.9
61.2
86.6
Both
13.1
26.2
48.7
68.0
95.3
Male
151.5
222.2
333.3
394.3
477.8
Multiple
Female
169.4
233.9
282.3
347.9
417.9
Both
161.4
228.7
304.9
368.5
446.8
male
17.2
34.1
60.0
82.3
112.8
All Birth
Female
17.0
29.2
51.2
70.0
96.8
95.3/1000 live-births (estimated probability of surviving to 364 days .9047, with s.e. 0.003). For
multiple-births it was 446.8/1000 live-births (estimated probability of surviving to 364 days .5532,
with s.e. 0.034). Mortality rate was higher for males (103.7 and 477.8) compared with females (86.6
and 417.9) for singletons and multiple-births. For all types of births, infant mortality rate for the whole
study area was 104.8/1000 (estimated probability of surviving to 364 days .8952, with s.e. 0.004). It
was 96.1/100 (estimated probability of surviving to 364 days .9039, with se. 0.005) for urban and
112.5/100 )estimated probability
288 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 2: Infant Mortality Rate per 1000 live-births at Different ages at Urban and Rural Settings
Age (in days)
Rural
Urban
Total
p
q
p
q
p
7
.9814
18.6
.9846
15.4
.9829
(.002)
(.002)
(.001)
q
17.1
28
.9643
(.003)
35.7
.9730
(.003)
26.0
.9683
(.002)
31.7
120
.9269
(.004)
73.1
.9465
(.004)
53.5
.9358
(.003)
64.2
180
.9137
(.004)
86.3
.9360
(.004)
64.0
.9238
(.003)
76.2
364
.8875
(.005)
112.5
.9039
(.005)
96.1
.8952
(.003)
104.8
of surviving to 364 days .8875, with s.e. 0.005) for rural (Table 2). The mean age of mothers was
found to be 26.34(6.29) for singletons and 27.87(6.10) for multiple-births (Table 3).
The peak age for occurrence of maximum frequency of multiple births was 25 to 29 years. As
shown in table 3 most mothers were in the low income group, married, delivered at home and illiterate.
In this cohort data collection was mainly conducted in the field. Anthropometric measurements,
gathered within seven days of birth, were considered as reflecting birth measurements. It should,
however, be borne in mind that newborns lose up to 10% of their weight in this period (12).
Accordingly birth-weight was collected for 7426 singletons and 126 (each) of the multiple
Infant survivorship and occurrence of multiple-births 289
────────────────────────────────────────────────────────────
Table 3: Socio-demographic Characteristics and Fertility History of Mothers by Birth Type
Birth type
Characteristics
Age group
15-19
20-24
25-29
30-34
35-39
40-44
45-49
Mean
SD
Marital Status
Married
Single
Divorced
Widowed
Ethnic group
single
%
multiple
%
962
2158
2090
1475
1076
215
51
26.34
6.29
12.0
26.9
26.0
18,4
13.4
2.7
0.6
7
22
34
24
20
4
0
27.87
6.1
6.3
19
30.6
21.6
18.0
3.6
0.0
7467
297
204
70
92.9
3.7
2.5
0.9
100
6
3
2
90.1
5.4
2.7
1.8
Oromo
Amhara
Tigre
Daworo
Keffa
gurage
Yem
Other
Religion
5387
650
110
492
601
410
289
89
67.1
8.1
1.4
6.1
7.5
5.1
3.6
1.1
68
11
3
11
6
8
3
1
61.3
9.9
2.7
9.9
5.4
7.2
2.7
0.9
Muslin
Christian
Monthly Income (birr)
5550
2476
69.2
30.8
63
48
56.8
43.2
<150
150-299
300+
Place of delivery
Health Institutions
Home
Educational status
6065
1017
968
75.3
12.6
12.0
82
21
8
73.9
18.9
7.2
1316
6734
16.3
83.7
30
81
27
73.0
Illiterate
1-6 grade
7-8 grade
9-12 grade
Higher
Total live birth
4890
1806
572
697
67
60.9
22.5
7.1
8.7
0.8
65
30
8
8
0
58.6
27.0
7.2
7.2
0.0
1-3
4-6
>7
# of Abortions
4592
2375
1073
57.1
29.5
13.3
40
37
34
36.0
33.3
30.6
None
One or more
# of still births
7438
612
92.4
7.6
97
14
87.4
12.6
None
One or more
7882
168
97.9
2.1
107
4
96.4
3.6
births. For singletons the mean (Sd) birth weight was 3081.8 ± 488 g, and for multiple-births 2140.1 ±
410 g. Babies <2500 g were 9.5%, 76.2% and 9.8 among singletons, multiple-births and all types of
births, respectively.
290 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Table 4: Socio-demographic characteristic and infant mortality rate per 1000 live-births
Factor
n
Death Incidence
RR
8263
3792
4471
Cum
Surv Proprn.
.8952
.9039
.8875
Total
Urban
Rural
95% C.I
104.8
96.1
112.5
1.17
1.03-1.33
Amhara
Tigre
Daworo
Keffa
Gurage
671
116
513
613
426
.9139
.9200
.8790
.8919
.9215
86.1
80.0
121.0
108.1
78.5
1.27
1.41
.91
1.02
1.42
0.98-1.64
0.75-2.66
0.71-1.16
0.80-1.30
1.01-1.98
Yem
Oromo
295
5516
.9041
.8904
95.9
109.6
1.16
0.81-1.16
Christian
Muslim
2560
5669
.9031
.8914
96.9
108.6
1.12
0.98-1.29
300+
150-299
<150
983
1060
6220
.9395
.9043
.8869
60.5
95.7
113.1
1.87
1.18
1.46-2.43
0.97-1.45
Literate
Illiterate
3230
5015
.9156
.8823
84.4
117.7
1.39
1.21-1.60
6802
8263
.9237
.7705
76.3
229.5
3.00
2.57-3.51
Population group
Income birr
Birth weight
2500+ gm
<2500 gm
Qualitative information was gathered from mothers (multiple-births) to see their views about
multiple-births. Majority of mothers feel multiple-births are due to god’s will and family trend. Again
majority of mothers and their family members were not happy of having multiple-births. Most mothers
complain that multiple-births are economic burden and difficult to care for. None wished to have more
multiple-births. Discussion with key informants reflect a similar view.
The frequency of births was seen with respect to the months at which the births occurred (Figure 2).
Multiple-births most frequently occurred in the months of March, September and December, but it
was relatively constant in all the months for singletons.
Infant survivorship and occurrence of multiple-births 291
────────────────────────────────────────────────────────────
Figure 1: Sketch map of the study area
Discussion
In this country mortality reports rely on health services data and on surveys based on the preceding
12 months recall period which could bias due to omissions and vaguely defined period (1, 3, 4, 5, 7).
This paper discusses the results of a longitudinal community-based generated data which was
undertaken to overcome the above mentioned problems. This study is the first of its kind which covers
diversified populations and bigger areas and bigger sample size (Table 3, 4). The estimated infant
mortality was 104.8/1000 live-births for the whole study area and 96.1 and 112.5 for urban and rural
areas, respectively. As depicted in Table 4, there is variation in infant mortality rates among the
different population groups. Those with lower infant mortality rate are mainly residents of urban
settings. The proportion of people with better income and education is relatively higher in urban
populations compared to rural (13, 14). The observed lower infant death rate among the above
mentioned groups could be explained by income and education factors as shown in Table 4. The
influence of these factors has also been shown in other studies (15, 16, 17).
According to the Central Statistical Authority’s 1984 census, based on the preceding 12 months
recall period, the infant mortality rate was 105/1000 for the whole country (1). In Butajira district,
Southern Shewa, a survey involving nine rural and one urban kebeles, from 1987 to 1990, based on
follow-up approach, reported infant mortality rate of 114/1000 (18). According to the 1994 National
Census, for regions 3 and 4 which comprise the major bulk of the nation’s population, the infant
mortality rate was 116 and 118 per 1000, respectively (13, 14). The same census result also showed
102 for urban and 117 for rural (region 3), and 93 for urban, 121 for rural (region 4). Based on
hospital data, the occurrence of multiple-births vary. In Addis Ababa deliveries in health facilities in
1973 and 1982 showed 43 and 33 per 1000, respectively (8). It was 14.9 per 1000 deliveries in
Gondar, based on the hospital deliveries between 1977 and 1985 (9). The above findings are mainly
for urban settings. In the present study the finding of 13.6 multiple-births per 1000 live-births is based
on both rural and urban community data of one year live-births cohort follow-up study. Reports from
other African countries (Nigeria, Tanzania, Zimbabwe) is about or over 30 per 1000 live -births (19,
20). For this paper, in all 46 urban and 65 rural kebeles, the rigorous method we followed to recruit all
live-birth cohorts has helped to avoid omission and recall bias.
Figure 2: Percent distribution of births by month of birth, South West Ethiopia
The infant mortality of 446.8/1000 among multiple-births is over four folds compared to singletons
in this study. This is a very high death toll and a matter of serious concern. The contributing factors
could be both social and biological. As mentioned in the results section, the attitude of people
292 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
(including the mother and her family) is not favourable towards multiple-births. The low birth-weight,
among this group could also partly explain the high mortality. Previous studies (17,21) had clearly
indicated birth-weight as an important factor for survival differentials as is also shown in this study
(Table 4). This entails better approach and care for multiple gestation during antenatal and early
infancy periods.
The seasonality of births, as shown in figure 2, was relatively constant in all the months for
singletons but there was a peak in the months of March, September and December for multiple-births.
Reviews of studies on seasonality of births showed agricultural cycle, economic variables, marriage
rate, weather..... as possible proximate influencing factors (22).
The review of studies clearly showed that there is birth seasonality in different countries like South
Africa, USA, India, Japan (22) but the relative constant birth in all the months for singletons in this
study seems a unique pattern. It would be better to have serial data on births over the years to make a
meaningful discussion.
The experience of this study on vital events registration using the existing grass-root level human
resources could lend itself for a wider practice. The high death toll among multiple-births needs the
concerted effort of the different sectors. The sound and concrete findings of this study on infant
mortality and survivorship could help for proper policy formulation and health care planning. This
study will also be a base for further analytical study on the influencing factors of survival differentials.
Acknowledgement
Very many people have been involved in this community-based and longitudinal study. The list is
long, and we are indebted to many people for their help and support throughout. The study would have
been impossible without the dedication and tireless work of Dr. Fekadu Ayele, Dr. Sileshy T.Mariam,
Sileshy Demessie, Kassahun Melese and Mirgissa Kaba towards its proper management. We would
like to extend our greatest appreciation to the mothers of this cohort, who showed very high
cooperation and patience throughout our repeated follow-up visits. Our appreciation also goes to the
community leaders of the study kebeles. This project was funded by the Rockefeller Foundation,
UNICEF, Ministry of Health, and Jimma Institute of Health Sciences to all whom we are most grateful.
Reference
1. Central Statistical Authority. Population and Housing Census 1984: Analytical Report at National
Level. Addis Ababa: CSA, 1991.
2. Transitional Government of Ethiopia. National population policy of Ethiopia, April 1993 Addis
Ababa.
3. Central Statistical Authority. The 1990 National Family and Fertility Survey Report June 1993,
Addis Ababa.
4. Kosa J, Alpert J, Haggerty R. On the Reliability of Family Health Information. Soc Sci and Med
1967;1:165-181.
5. Hoekelman R, Kelly J, Zimmer A. The reliability of Maternal Recall. Clinical Pediatrics
1976;15(3):261-265.
6. Kroeger A. Health Interview Surveys in Developing Countries: A Review of Method and Results.
Int J Epidemiol 1983;1294:465-481.
7. Office of the National Committee for Central Planning Office for Western Ethiopia. Baseline
Survey for Jimma town. Jimma: RPOWE/UNICEF, 1986.
8. Green-Abate C. Changes in Birth-weight Distribution from 1973 to 1982 in Addis Ababa. Bull
WHO 1986;64(5):711-714.
9. Ahmed Z. The frequency of Multiple-births in Gondar Hospital North-Western Ethiopia. Ethiop
Med J 27:21-26.
Infant survivorship and occurrence of multiple-births 293
────────────────────────────────────────────────────────────
10. Asefa M, Drewett R, Hewison J,. An Ethiopian birth Cohort Study. Paediatric and Perinatal
Epidemiology 1996;10(4):443-462.
11. Norvsis M. SPSS/PC+ Advanced Statistics 4.0 Spss Inc. 1990. Chicago.
12. Behrman R et al., editors. Nelson Text book of paediatrics. London: Saunders Company, 1992.
294 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
13. Central Statistics Authority. The 1994 Population and Housing Census of Ethiopia: Results for
amahara Region. Vol 1, Part 1, Addis Ababa, April 1996.
14. Central Statistics Authority. The 1994 population and Housing Census of Ethiopia: Results for
Oromiya Region. Vol I, Part I, Addis Ababa, April 1996.
15. Gajanayake I. Infant mortality in Seri Lanka. J Biosoc Sci 1988;20:79-88.
16. Das Gupta M. Death clustering, Mothers Education and Determinants of Child Mortality in Rural
Punjab, India. Population Studies 1990;44:489-505.
17. Victora CG, Barros FC, Huttly SRA, Maria AB, Teixeira AMB and Vaughn JP. Early childhood
Mortality in a Brazilian Cohort: The role of birth-weight and socio-economic status. Int J Epidemiol
1992;21(5_:911-915.
18. Shamebo D, Sandstorm A, and Wall S. The Butajira Rural Health Project in Ethiopia:
Epidemiological Surveillance for Research and Intervention in Primary Health Care. Scand J Prim
Health Care 1992;10:199-205.
19. Tengio FU. A Survey of 43 Twin Deliveries at Bagmyou District Hospital, Tanzania. East Afr
Med J 1983;60:622-625.
20. Harrison KA, Rossiter CE. Multiple pregnancy. Brit J obstet. Gynaecol 1985;(Supp 5) 92:49-60.
21. Wilcox AJ and Russell IT. Birth-weight and perinatal Mortality: 111 Towards a New Method of
Analysis. Int. J Epidemiol 1986;15(2):188-196.
22. Lam DA and Miron JA. Seasonality of Births in Human Populations. Social Biology 1991;38(12):51-78.
Original article
Intestinal helminth infections in school children in Adurkay District 295
────────────────────────────────────────────────────────────
Intestinal helminth infections in school children in
Adarkay District, Northwest Ethiopia, with special
reference to Schistosomiasis mansoni
Leykun Jemaneh1
Abstract: In a survey carried out in five schools of five rural towns in Adarkay district in Northwest
Ethiopia, 519 children had their stool specimens examined for Schistosoma mansoni and other
intestinal helminth infections of man by the Kato thick smear technique. Infection due to S. mansoni
was the most prevalent (54.3%), ranging from 16.7% in Deb Bahir to 55.3% in Adarkay and Buya,
67.4% in Kerenejan and as high as 88.9% in Zarema, followed by Ascaris lumbricoides (43.0%) and
hookworm spp (23.3%). The least prevalent was Trichuris trichiura infection (11.8%). Triple, double,
and single infections were found in 49 (9.4%), 180 (34.7%) and 172 (33.1%) specimens respectively.
Most of the double infections were a combination of S. mansoni and A. lumbricoides (90=17.3%).
The highest prevalence for a single infection was recovered for S. mansoni (103=19.8%). Age specific
analysis of prevalence due to S. mansoni, the hookworms and A. lumbricoides showed the presence of
infection in all ages under consideration, but with no significant difference among the age groups.
Neither was there any significant difference in infection rates between the sexes. The average egg
counts were generally higher for S. mansoni and A. lumbricoides. Younger age groups appear to have
higher average egg counts, particularly for the hookworms. Sex was not related to egg output. The
high infection rate of intestinal helminth infestation observed in this study among school children
signifies the need for prompt intervention measures. [Ethiop. J. Health Dev. 1997;11(3):289-294]
Introduction
Infections due to intestinal helminths are most common in the developing world, particularly in
tropical regions where the environment, socio-economic status, human behaviour and cultural practices
favour transmission. Knowledge of the distribution and extent of helminthic diseases in these areas is
essential for prevention and control programmes.
In Ethiopia, intestinal helminthic infection is of major public health concern (1). Schistosomiasis
mansoni is endemic in many localities (2,3). The major soil-transmitted helminths, Trichuris trichiura,
Ascaris lumbricoides, and the hookworms are also frequently encountered in surveys (2-6).
There are well known causes of disease and contribute, among others, to the high proportion of
childhood morbidity (4).
Health strategy for the attainment of effective parasitic diseases control programmes demand
knowledge of the prevalence and distribution of the diseases and their changes in the course of time as
related to ecological, cultural, behavioural and other factors.
The present study is aimed at providing epidemiological information on schistosomiasis mansoni
______________________________________
1
From the Department of Microbiology and Parasitology, Faculty of Medicine, P.O. Box 9086, Addis
Ababa, Ethiopia
and other major intestinal helminthic parasites with respect to prevalence, species distribution and
intensity of infection among elementary schoolchildren in Adarkay Woreda (District) in Northwest
Ethiopia. It is anticipated that the information generated through this study will serve as a baseline
data for future health programme in the region.
296 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Methods
This study was undertaken in five randomly selected elementary schools (Deb Bahir, Zarema,
Adarkay, Kerenejan, Buya) out of the total of sixteen found in Adarkay Woreda (District), North
Gonder, Ethiopia, in 1995. Accessibility by a fourwheel drive was a factor in the selection of the
schools. The District has a population of approximately two hundred thousand people who are mainly
static and engaged in subsistence mixed agriculture. The altitude ranges from 1,000 to 2,400m above
sea level. The topography shows hills and plain land with springs, streams and rivers, which are often
the sources of water for domestic and other uses for the population. The meteorology, soil types and
other geographical features of the area are dealt with in detail by various authors (7-11 ). In the rural
areas and in most parts in the towns, defecation is mainly in open fields and in ditches.
The study population constituted children attending grades five and six in these five schools.
Children in these grades were assumed to show age of peak infection rate and hence expected to
indicate the situation of helminthiasis in the area. From the lists of children in these grades, lists of
prospective examinees (those who lived in the area for three or more years and had no treatment for
any intestinal ailment during the last one month) were drawn. Based on earlier reported prevalence
rates of 30-50% (4,12) for Gondar region and a confidence interval of 95% using appropriate statistical
formulae, 519 pupils were selected, using systematic sampling with a random start, to constitute the
sample populaton. The compliance rate was 100%. Their ages and sexes were registered.
For the parasitological diagnosis of intestinal helminths, stool samples were obtained from all 519
pupils and the Kato-Katz cellophane faecal thick smear technique (13) was employed using a 50 mg
template. Double Kato-Katz slides were prepared for each specimen. Examination for hookworm ova
under the microscope was performed immediately after the double Kato-Katz slide preparations
following which the slides were kept for at least one hour at ambient temperature to clear the faecal
material prior to examination for Ascaris, Trichuris, and Schistosoma eggs. Stool specimens were
considered positive when the characterstic eggs of S. mansoni, A. lumbricoides, T. trichiura, and
hookworms were noted in any one of the double Kato-Katz slides. The number of eggs of each species
was recorded and converted into the number of eggs per gram of faeces (EPG) in order to analyze
intensity of infection. The average number was taken when eggs were found on the two Kato slides.
Individuals positive for S. mansoni were treated on the spot with a single dose of praziquantel at 40
mg/kg body weight. Children positive for the other helminths were informed of the type of parasite
they harboured and advised to get treatment from the nearest health institution. Their names along
with the parasitological results were also communicated to respective home room teachers who were
requested to follow-up their treatment.
Results
In these five schools infection due to Schistosoma mansoni was the most prevalent (54.3%). The
least prevalent was Trichuris trichiura infection (11.8%) (Table 1).
Table 1: Prevalence (%) and intensity (average EPG) of intestinal helminth infections in children attending five elementary schools in Adarkay district Northwest Ethiopia, in 1995.
Values in parenthesis are average egg count per gram per positive individual (EPG)
School
No.
Examined
Adarkay
114
Buya
103
S. mansoni
A. lumbricoides
Hook worms
T. trichiura
#
63
(507.14)
%
55.3
#
49
(1579.59)
%
43.0
#
23
(563.04)
%
20.2
#
2
(200.00)
%
18
57
(584.21)
55.3
55
(1718.18)
53.4
31
(1319.35)
30.1
19
(621.05)
18.4
Intestinal helminth infections in school children in Adurkay District 297
────────────────────────────────────────────────────────────
Deb Bahir
120
20
(862.50)
16.7
62
(1179.84)
51.7
37
(687.84)
30.8
27
(618.52)
22.5
Kerenejan
92
62
(857.26)
67.4
32
34.8
25
27.2
13
14.1
Zarema
90
80
(926.25)
88.9
25
(1787.50)
27.8
5
(1808.00)
5.6
0
(773.08)
0.0
Of the four parasites studied, infection due to Schistosoma mansoni is more pronounced in all
communities except in Deb Bahir. The prevalence ranged from 16.7% in Deb Bahir to as high as
88.9% in Zarema. The prevalence of A. lumbricoides infection was highest in Buya (53.4%).
Hookworm infection prevalence was highest in Deb Bahir (30.8%) and Buya (30.1%). No T. trichiura
infection was registered in Zarema while the prevalence in others ranged from 1.8% in Adarkay to
22.5% in Deb Bahir.
Table 2: Prevalence (%) and intensity (average EPG) of intestinal helminth infection by age group in children attending five elementary schools in Adarkay District, Northwest
Ethiopia, in 1995. Values in parenthesis are average egg count per gram per positive individual (EPG).
Age Group
(yrs)
No
examined
5-9
82
10-14
361
15-19
76
Total
519
S. mansoni
#
47
(784.04)
203
(790.89)
32
(696.88)
282
(779.08)
A. lumbricoides
%
57.3
56.2
42.1
54.3
#
34
(1576.47)
158
(1559.18)
31
(1343.55)
223
(1531.84)
Hookworms
%
41.5
43.8
40.8
43.0
#
24
(1181.25)
83
(1095.78)
14
(628.57)
121
(1058.68)
T. trichiura
%
29.3
23.0
18.4
23.3
#
12
(729.17)
42
(632.93)
8
(531.25)
62
(638.52)
%
14.6
11.4
10.5
11.8
prevalence due to each of the four parasites showed the presence of infection in all ages with no
significant difference among the age groups. Nevertheless, in all but A. lumbricoides, a gradual decline
was noted in occurrence with increasing age (Table 2). Similarly no significant difference was seen in
infection rates of these intestinal helminths between the sexes (Table 3).
The average EPG due to S. mansoni was highest in Zarema, Deb Bahir and Kerenejan. With
hookworms and T. trichiura, the EPG tended to be higher in Buya and Kerenejan
(Table 1). Younger age groups appear to have higher average egg counts for the intestinal helminths,
particularly for the hookworms, but the difference is not significant (Table 2). Considering the sex, no
significant difference was observed in the intensity of infection among all helminths (Table 3).
Triple, double and single infections were found in 49 (9.4%), 180 (34.7%) and 172 (33.1%)
specimens, respectively as can be seen in Table 4. In addition, in two children there had quadruple
infections. Most of the double infections were a combination of S. mansoni and A. lumbricoides
(90=17.3%). The highest prevalence for a single infection was recorded for S. mansoni (103=19.8%).
Discussion
The outcome factors of interest in this study were positivity for the major intestinal helminths,
intensity and multiplicity of infection. The results of the study confirm the high prevalence of S.
mansoni and the soil-transmitted helminths.
In this study S. mansoni was present in 54.3% of the examined children with the prevalence
reaching close to 90% in Zarema. Although previous reports on S. mansoni are lacking for Buya, Deb
Bahir and Kerenejan, two decades ago McConnell and Armstrong (2), using the merthiolate-iodineformaline concentration technique (14) in Zarema and the merthiolate-iodine-formaline direct smear
method (15) in Adarkay, have reported the prevalence of S. mansoni to be 94% and 24% for Zarema
and Adarkay primary schoolchildren, respectively. It is vividly
Table 3: Prevalence (%) and intensity (MEPG) of intestinal helminth infection by sex in children attending five elementary
District, Northwest Ethiopia, in 1995.
Parasites
Male N = 272
Female N = 247
Both sexes N = 519
No
%
MEPG**
No.
MEPG
No.
%
S. mansoni
163
59.9
775.7
119
48.2
783.6
282
54.3
A. lumbricoides
118
43.4
1420.6
105
42.5
1656.7
223
43.0
schools in Adarkay
MEPG
889.1
1531.8
298 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
Hookworms
T. trichiura
69
28
25.4
10.3
1073.2
698.2
52
33
21.1
13.4
1039.4
587.8
121
61
23.3
11.8
1058.7
638.5
Table 4: Multiple intestinal helminthic infection in children attending five elementary school in Adarkay District, northwest Ethiopia, in 1995.
Multiplicity of infections
Specimens with 4 helminth parasites
SM, HW, AL and TT
Specimens with 3 helminth parasites
SM, AL and HW
SM. AL, and TT
SM, TT a nd HW
AL, TT and HW
Specimens with 2 helminth parasites
SM and AL
AL and HW
SM and HW
SM and TT
TT and HW
AL and TT
Specimens with 1 helminth parasite
SM
AL
HW
TT
Total number positive
Total number negative
Total number examined
0SM = S. mansoni
AL = A. lumbricoides
No.
Males
%
Females
No.
%
Both sexes
No.
%
2
0.7
0
0.0
2
0.4
14
7
5
1
5.1
2.6
1.8
0.4
7
11
1
3
2.8
4.4
0.4
1.3
21
18
6
4
4.0
3.5
1.2
0.8
53
20
16
6
3
1
19.5
7.3
5.9
2.2
1.1
0.4
37
15.0
15
5
1
8
15.0
6.1
6.1
2.0
0.4
3.2
90
35
31
11
4
9
17.3
6.7
6.0
2.1
0.8
1.7
43
24
10
4
184
63
247
17.4
9.7
4.0
1.6
74.5
25.5
100.0
103
44
18
7
403
116
519
19.8
8.5
3.5
1.3
77.6
22.4
100.0
60
22.1
20
7.3
8
2.9
3
1.1
219
80.5
53
19.5
272
100.0
HW = Hookworm
TT = T. trichiura
evident that when one compares the outcomes of the present study and that of McConnell and
Armstrong (2) the situation of S. mansoni has not changed in Zarema. On the other hand S. mansoni
prevalence has more than doubled in Adarkay. This increase can be explained by, among many other
factors, a more sensitive stool examination technique used in this study over the merthiolate-iodineformalin direct smear technique on the one hand, and by sampling variation on the other. Moreover a
finding of 55.3% prevalence for S. mansoni in Buya, an area well-removed from the main highway,
and 16.7% in Deb Bahir has been noted. These high prevalence figures of S. mansoni infections
registered in the different localities of the district are alarming.
The prevalence of the three soil-transmitted helminthic infections in this study can be compared with
a variety of other surveys carried out in schoolchildren or in communities in Gondar and other parts of
Ethiopia. Prevalence rates ranging from as low as 14% to as high as 85% for Gondar region (2,12, 1617) and from 4% to 72 % for other parts of Ethiopia (3-5) have been reported for A. lumbricoides.
Leykun Jemaneh and Shibru Tedla (18) found an overall hookworm infection rate of 18.1% among
schoolchildren in Gondar region. Trichuriasis, with infection rates ranging from 3% to 100%, has also
been reported from various localities (2,3).
There are a couple of published information that deal with the situation of these geo-helminths in
the study localities. Two decades ago McConnel and Armstrong (2) indicated prevalence rates of
82% and 52% for A. lumbricoides, 77% and 20% for hookworms and 47% and 0% for T. trichiura,
respectively, in schoolchildren for Zarema and Adarkay. On the other hand, in 1984 Leykun Jemaneh
and Shibru Tedla (18) noted hookworm prevalence rates of 32.4%, 39.2% and 21.4% for Zarema,
Adarkay, and Deb Bahir, respectively. The infection prevalence rates of A. lumbricoides, T. trichiura,
and the hookworms registered for the study localities in this study concur to the outcomes of the
previous investigations, while a much lower prevalence rate was obtained for A. lumbricoides
(27.8%), the hookworms (5.6%), and T. trichiura (0.0%) in Zarema. These differences in prevalence
rates may be due to differences, among others, in the method of examination used, on the one hand,
and to environmental factors on the other.
The intensity of helminthic infection in this study has been assessed indirectly by egg counts in
faeces. This method, although susceptible to errors of sampling due to periodicity of egg production
by female worms which may lead to uneven distribution of eggs in faeces (19) and density-dependent
constraints on fecundity that may mask the number of worms present (20), is still widely used as a
measure of intensity of intestinal helminthic infections. In this study, although the reported prevalence
of the geo-helminths is high, the intensity of infection as measured by eggs per gram of faeces is low.
It is comparable with the findings of a study carried out in a small farming village, near Lake Tana,
Gondar Region, Ethiopia (21). On the other hand, the intensity of S. mansoni infection was found to
be much higher than that reported for schoolchildren of Zeghie Junior Secondary School in Lake Tana,
an area endemic for S. mansoni (22). Such patterns of high prevalence of infection along with low
intensity of infestation and vice versa, have been observed elsewhere (23,24). Disparate exposure to
infection probably plays an important role in affecting intensity of infestation and helminth
distribution in different communities.
In localities where numerous kinds of intestinal helminths are found multi-parasitism is frequently
encountered. The most common combinations in some regions are infections which involve A.
lumbricoides, T. trichiura, and hookworms (3,25). Although a similar phenomenon is noted in the
present study, S. mansoni tended to appear more with A. lumbricoides. This is probably due to the
high prevalence of the two helminths in the area.
In conclusion, the high prevalence rate of intestinal infestation established in this study among
school children indicates the need for timely control measures. In many communities, as is the case in
this study, the majority of children aged between 5 and 15 years are not only infected with at least one
species of worm but they also tend to harbour the heaviest burdens (25). These helminth parasites are
identified with malnutrition as causes of compromised growth (26) and reduced physical and mental
fitness (27) that in turn affect the educational achievements of school children (28). Periodic treatment
measures targeted at school-aged children may be implemented within the capability of the local health
care system. Targeted chemotherapy with long term improvements of sanitation and incorporation of
health education in the schooling system should be exercised. Safe, low cost, single dose drugs are
now available to treat parasitic worms and the fact that many children aged 5 to 15 years assemble in
schools provides ample apportunity in which periodic treatments could be delivered.
Acknowledgement
The WHO/UNDP/WORLD BANK Special Programme for Research and Training in Tropical
Diseases (TDR) provided the funding for this study. The Institute of Pathobiology and the Gondar
College of Medical Sciences Addis Ababa University provided the parasitological staff and the
logistical support for the smooth running of the study. The administrators, educational officers, school
directors, teachers, and students of the study areas are highly acknoweldged for their utmost
cooperation.
References
1. Ministry of Health (1991). Comprehensive Health Service Directory. Planning and Programming
Department, Addis Ababa.
2. McConnell E, Armstrong JC. Intestinal parasitism in fifty communities on the Central Plateau of
Ethiopia. Ethiop Med J. 1976;14:159-168.
3. Hailu B, Berhanu E, Shibru T. Intestinal helminthic infections in the southern rift valley of
Ethiopia with special reference to schistosomiasis. East African Medical Journal 1994;71:447-452.
4. Shibru T. Intestinal helminthiasis of man in Ethiopia. Helminthologia 1986;23:43-48.
5. Shibru T, Leykun J. Distribution of Ancylostoma duodenale and Necator americanus in Ethiopia.
Ethiop Med J. 1985;23:149-158.
6. Shibru T, Teklemariam A. Ascariasis distribution in Ethiopia. Ethiop Med J. 1986;79-86.
7. Polderman AM. The transmission of intestinal schistosomiasis in Begemedir province, Ethiopia.
Dissertation, Leyden University, the Netherlands, 193pp, 1974.
8. Shibru T and Leykun J. Distribution of Ancylsotoma duodenale and Necator americanus in
Ethiopia. Ethiop Med J. 1985;23:149-158.
9. Schaller KF, Kuls W. Ethiopia: Geomedical Monograph Series. 1972;3:109-126.
10. Ethiopian Meteorological Service. Meteorological Maps of Ethiopia. Addis Ababa, 1979.
11. Galperin G. Ethiopia: Population, Resources, Economy. Progress Publishers, Moscow, 1981.
12. Zein AZ and Mekonnen A. The prevalence of intestinal parasites among farming cooperatives,
Gondar region, North-western Ethiopia .Ethiop Med J. 1985;23:159-167.
13. World Health Organization. Prevention and control of intestinal parasitic infections. WHO
technical report series 1987, No 749.
14. Blagg W, Schloegel EL, Mansour NS, Khalaf GI. A new concentration technique for the
demonstration of protozoa and helminth eggs in faeces. Am J Trop Med Hyg. 1955;4:23-31.
15. Dunn FL. The MIF direct smear as an epidemiological tool, with special reference to counting
helminth eggs. Bull. WHO 1968;39:439-449.
16. Wang L. Helminthiasis in Begemedir and Siemen province. Ethiop Med J. 1965;4:19-26.
17. Kloos H, Bedri A, Addus A. Intestinal parasitism in three resettlement farms in
western Ethiopia. Ethiop J Health Dev. 1991;5:51-56.
18. Leykun J, Shibru T. The distribution of
Necator americanus and Ancylostoma duodenale in school populations, Gojam and Gondar
administrative regions. Ethiop Med J. 1984;22: 87-92.
19. Croll NA, Anderson RM, Gyorkos TW, Ghadirian EA. The population biology and control of
Ascaris lumbricoides in a rural community in Iran. Trans Roy Soc Trop Med Hyg 1982;76:187-97.
20. Hall A. Intestinal Helminths of man: The interpretation of egg counts. Parasitology 1982;85:605613.
21. Melake BD, Wondwossen H, Tesfaye W, Elias GK, Sissay Y, Tariku A, Tibabu D. Intensity of
intestinal parasite infestation in a small farming village, near Lake Tana, Ethiopia. Ethiop J Health
Dev. 1993;7:27-31.
22. Berhanu E, Shibru T. Intestinal helminth infections at Zeghie, Ethiopia, with emphasis on
schistosomiasis mansoni. Ethiop J Health Dev. 1993;7:21-26.
23. Higgins AD, Jenkins JD, Lilians Kuriawan P, Harun S, Sundraiju-wone S. Human intestinal
parasitism in three areas of Indonesia: A survey. Ann Trop Med Parasit. 1984;78:637-648.
24. Robertson LJ, Crompton DWT, Walters DE, Nesheim MC, Sanjur D, Walsh EA. Soil-transmitted
helminth infections in school children from Cocle Province, Republic of Panama. Parasitology
1989;99:287-292.
25. Tilahun W, Tsehay A, Tareke S. Intestinal parasitism among the student population of the WonjiShoa Sugar Estate. Ethiop Med J. 1990;4:45.
26. Bundy DAP, Hall A, Medley GF, Savioli L. Evaluating measures to control intestinal parasitic
infections. World Health Statistics Quarterly 1992;45:168-179.
27. Hall A. Intestinal parasitic worms and the growth of children. Trans Roy Trop Med Hyg
1993;87:241-242.
28. Crompton DWT. Ascariasis and childhood malnutrition. Tran Roy Soc Trop Med Hyg
1992;86:577-579.
Review article
Evaluation of Hutchinson’s sign in HIV associated
herpes zoster ophthalmicus
Samson Bayu1 and Wondu Alemayehu1
Abstract:A prospective study of 100 serial HIV infected herpes zoster ophthalmicus patients, was
done from December 1993 to June 1995, to determine the correlation between Hutchinson’s sign and
ocular involvement. A mean age of 35.2 ±10.7 years and a male to female ratio of 2.22 to 1 was
found. Ocular complications resulted in 78% of the patients. The sensitivity, specificity, accuracy,
positive and negative predictive value of the sign in detecting ocular involvement is 74%, 68%, 73%,
89% and 43% respectively. An H.Z.O. patient with Hutchinson’s sign is 2.33 times likely to have
ocular involvement. However in areas where HIV associated H.Z.O. is highly prevalent, Hutchinson’s
sign is not a reliable predictor of ocular involvement; therefore, it should be used with prudence. A
future study on adequate number of patients is also recommended. [Ethiop. J. Health Dev.
1997;11(3):295-298]
Introduction
Herpes zoster ophthalmicus (H.Z.O.) is a maculopapular rash on the dermatomal distribution of the
ophthalmic division of the trigeminal nerve which is said to be due to reactivation of Varicella-zoster
virus (1). H.Z.O. was the 2nd common cause of attendance at a neuro-ophthalmic clinic in an Ethiopian
tertiary eye care centre (2). It is recognized as an early clinical marker of HIV infection (3, 4). A
prospective study on Ethiopian H.Z.O. patients revealed 95.3% seropositivity to HIV infections (5).
H.Z.O. affects all layers of the eye. Ocular complications occur in 50-89 % of patients (3, 5, 6, 7).
The incidence and severity of ocular complications and post herpetic neuralgia is markedly increased
in HIV infected patients with H.Z.O. (3, 4, 5).
Hutchinson in 1885 observed that ocular involvement was much more common in patients who had
zoster involvement of the nasociliary branch which is manifested by cutaneous involvement on the side
of the tip of the nose (8). Hutchinson's sign is convenient, and applicable early during the eruption
phase.
A recent study, in 1987, also showed statistical association of nasociliary nerve involvement with
subsequent ocular disease (7). Oral acyclovir, especially if given in the first 72 hours, protects against
ocular complications (9).
The purpose of this study is to evaluate Hutchinson's sign as a diagnostic test for screening of ocular
involvement in HIV associated H.Z.O.. Early detection of H.Z.O. patients at risk for the development
of ocular involvement is important, so that antiviral (acyclovir) would be used early in the course of
the illness. The efficient utilization of antiviral is of paramount importance in developing countries
where its use is limited due to unavailability of the drug and its costliness.
_____________________________________________
From the Department of Ophthalmology, Medical Faculty , Addis Ababa University, Addis Ababa,
Ethiopia.
Methods
A prospective case series study of 100 consecutive Ethiopian H.Z.O. patients attending the eye
department of Menelik II hospital, from December 1993 to June 1995, who are seropositive to HIV
infection by double ELISA technique using first generation Welcozyme kits, which were replaced by
second generation kits (Welcome diagnostics, Dartford, England) were studied. Serologic
examination to HIV infection was done following verbal informed consent. Repeat ELISA positive
patients were considered to have HIV infection. This is in accord with 64% and 94.2% specificity of
302 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
the ELISA test in the low risk Ethiopian population for the first and second generation Welcozyme kits
respectively (10). Counselling of the patients was done based on the serologic result. The patients were
examined by the two authors separately. Age, sex, right or left dermatomal distribution, presence or
absence of ocular involvement and Hutchinson's sign was recorded in each patient. Patients were
followed for a minimum period of three months.
H.Z.O. was defined as a maculopapular rash which leaves a scar on the dermatomal distribution of
the ophthalmic division of the trigeminal nerve.
Hutchinson's sign was defined as cutaneous involvement of zoster rash or scar on the side of the tip
of the nose.
Ocular involvement was defined as abnormalities in the globe due to H.Z.O. which is detected by
detailed ocular examination using slit lamp biomicroscope and ophthalmoscope. It was done by a
senior ophthalmologist and taken as a standard for evaluation.
The sensitivity, specificity, accuracy, positive and negative predictive value of Hutchinson's sign as
a marker of ocular involvement is calculated. The pre test likelihood (ocular complications
prevalence), post test likelihood and the likelihood ratio when Hutchinson's sign is present or absent is
analyzed as reviewed in literatures (11).
Results
The age and sex distribution of the patients are as seen in Table 1. The age ranges from 18 to 70
years with a mean age of 35.2 + 10.7 years. Eighty four percent of the patients were aged 45 years or
less. Males constitute 69 % of the patients (male: female is 2.2 to 1). Right eye was involved in 48
patients and left eye in 52 patients. There was no bilateral involvement or systemic zoster seen.
Table 1: Age and Sex distribution of HIV infected H.Z.O. patients
Age
Male
Female
Total
15-19
-1
1
20-24
7
8
15
25-29
10
8
18
30-34
14
3
17
35-39
19
4
23
40-44
5
2
7
45-49
5
2
7
50-54
4
_
4
55-59
3
2
5
> 60
2
1
3
Total
69
31
100
As seen in Table 2, the prevalence of ocular involvement is 78%. Corneal complications, anterior
uveitis, secondary glaucoma and scleral involvement constitute the majority of ocular involvement.
The sensitivity, specificity, accuracy, positive and negative predictive value of Hutchinson's sign in
detecting ocular involvement is 74%, 68%, 73%, 89%, and 43% respectively.
Evaluation of Hutchinson’s sign in HIV associated herpes zoster 303
────────────────────────────────────────────────────────────
Table 2: The correlation of Hutchinson is sign and ocular involvement
ocular involvemenrt
Present
Hutchinson' s
present
58%
Absent
20%
Total
78%
Hutchinson' s
58%
Sensitivity= 74.35 %
Post test likelihood
Specificity= 68.18%
if (+)= 89 %
Accuracy= 73 %
if (-)= 57 %
Pre test likelihood= 78%
Likelihood ratio= 2.33
Positive predictive value= 89%
Negative predictive value= 43%
Absent
7%
15%
22%
7%
Discussion
In the pre HIV period, H.Z.O. was said to be the disease of the elderly and the aged are particularly
susceptible for its development (1). Various studies showed the mean age to be in the sixties (6,7, 12).
After the HIV pandemic, the pattern changed and HIV related H.Z.O. developed in the younger age
groups with a mean age ranging from 28-34 years (3, 4). The mean age in our series is 35 years which
is in agreement with the above studies and nearer to the mean age of Ethiopian AIDS patients of 30
years (13).
The presence or absence of gender predilection for the acquisition of H.Z.O. is still in debate. Some
studies have shown female preponderance (6, 7, 12) and others contend that there is no sexual
predilection (1). These studies were undertaken prior to the HIV era. The male to female ratio in our
series is 2.2 to 1. A recent report of AIDS cases in Ethiopia showed a male to female ratio of 1.6 to 1
(13). This might be due to a difference in utilization of health institution favouring males; as is also
seen in other neuro-opthalmic patients (2).
Studies in pre HIV era showed ocular complications ranging from 50-71% (6, 7, 12). The
prevalence of ocular complications seen in our HIV infected H.Z.O. patients is markedly increased
(78%). An African study in HIV infected young individuals showed 89% incidence of ocular
complications (3) which is slightly greater than in our series.
The positive predictive value of Hutchinson's sign reported ranges from 76-85% (7, 14). In our
series of patients, the likelihood of ocular complications in patients with the sign rise from 78% to
89%. In its absence it falls to 57 %. Thus, an H.Z.O. patient with the sign is 2.33 times likely to have
ocular involvement than without. An absence of Hutchinson's sign is 2.66 times likely to be seen in
patients without ocular involvement.
Hutchinson's sign can be observed easily by all health workers early in the course of the illness.
Based on our finding, it gives a clue to the increased risk of subsequent ocular involvement. On the
contrary, absence of Hutchinson's sign occur in one-fourth of the patients with ocular involvement. In
accord with the reported literature, our study substantiates that Hutchinson's sign alone is not a reliable
predictor of ocular involvement (8) and severe ocular complications may occur even in patients with
slight rash anywhere in the forehead (10, 15). This would worsen in HIV related H.Z.O. Thus, in
areas where HIV associated H.Z.O is highly prevalent, Hutchinson's sign should be used with caution.
A future study on a large number of patients is also recommended.
Acknowledgement
We sincerely thank Ato Fikre Enquoselassie of the Department of Community Health, A.A.U., for
reviewing the manuscript.
304 Ethiop.J.Health Dev.
────────────────────────────────────────────────────────────
References
1. Karabassi M, Raizman MB, Schuman JS. Herpes zoster ophthalmicus. Surv ophthalmol
1992;36:395-410.
2. Samson B, Wondu A. Pattern of neuro-ophthalmic disorders in a tertiary eye care centre. In Addis
Ababa Ethiop Med J. 1997;35:43-51.
3. Kestelyn P, Stevens AM, Bakkers E, Rouvray D, De-Perre UP. Severe herpes zoster ophthalmicus
in young African adults: a marker for HTLV-III seropositivity. Br J Ophthalmol 1987;71:806-809.
4. Sandor EV, Millman A, Croxson TS, Mildvan D. Herpes zoster ophthalmicus in patients at risk for
the acquired immune deficiency syndrome (AIDS). Am J Ophthalmol 1986;101:153-5.
5. Samson B, Wondu A. Clinical profile of herpes zoster ophthalmicus in Ethiopians. Clin Infect Dis
1997;24:125601260.
6. Womack LW, Liesegang T. Complications of herpes zoster ophthalmicus. Arch Ophthalmol
1983;101:42-45.
7. Harding S, Lepton J, Wells CD. Natural history of herpes zoster ophthalmicus: predictors of post
herpetic neuralgia and ocular involvement. Br J Ophthalmol 1987;71:353-358.
8. De-Luise V, Wilson II FM. Varicella and Herpes zoster ophthalmicus. In: Duane T, Jaeger EA
editors. Clinical ophthalmology, Philadelphia: J.B. Lippincott, 1988: vol.4 chap.20;p 1-20.
9. Cobo LM, Foulks GN, Liesegang T, Lass J, Sutpin JE, Wilhelmus K, et.al. Oral acyclovir in the
therapy of acute herpes zoster ophthalmicus. Ophthalmology 1986;93:763-770.
10. Tsehaynesh M and Debrework Z. Comparison of two generations of Wellcozyme kits. Ethiop J
Health Dev 1990;4:207-211.
11. The Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences
Centre. How to read clinical journals: II. To learn about a diagnostic test. Can Med Assoc J
1981;124:703-710.
12. Marsh RJ, Cooper M. Ophthalmic herpes zoster. EYE 1993;7(3):350-370.
13. National HIV/AIDS Update. AIDS/STD control program. Ministry of Health, Addis Ababa.
Ethiop J Health Dev 1995;9:63-66.
14. Liesegang TJ. Corneal complications from herpes zoster ophthalmicus. Ophthalmology
1985;92:316-324.
15. Kanski JJ. Clinical ophthalmology. 2nd ed. London, Butterworth-Heinemann, p.101-105, 1989.