Prostate Cancer: A New Brunswick Tale*
Peter Sephton
School of Business
Queen’s University
Kingston, Ontario
K7L 3N6
Psephton@business.queensu.ca
613 533 3013
613 533 6589
*
VOICE
FAX
This work is based in part on data provided by the
New Brunswick Department of Health and Wellness.
The interpretation and conclusions contained
herein do not necessarily represent those of the
Government of New Brunswick or the New
Brunswick Department of Health and Wellness.
I thank seminar participants at the Radiation Oncology
Research Group at the Kingston General Hospital and Patti
Groome and Julian Barling for very helpful comments. Financial
assistance from the Canadian Cancer Society Regional Research
Development Program and the Dean’s Research Fund is
gratefully acknowledged.
Abstract
From 1988 until 1997 there were 947 deaths from
prostate cancer in New Brunswick and 4382 new cases
diagnosed. This paper provides a descriptive view of the disease
in New Brunswick and examines i) whether the actual number
of deaths was different from what could have been expected for
each health region, and ii) whether there was a link between
occupation and death by prostate cancer. The results suggest
that two health regions, and New Brunswick as a whole,
experienced more deaths than expected. They also indicate
there were occupations in which the number of prostate cancer
deaths was significantly different from their expected values.
2
I
Introduction
From 1988 until 1997, 28 percent of all male deaths in New Brunswick
were due to cancer. Of these 8540 deaths, 2997 were caused by lung cancer,
with 947 attributed to prostate cancer. During this period there were 4382
new prostate cancer cases diagnosed, the vast majority in men aged 55-74.
Treatment regimens depend on a number of factors including age, life
expectancy, stage of tumour, and the individual’s preferences over palliative
care rather than attempted curation. Options include radiation, surgery,
chemotherapy, watchful waiting, or combinations thereof, all of which impose
direct and indirect costs on those affected.
The purpose of this note is to provide a descriptive view of mortality
related to the disease in New Brunswick. Given mortality rates, did New
Brunswick experience more deaths than could have been expected? What
was the regional distribution of mortality, and what was the distribution by
age at death? Is it true that the majority of prostate cancer deaths occur in
the elderly? While some researchers attribute prostate cancer to genetic
predisposition1, there are other factors such as diet, sexual history, and
occupation that may play a role. Was there a link between prostate cancer
mortality and occupation in New Brunswick? This latter question is of
particular interest given that occupations with exposure to polycyclic
aromatic hydrocarbons - such as farming, construction, and others – have
Much of the literature confuses absolute risk: the chance of developing the disease over a
particular time period, with relative risk: the ratio of the absolute risk of those containing a
certain risk factor to the absolute risk of those without the risk factor. Having a first line
relative (brother, father) with prostate cancer is thought to double relative risk. Having two
first line relatives diagnosed with prostate cancer quadruples relative risk. Persons of colour
have a higher absolute risk than Caucasians. See Gallagher and Fleshner (1998) and
Vetrosky and White (1998) for more on the clinical perspectives of the disease, and Dupont
and Plummer (1996) on the differences between absolute and relative risk.
1
3
been identified in previous studies as having a higher incidence of the
disease.2
As Gallagher and Fleshner (1998) argue, even if there are occupations
in which incidence is higher than the norm, the implementation of
preventative measures would face a number of challenges, particularly given
that the etiology of the disease is unknown. This notwithstanding, it would
be helpful for the authorities to understand whether there are specific needs
for prostate cancer related services in certain areas of the province. For
example, if farmers experienced a higher than expected mortality rate, the
government might be expected to provide enhanced services in rural areas to
accommodate the needs of the farming community.3 Failing this, the
authorities may want to encourage physicians to follow certain patient
groups more closely than others, with perhaps a more thorough screening
program or protocol than offered to the general population. All have
implications over resource allocation.4
II
Age and Regional Patterns
Vital Statistics data from the Department of Health and Wellness was
used in the present analysis.5 It covered all male deaths in New Brunswick
from 1988 until the end of 1997. Information on the health region of the
deceased, their occupation, date of birth, and cause of death was available.
During this period there were over 30,000 male deaths, 8540 the result of
See Aronson et al (1996) and Seidler et al (1998) for evidence that exposure to diesel fuel or
fumes may be associated with the development of prostate cancer.
3 This presumes that the higher than expected mortality rate is associated with a higher
incidence of prostate cancer in farmers, which is outside the scope of the present analysis,
and may not be the case.
4 Krahn et al (1999) estimate that the costs of PSA screening programs for all eligible men in
Canada could be as high as $412 million by 2005.
5 I thank Chris Heissner and John Boyne for providing this data.
2
4
cancer. There were 947 deaths in which prostate cancer was cited as the
cause of death in the vital statistics records.
A cursory search of the term “prostate cancer” in internet search
engines provides thousands of hits. Public awareness has risen substantially
over the past decade, partly as a result of a number of public figures
admitting they had the disease. Much of the literature appears to suggest
that the majority of prostate cancers are slow-growing and that in all but a
small number of cases, a diagnosis of prostate cancer, if made early enough,
allows one to effectively treat the disease. The survival rate is relatively high,
suggesting that those who do die as a result of prostate cancer are relatively
old.
Was this the case in New Brunswick? Table 1 contains an agedistribution of deaths from prostate cancer from 1989 until 1997.6 Of the 859
deaths, over 80 percent occurred in patients over 70 years of age. Nearly half
of all deaths were in males over 80.
What of those individuals who were diagnosed with prostate cancer?
Did a majority die from prostate cancer, or was there some other cause of
death? Table 2 indicates that 39 percent of those who had been diagnosed
with prostate cancer actually died as a result of the disease. A majority died
from other causes.
These statistics appear to validate the conventional wisdom: over 80
percent of prostate cancer deaths occurred in those aged 70 and over, with
nearly half of all deaths occurring at 80 years of age and over. Men
Subsequent analyses examine the period from 1989 until 1997. This is because a consistent
measure of the male New Brunswick population by health regions was available only back to
1989.
6
5
diagnosed with prostate cancer were more likely to have died as a result of
something other than the disease.7
New Brunswick is divided into a number of different health regions.
Figure 1 describes the geographic areas associated with each region. From a
health policy standpoint, it is interesting to determine whether the actual
number of deaths within each region was as might have been expected. Were
there “pockets” of prostate cancer, or were there regions in which the number
of deaths from prostate cancer was lower than expected? If differences
emerge, the Department of Health and Wellness might want to look at those
regions to determine if there are structural differences in the management of
the disease. For example, differences in clinical practice across the regions or
in access to capital-intensive facilities (such as radiological suites not
available in each health region) may affect mortality. While prostate cancer
mortality appears to have fallen in the late 1990s, Perron et al (2002) suggest
that this cannot be attributed to PSA (prostate specific antigen) screening
programs, so there may be other factors the Department of Health and
Wellness can identify as having a protective influence.
Figure 2 presents a regional breakdown of the number of deaths along
with the number that would have been expected to occur given the male
population within the health region and the mortality rates associated with
prostate cancer over this period.8 For the province as a whole there was a
statistically significant difference between actual and expected mortality,
with more deaths than expected. Health Regions 1 and 6 contributed to this
It is worth noting that these results depend critically on the accuracy of the reporting
mechanisms in place. If vital statistics data was not collected and coded properly, or if there
were variations across regions in the manner in which reports were constructed, then the
findings reported herein may not be truly reflective of the disease in New Brunswick.
8 The expected number of deaths is constructed using age-specific mortality rates,
standardized to the New Brunswick male population. Age-specific mortality rates were
obtained from Health Canada (http://cythera.ic.gc.ca/dsol/cancer/d_age_e.html).
7
6
finding. This suggests the Department of Health and Wellness might benefit
from identifying how these two Regions differ from the rest of the province,
either in terms of the characteristics of the underlying population (eg.,
smoking prevalence) or in clinical practice and treatment regimens.
III
Occupational Factors
A number of studies have found evidence supporting an association
between occupation and prostate cancer. Exposure to diesel fuel or fumes has
been linked to prostate cancer. Farmers and agricultural workers, airline
pilots and some chemical workers have experienced higher than expected
incidence and mortality rates. Was there a link between occupation and
death from prostate cancer in New Brunswick?
Vital statistics death records include the occupation of the deceased.
To the extent that this information correctly identifies the occupation of the
individual (many people work in several different industries over their
working lives), it may be possible to determine whether there were more
deaths than expected by occupation. To calculate the expected number of
deaths, data from the Labour Force Survey was used to establish the
distribution of male workers by occupation in New Brunswick. The data from
the 1996 Census was used to allocate workers to occupational categories that
matched as closely as possible the occupational groupings in the vital
statistics records. This provided the proportion of New Brunswick males
working in each occupation, which was assumed to represent the distribution
of employment over the entire sample period. Given age-specific mortality
rates standardized to the New Brunswick population, it is possible to
determine the expected number of deaths due to prostate cancer in each
region for each occupation.
7
Table 4 presents the actual and expected number of prostate cancer
deaths over this nine-year period. For those death records where occupation
category was missing or uncoded, the cases were deleted from the analysis
(the expected number of deaths also excluded those workers whose
occupations were not classified). This reduced the number of actual deaths to
727 cases (there were 132 actual deaths due to prostate cancer where
occupational class was not listed or was unclassified). Several results are
striking. Occupations in medicine, sales, machining, product fabricating,
transportation equipment, and in material handling experienced mortality
from prostate cancer as was expected. Many occupations had substantially
fewer deaths than expected, notably those in management, the natural and
social sciences, teaching, clerical work, services, processing, and artistic
fields. Evidence supportive of previous studies into occupational factors and
prostate cancer mortality is reported here, with farmers, fishers, foresters,
those in mining and construction, and religion facing higher mortality than
expected.
Table 5 presents Standardized Mortality Ratios and the associated
probability values, and demonstrates that these results are statistically
significant. The occupations associated with a higher than expected number
of deaths are very similar to those identified through case-referent studies
and the analysis of more finely refined patient-level data, such as Aronson et
al (1996) and Seidler et al (1998). The empirics support the view that certain
occupations and prostate cancer mortality – either higher or lower than
expected - were associated in New Brunswick during this period.
8
IV
Final Remarks
The purpose of this note was to provide a descriptive view of prostate
cancer mortality in New Brunswick. Nearly three percent of all male deaths
during this period were due to prostate cancer, with most occurring in the
population aged 70 and over. For those who had been diagnosed with
prostate cancer and subsequently died, the vast majority died from
something other than prostate cancer.
Vital statistics data and information on employment in the province
were used to construct estimates of the expected number of deaths from
prostate cancer by region and by occupation. For the province as a whole,
there were more deaths than expected, mainly in Health Regions 1 and 6.
This suggests the Department of Health and Wellness might benefit from
determining how these Health Regions differ from the rest of the province,
either in the clinical management of the disease or in the underlying
characteristics of the population.
Support for an association between occupation and death by prostate
cancer was found for several job categories. Some occupations experienced
fewer deaths than expected. Farmers, foresters, construction workers, and
miners experienced more deaths than expected. Given that employment in
these areas is high in New Brunswick, there are important implications for
health care resource management in the province. In addition, to the extent
that farmers, foresters, construction workers and miners have relatively little
control over their work environments than workers in other occupations,
these findings argue for greater emphasis on workplace safety and an
enhanced protective role for governmental agencies.
9
Health policymakers should be interested in the empirical results. The
age distribution of mortality appears to indicate that existing protocols offer
an effective approach to disease management. There appears to be evidence
that Regions 1 and 6 experienced a larger number of deaths than could have
been expected, and suggests there may be differences in the underlying
characteristics of the population or in clinical practice that distinguish these
regions from the rest of the province. The appearance of an association
between death from prostate cancer and employment in certain industries
requires additional study to determine both the exact transmission
mechanism as well as protective measures which can reduce exposure to
carcinogens. Physicians might be expected to provide additional screening to
men working in occupations identified as being high risk.
Further study in this area should focus on patient-level data to
determine whether there are unique associations between occupation,
location, and even socio-economic factors that contribute to prostate cancer
mortality in New Brunswick.
10
References
Aronson, K., J. Siemiatycki, R. Dewar, and M. Gerin, (1996), “Occupational
Risk Factors for Prostate Cancer: Results from a Case-Control Study in
Montreal”, American Journal of Epidemiology 143, 363-373.
Dupont, W., and W. Plummer (1996), “Understanding the Relationship Between Relative
and Absolute Risk”. Cancer 77, 2193-2199.
Gallagher, R., and N. Fleshner (1998), “Prostate Cancer: 3. Individual Risk
Factors”, Canadian Medical Association Journal 159 (7), 807-813.
Krahn, M., Coombs, A., and I. Levy (1999), “Current and Projected Annual
Direct Costs of Screening Asymptomatic Men for Prostate Cancer Using
Prostate-Specific Antigen”, Canadian Medical Association Journal 160 (1),
49-57.
Perron, L., L. Moore, I. Bairati, P-M. Bernard, and F. Meyer (2002), “PSA
Screening and Prostate Cancer Mortality”, Canadian Medical Association
Journal 166 (5), 586-591.
Seidler, A., Heiskel, H., R. Bickeboller, and G. Elsner (1998), “Association
Between Diesel Exposure at Work and Prostate Cancer”, Scandinavian
Journal of Work, Environment, and Health 24(6), 486-494.
Vetrosky, D., and G. White (1998), “Prostate Cancer: Clinical Perspectives”,
AAOHN Journal, 46 (9), 434-440.
11
Table 1: Actual Mortality by Age Group, 1989-1997
Age Groups
Number
%
40-49
8/859
1
50-59
21/859
2.5
60-69
116/859
13.5
70-79
300/859
35
80+
414/859
48
12
Table 2: Of Those Who Were Diagnosed with Prostate Cancer and Died in
New Brunswick: Underlying Cause of Death
ICD9
185
410
436
486
496
1539
1629
1991
2500
3310
4140
4149
4275
4280
Description
Percentage %
Malignant neoplasm of prostate
Acute myocardial infarction
Acute, but ill-defined, cerebrovascular disease
Pneumonia, organism unspecified
Chronic airway obstruction, not elsewhere classify
Colon, unspecified
Bronchus and lung, unspecified
Malignant neoplasm without specification of site
Diabetes mellitus
Other cerebral degenerations
Coronary atherosclerosis
Chronic ischemic heart disease, unspecified
Cardiac arrest
Heart failure
39
6
4
1
5
1
4
2
2
1
2
5
3
1
Subtotal
76
Other
24
Total
100
ICD9 refers to the 9 th edition of the International Classification of Diseases
coding system.
13
Figure 1:
New Brunswick Health Regions
14
Figure 2: Actual and Expected Prostate Cancer Deaths
Actual and Expected Prostate Cancer Deaths
300
250
200
Actual
150
Expected
100
50
0
1
2
3
4
5
6
7
Health Region
Table 3: Standardized Mortality Ratios with Probability Values
Health Region
Standardized Mortality
Ratio
Probability
Value
1
2
3
4
5
6
7
1.26
0.97
0.98
0.96
0.95
1.25
1.12
0
0.67
0.76
0.80
0.76
0.02
0.40
Total
1.08
0.02
Note: The Standardized Mortality Ratio is calculated as the ratio of the
actual to the expected number of deaths. Probability values associated with a
test that examines null hypothesis that the expected values equal the actual
value. There were 859 deaths due to prostate cancer from 1989 to 1997.
15
Table 4: Actual vs Expected Deaths by Occupation
Occupation
Managerial, administrative and related occupations
Proportion of
Male
Employment (%) Actual
Expected
12.05
60
95
Occupations in natural sciences, engineering and mathematics
5.95
16
47
Occupations in social sciences and related fields
1.67
4
13
0.6
11
5
3.01
14
24
Occupations in medicine and health
1.7
10
13
Artistic, literary, recreational and related occupations
1.6
1
13
Clerical and related occupations
5.23
26
41
Sales occupations
7.92
51
62
12.02
55
95
3.83
90
30
2.4
33
19
Forestry and logging occupations
2.99
41
24
Mining and quarrying including oil and gas field occupations
0.67
10
5
Processing occupations
5.65
23
44
Machining and related occupations
2.81
16
22
Product fabricating, assembling and repairing occupations
6.66
58
52
Construction trades and occupations
9.38
126
74
Transport equipment operating occupations
6.59
48
52
Material handling and related occupations
1.92
17
15
Other crafts and equipment operating occupations
5.35
17
42
Occupations in religion
Teaching and related occupations
Service occupations
Farming, horticultural and animal husbandry occupations
Fishing, trapping and related occupations
Note: There were 859 deaths due to prostate cancer from 1989 to 1997. Of
those, 132 had occupation classes that were not listed or listed as
unclassified. These observations were dropped from the occupational
analysis.
16
Table 5: Standardized Mortality Ratios with Probability Values
Occupation
Managerial, administrative and related occupations
Occupations in natural sciences, engineering and mathematics
Occupations in social sciences and related fields
Occupations in religion
Teaching and related occupations
Occupations in medicine and health
Artistic, literary, recreational and reelated occupations
Clerical and related occupations
Sales occupations
Service occupations
Farming, horticultural and animal husbandry occupations
Fishing, trapping and related occupations
Forestry and logging occupations
Mining and quarrying including oil and gas field occupations
Processing occupations
Machinig and related occupations
Product fabricating, assembling and repairing occupations
Construction trades and occupations
Transport equipment operating occupations
Material handling and related occupations
Other crafts and equipment operating occupations
Standardized Mortality
Ratio
Probability
Value
0.63
0.34
0.31
2.2
0.58
0.77
0.08
0.63
0.82
0.58
3
1.74
1.71
2
0.52
0.73
1.12
1.7
0.92
1.13
0.4
Note: The Standardized Mortality Ratio is constructed as the ratio of the
actual number of deaths to the expected number of deaths. Probability values
associated with a test that examines null hypothesis that the expected values
equal the actual value. There were 859 deaths due to prostate cancer from
1989 to 1997. Of those, 132 had occupation classes that were not listed or
listed as unclassified. These observations were dropped from the occupational
analysis.
17
0.0001
0
0.004
0.02
0.03
0.41
0
0.01
0.16
0
0
0
0
0.04
0
0.09
0.4
0
0.59
0.58
0