THE COST-EFFECTIVENESS OF PNEUMOCOCCAL VACCINATION FOR OLDER
PEOPLE: A STUDY IN FIVE WESTERN EUROPEAN COUNTRIES
Ament A1, Baltussen R1,2, Duru G3, Rigaud-Bully C3, de Graeve D4, Orqtvist A5, Jönsson B6,
Verhaegen J7, Gaillat J8, Christie P9, Salazar Cifre A10, Vivas D11, Loiseau C12, Fedson DS12
1
Maastricht University, Maastricht, The Netherlands
2
3
Heidelberg University, Heidelberg, Germany
Laboratoire d'Analyse des Systèmes de Santé, Villeurbanne Cedex, France
4
University of Antwerp, Antwerp, Belgium
5
6
Danderyd Hospital, Danderyd, Sweden
Stockholm School of Economics, Stockholm, Sweden
7
University Hospital Leuven, Leuven, Belgium
8
9
Hospital Center, Annecy, France
Scottish Center for Infection and Environmental Health, Glasgow, Scotland
10
Center for Public Health, Valencia, Spain
11
University of Valencia, Valencia, Spain
12
Pasteur Mérieux MSD, Lyon, France
Correspondence and reprint requests to:
André Ament, Ph.D.
Department of Health, Organization, Policy and Economics
Maastricht University
PO Box 616
6200 MD Maastricht
The Netherlands
Phone: ++ 31 43 3881723
Fax: ++31 43 3670960
E-mail: andre.ament@BEOZ.unimaas.nl
1
ABSTRACT
Several studies have shown that pneumococcal vaccination of older persons would be cost-effective in
preventing pneumococcal pneumonia, but evidence of clinical protection for this condition has been
uncertain. Given much better evidence of vaccination effectiveness against invasive disease, studies
showing that vaccination is cost-effective in preventing invasive disease alone could provide strong
support for public policies to vaccinate older persons. Using a cohort model approach, we examined the
cost-effectiveness of pneumococcal vaccination in five western European countries: Belgium, France,
Scotland, Spain, and Sweden. For each country, we determined the cost-effectiveness of pneumococcal
vaccination of persons  65 years in age in preventing hospital admission for both invasive
pneumococcal disease and pneumococcal pneumonia. The analyses used clinical and economic variables
unique to each country in 1995 and considered the cost savings from simultaneous administration of
pneumococcal and influenza vaccines. In the base case analyses, the cost-effectiveness ratios for
pneumococcal vaccination in preventing invasive disease varied between ~ ecu 11,000 to ~ ecu 33,000
per quality adjusted life year (QALY) among the five countries. Using more plausible epidemiological
assumptions on the incidence (50 cases per 100,000) and mortality (20% to 40%) of invasive disease, the
cost-effectiveness ratios were ~ ecu 12,000 or less per QALY in all five countries. If vaccination is
also clinically effective in preventing pneumococcal pneumonia, vaccinating all elderly persons would
be highly cost-effective to cost saving. Because of the similarity of the results in all five countries,
pneumococcal vaccination should be acceptably cost-effective in other western European countries. On
the basis of our findings, we believe public health authorities should consider policies for encouraging
pneumococcal vaccination for all persons  65 years in age.
2
INTRODUCTION
Pneumococcal infections are common in western European countries, accounting for a large
number of hospitalisations and deaths each year, especially among persons  65 years in age (1).
Retrospective studies in the United States have shown that pneumococcal polysaccharide vaccine is 5080% effective in preventing hospitalization for invasive pneumococcal disease in older persons (2-6).
Earlier prospective studies determined that vaccination also prevents pneumococcal pneumonia in
healthy young adults (7,8), but clinical trials in older adults have been inconclusive (9-14).
Pneumococcal vaccine has not been widely used since it was first introduced 20 years ago. Most of
the vaccine has been used in the United States, but even so only ~ 45% of elderly Americans have ever
been immunized (15). Until recently, few people in western European countries had ever received
pneumococcal vaccine (16).
The evidence that pneumococcal vaccination is clinically effective is becoming more widely
accepted in western Europe. This is reflected in recent changes in vaccination policies in the United
Kingdom, Sweden, Belgium, Norway, Denmark, Finland and Germany (16,17). Moreover, increases in
vaccine use have also been observed in most of these countries. More rapid change in vaccination
policies and increases in vaccine uptake might occur if additional evidence were to become available on
its cost-effectiveness.
Several studies in western Europe have shown that pneumococcal vaccination of older persons
would be cost-effective in preventing pneumococcal pneumonia (18-20). Given limited evidence that
vaccination is clinically effective in preventing all cases of pneumococcal pneumonia, nonbacteremic as
well as bacteremic (21,22), and better evidence that it is effective in preventing invasive disease,
information on the cost-effectiveness of vaccination in preventing invasive disease could provide strong
support for public policies to vaccinate older persons. In the U.S., pneumococcal vaccination of older
persons has been shown to be highly cost-effective in preventing hospitalization for pneumococcal
bacteremia, and in most instances it is cost saving (23). Because of differences between the U.S. and
countries in western Europe regarding the organisation and costs of health care, these findings may not
be persuasive to vaccination policy makers in western Europe. They may be more likely to accept costeffectiveness studies if the studies reflect European experience, are multinational in scope, and consider
3
the cost-effectiveness of vaccination in preventing invasive disease. In this study we examined the costeffectiveness of pneumococcal vaccination in preventing invasive pneumococcal disease because there is
convincing evidence of its clinical effectiveness for this outcome. However, in order to compare our
results with these of other studies, we also estimated the cost-effectiveness of vaccination in preventing
pneumococcal pneumonia. Our study is one of few multinational cost-effectiveness studies of health
care interventions (24).
METHODS
Model. In performing our economic evaluation we followed the principles of 'reference case
analysis' (25). We developed a cohort model and applied it to each of five countries: Belgium, France,
Scotland, Spain, and Sweden. Two hypothetical cohorts of individuals  65 years in age were
considered: one that received pneumococcal vaccine and one that did not. Both cohorts were followed
during the rest of their life spans. The life span of the cohort without vaccination was calculated on the
basis of age-specific mortality rates (Eurostat data). Vaccination would influence the age-specific
mortality rates during the period of vaccine efficacy. The model allowed us to calculate the
consequences of vaccination for single years of each age group above 65 year. The difference in life
years between the two cohorts with and without vaccination was calculated for each age group for each
year of life. The life years gained due to vaccination were then corrected for quality of life and
discounted, leading to the number of quality adjusted life years (QALYs) gained. The differences in
hospitalization costs for invasive pneumococcal disease and pneumococcal pneumonia were calculated
for the two cohorts. These costs were calculated for each age group and for each year of life as a
function of admission rate, average length of stay, cost of one hospital day, and vaccine efficacy. Finally
the cost per QALY gained as a result of vaccination was determined. All results were expressed as costeffectiveness ratios (CERs); i.e., the costs of achieving one additional quality-adjusted life-year in the
vaccinated cohort. We carried out our analyses using a societal point of view. To reflect the time
preference of society, we discounted all future costs and QALYs at 3% in the base case analyses (25).
Because other studies have demonstrated that the cost-effectiveness of pneumococcal vaccination varies
considerably with age (20,23), we performed separate analyses for persons 65-74 years, 75-84 years and
4
85 years of age and older, as well for all persons  65 years in age. The base case analyses used the
specific epidemiologic and economic values for each variable in each country. We also performed
sensitivity analyses for persons  65 years in age and varied the values of uncertain variables over their
reasonable ranges. A full description of the model is available at request.
Health effects. To calculate QALYs, we incorporated data on the incidence and mortality of
invasive pneumococcal disease and pneumococcal pneumonia, data on the effectiveness of
pneumococcal vaccination, and data on age-specific mortality rates and age-specific differences in
quality of life. Data on the effectiveness of pneumococcal vaccination were drawn from a case-control
study of invasive disease that provides the only available information on vaccination-effectiveness by
age group and on declining protection over time(3); for example, the effectiveness for persons  65 years
ranges from 75% in the first year to 33% in the sixth year after vaccination. These data were used in the
U.S. study on the cost-effectiveness of vaccination in preventing pneumococcal bacteremia (23). We
assumed the same effectiveness of vaccination in preventing pneumococcal pneumonia as in preventing
invasive disease, an assumption supported by the results of earlier clinical trials in South Africa (17).
We further assumed that the vaccine covers 88% of capsular polysaccharide types causing invasive
disease and pneumococcal pneumonia (23,26). We did not include adverse events following vaccination
in our analyses, as the incidence of 5 cases of anaphylaxis per million doses of vaccine administered,
used in the earlier U.S. study, had very little effect on the results (23). We adjusted life-years gained for
quality of life to reflect less-than-perfect health during surviving years, using weights obtained by the
U.S. National Health Interview Survey (27) because similar data for western European countries are
unavailable. The weights were .76, .74, .70, .63, and .51, for persons 65-69, 70-74, 75-79, 80-84, and 85
years and older, respectively.
Epidemiologic variables. Pneumococcal pneumonia. A hospital discharge diagnosis of
pneumococcal pneumonia (ICD 9 CM 481) is not a reliable indicator of pneumococcal pneumonia, but
a discharge diagnosis of pneumonia due to any cause (ICD 9 CM 480-486; first-listed diagnosis) is a
very reliable indicator for pneumonia (17,28). Numerous clinical studies have shown that definite
pneumococcal pneumonia (i.e., blood-culture positive) and probable pneumococcal pneumonia
(variously defined, but usually including purulent sputum with a gram stain and/or sputum culture
5
positive for Streptococcus pneumoniae) account for approximately 30-50% of all adult cases of
community-acquired pneumonia requiring hospitalisation (17). We used 40% as an estimate of incidence
in our base case analyses.
For each country, we obtained information on the number of hospital discharges for pneumonia
(ICD 9 CM 480-486; first-listed diagnosis), average length of stay (ALOS) and mortality rates for each
age group (Table 1). For Belgium, Spain, and Sweden the hospital data were obtained for 1995 from
national registrations maintained by the Ministries of Health (see Appendix; B1, Sp1, Sw1). For France,
hospital data were obtained from 35 hospitals in the Rhône-Alpes region for 1995 (F1). For Scotland,
hospital data were obtained from all non-obstetric, non-psychiatric National Health Service hospital
discharge records in Scotland for 1995 (Sc1). Because the Scottish ALOS and mortality data included
data from long-stay medical facilities and geriatric units, we truncated the ALOS and mortality data sets
to exclude those cases with more than twice the median ALOS.
Invasive pneumococcal disease. All laboratory isolates of Streptococcus pneumoniae obtained
from normally sterile extra-pulmonary sites (e.g., blood, cerebrospinal fluid, joint space, etc.) were
considered as invasive isolates. When multiple isolates were obtained from the same patient during the
same hospital stay, they were counted as a single case. Reports of invasive isolates also included data on
ALOS and the outcome of hospital care (survived or died) (Table 1). In all countries, data on the
incidence of invasive pneumococcal disease were obtained for one year from the clinical microbiology
laboratories of hospitals that serve a population of a region (or a country) of at least 3 million persons.
For smaller regions, a longer period was chosen in order to obtain a similar number of person-years of
observation. Where complete reports from all clinical microbiology laboratories in a region were not
available, reports were accepted as representative if they came from a subset of hospitals that accounted
for > 90% of all discharges of patients  65 years in age in the region.
In Belgium, data on the incidence of invasive pneumococcal disease were obtained from 10 hospital
laboratories in Flemish Brabant (1,100,000 inhabitants) for the period 1995-1997 (B2). Data on
mortality and ALOS for pneumonia were applied as a proxy for mortality and ALOS for invasive
pneumococcal disease because data on these latter were not available (B1). For France, data on the
incidence of invasive pneumococcal disease were obtained from the National Network for Public Health
6
for 1996. Mortality and ALOS data were based on a sample of five different hospitals in several regions
for 1995-1997 (F2). For Scotland (5,100,000 inhabitants), data on the incidence of invasive disease
(blood and CSF isolates) were obtained from all National Health Service hospital bacteriology
laboratories for 1995 (Sc2). ALOS and mortality data for pneumonia were applied as a proxy for ALOS
and mortality data for invasive disease because data on these latter were not available (Sc1). For Spain,
data on the incidence and ALOS of invasive disease were based on all hospital laboratory reports in the
Valencian Autonomous Community (3,500,000 inhabitants) for 1995 (Sp2). Mortality data on
pneumonia were applied as a proxy for mortality data for invasive disease as the latter were not available
(Sp1). For Sweden, data on the incidence of invasive disease, mortality, and ALOS were obtained from a
prospective study of community-acquired pneumococcal bacteremia treated in all hospitals in Stockholm
county (1,600,000 inhabitants) for the two-year period from September 1, 1993 to August 31, 1995
(Sw2).
Economic variables. Given differences in health care systems in the five countries, the approach to
estimate costs varied from country to country. Wherever data were available, estimates were based on
real costs; i.e., on resources consumed rather than on charges (29). Medical costs included the costs for
purchasing and administering the vaccine and the hospital costs for the treatment of invasive
pneumococcal disease or pneumococcal pneumonia (Table 2). Costs were estimated in each national
currency and are presented in the 1995 European currency unit, the ecu (Table 2). (The ecu was
replaced by the Euro in January 1998, and at that time had nearly the same value as the 1995 ecu). For
Belgium, France, Scotland, and Spain, all costs were standardised for 1995 by considering each
country's medical price index, defined as the annual increase in medical prices during the period 19851995 (B3;F3;Sc3;Sp3). These data were not available for Sweden, so the general price increase was
applied (Sw3).
The prices for pneumococcal vaccine in each country were retail prices (Pasteur Mérieux MSD,
unpublished data). Although the study was based on 1995 epidemiologic and economic data, vaccine
prices for 1998 were used. Two approaches were used to estimate the costs of administering
pneumococcal vaccination. The base case analyses considered the cost savings from simultaneous
pneumococcal and influenza vaccination. Four of the study countries recommend annual influenza
7
vaccination for all elderly persons: France ( 70 years), Spain and Sweden ( 65 years) and Belgium (
60 years). In Scotland, influenza vaccination was recommended only for persons with high-risk medical
conditions in 1995, but routine vaccination for persons  75 years in age was introduced in 1998.
Pneumococcal and influenza vaccines can be administered concurrently at separate sites without adverse
effects and without compromising their immunogenicity (16,30), and most doses of pneumococcal
vaccine are administered during the influenza vaccination season each year (DS Fedson, unpublished
observation). In Spain and the Netherlands, adding pneumococcal vaccination to an influenza
vaccination program incurs extra administrative costs of ecu 2.59 and 4.02, respectively (M Postma,
unpublished observation). In our base case analyses, we used ecu 3.00 for all five countries.
The sensitivity analyses considered the costs of administering pneumococcal vaccine by itself.
Estimates of these costs varied between countries (Table 2). In Belgium, France, Scotland, and Sweden,
the vaccine is administered by general practitioners (GPs). In Belgium and France, these costs were
estimated by the fees that GPs receive for each consultation for vaccination (B4;F4). For Scotland and
Sweden, they were based on real cost estimates for GP consultations, and in the case of Sweden were
corrected for the number of vaccinations that can be carried out during one consultation (Sc4;Sw4). In
Spain, GP costs were estimated by dividing the total national GP costs by the total national number of
GP consultations for 1995 (Sp4).
The costs of treating invasive pneumococcal disease and pneumococcal pneumonia were limited to
costs associated with hospitalisation. Because reimbursement systems differ among the five countries, it
was not possible to develop a uniform approach to estimate costs for each hospital stay. In countries
where hospital care is financed according to diagnosis-related groups (DRGs), DRG reimbursements
were considered to be the best basis for estimating the costs of hospital care. To reflect age-specific
hospital care costs, these costs were divided by the ALOS for each specific age-group to arrive at daily
hospital care costs. This approach was used for Spain (DRGs 079, 080, 089, 090, 091, and 489), and
Sweden (DRGs 089, 090) for 1995 (Sp5;Sw5). The approach was also applied for France (DRG 129),
although DRGs are not used for hospital financing in this country (F5). In Belgium and Scotland, where
hospital care is financed through global budgets, the costs of each episode were estimated by multiplying
the ALOS by the average cost per day. For Belgium, these estimates were based on an evaluation of 410
8
detailed invoices for the care of pneumonia patients in 46 general and 4 university hospitals in 1995
(B5). For Scotland, the daily hospital costs were estimated by dividing total national hospital costs by
the total national number of hospital days for 1995 (Sc5). Estimates of hospital care costs included only
costs related to acute care: costs of long-term care and psychiatric care were excluded. Medication costs,
specialist salaries and hospital overhead and investment costs were included in the costs of hospital care.
All costs of outpatient care were excluded.
RESULTS
Base case analyses. To prevent invasive pneumococcal disease in persons  65 years in age, the
cost-effectiveness ratios (CERs) varied from ~ ecu 11,000 for Spain to ~ ecu 33,000 per QALY for
Sweden (Table 3). If vaccination is also clinically effective in preventing pneumococcal pneumonia,
vaccination was highly cost-effective to cost saving in all five countries (Table 3). The age-specific
results showed no clear pattern of increasing or decreasing cost effectiveness. Increases in incidence
rates, mortality rates, and hospital resource use in older age groups appeared to cancel out declining
vaccination effectiveness observed with increasing age.
Sensitivity analyses. We conducted univariate sensitivity analyses for all persons  65 years in age
for invasive pneumococcal disease (Table 4). When pneumococcal vaccine was administered by itself
and not with influenza vaccine, the CERs increased in some countries as much as two-fold. When
alternative values for the incidence of disease were assumed (e.g., 50 cases of invasive pneumococcal
disease per 100,000 elderly persons), the CERs decreased in four countries ( ~ ecu 8,000 per QALY for
France and Scotland and ~ ecu 16,000 for Belgium and Sweden), but not in Spain (~ ecu 13,000 per
QALY). When mortality from invasive disease for all persons  65 years was assumed to be 40%, the
CERs decreased considerably and varied between ~ ecu 4,000 per QALY for Spain and ~ ecu 12,000 per
QALY for Scotland. Alternative assumptions on vaccine prices, vaccination effectiveness, not applying
the quality-of-life adjustment, and the discount rate had relatively large impacts on the results. For
example, when cost and benefits were not discounted the CERs varied between ecu 7,000 (for Spain)
and 24,000 (for Sweden).
9
Because the epidemiologic data on the incidence and mortality of invasive disease differed
substantially among the five countries (Table 3), we also conducted a two-way sensitivity analysis for
each country. We assumed a common incidence of invasive pneumococcal disease of 50 cases per
100,000 persons  65 years in age and varied mortality rates from 20% to 40%; all other variables were
those used in the base case analyses. In all countries, the CERs were less than ~ ecu 12,000, ~ ecu 9,000,
and ~ ecu 6,000 for mortality rates of 20%, 30% and 40% respectively (Table 5).
DISCUSSION
The cost-effectiveness of pneumococcal vaccination varied considerably across the five countries.
When vaccination against invasive disease alone was considered, the CERs in the base case analyses
ranged from ~ ecu 11,000 to ~ ecu 33,000 per QALY. Policy makers generally consider interventions
with cost-effectiveness ratios below $20,000 (about Euro 20,000 on December 1999) per QALY as
acceptably cost-effective, while those between $20,000 and $100,000 per QALY are considered
moderately cost-effective (32). Thus our findings, based on country-specific data indicate that
pneumococcal vaccination to prevent invasive pneumococcal disease alone can be considered an
acceptably to moderately cost-effective intervention in all five countries.
In our analyses, pneumococcal vaccination to prevent pneumococcal pneumonia in older persons
was highly cost-effective to cost saving in all countries. Although pneumococcal vaccination clearly
prevents bacteremic pneumococcal pneumonia in older persons, whether it prevents all cases of
pneumococcal pneumonia (nonbacteremic as well as bacteremic) has been a topic of considerable
dispute (6,11,13,14,17,21,22,33,34). A recent report indicates that pneumococcal vaccination reduces
rates of hospitalization for pneumonia and all-cause mortality in persons with chronic lung disease (35).
Nonetheless, we believe that our results on the cost-effectiveness of pneumococcal vaccination in
preventing pneumococcal pneumonia should be interpreted with caution.
The results of our analyses were sensitive to alternative assumptions. The base case analyses
assumed that pneumococcal and influenza vaccines would be given concomitantly, and that the
preventive effect of pneumococcal vaccination would be in addition to that of influenza vaccination.
When the analyses considered that pneumococcal vaccine would be given by itself, the cost-
10
effectiveness ratios increased, but they could still be regarded as moderately cost-effective in all five
countries.
The results were sensitive to assumptions on the incidence of invasive disease. The large
differences in reported incidence rates among different countries (table 1) seem not be related to real
variations in the magnitude of disease occurrence but rather to differences in surveillance systems and
case ascertainment (DS Fedson, et al, unpublished observations). The large differences in invasive
disease incidence that have been found between different regions within individual countries in single
years strengthens this argument. Thus, the comparatively high incidence rate for invasive disease in
Spain (Table 3) may be a more realistic estimate of its true incidence in all five countries than the lower
rates reported from the other four countries. The sensitivity analyses showed that the economic
attractiveness of pneumococcal vaccination increases considerably when higher incidence rates are
applied to each country.
The cost-effectiveness of vaccination to prevent invasive disease was also sensitive to assumptions
on mortality rates. For Belgium, Scotland and Spain, mortality data on pneumonia were applied as
proxies for mortality data for invasive disease because the latter were not available. Because invasive
disease is the most severe manifestation of pneumococcal pneumonia, its mortality was most likely
underestimated in Belgium and Spain (see Table 1). Population-based studies and case series from
several other countries indicate that mortality rates for invasive disease in persons  65 years in age are
usually 20 to 40% (17). The sensitivity analyses showed that for most countries, vaccination becomes
economically more attractive when these mortality rates are applied.
The mutual application of presumed common incidence rates (50 cases per 100,000) and mortality
rates (20% to 40%) for invasive disease in the two-way sensitivity analyses showed that the costeffectiveness ratios for pneumococcal vaccination would be ~ ecu 12,000 or less per QALY in all five
countries. We believe these findings provide a useful corrective for the otherwise disparate results
obtained when the actual epidemiologic data reported for each country were used in the base case
analyses.
The results of our cost-effectiveness analyses for invasive disease in the five European countries
were different from the U.S. results which showed that vaccinating all elderly persons was cost saving
11
under most assumptions (23). Almost all of the epidemiologic and economic variables used in the U.S.
study were more favourable towards vaccination than those applied in our base case analyses. The U.S.
analysis assumed a higher incidence of bacteremic disease, a higher mortality rate (29%), lower vaccine
purchase and administration costs (together ecu 9.3), and higher hospitalisation costs (ecu 6,971 per
episode of hospitalization). In addition, the U.S. study did not assume simultaneous administration of
pneumococcal and influenza vaccines. However, the differences in cost-effectiveness estimates between
the U.S. and western European analyses cannot be ascribed wholly to the use of different values for
these variables. When we applied the epidemiologic and economic data used in the U.S. study to our
model, we obtained a CER of ~ ecu 1,500 per QALY. Part of this difference might be ascribed to the
use of different values for other variables and/or the use of a different model. We used a cohort model in contrast to the Markov model used in the U.S. study. Nonetheless, like the investigators who
conducted the U.S. study, we believe that our cost-effectiveness estimates are conservative; both
analyses covered only hospitalization for invasive disease and disregarded health improvements or cost
savings stemming from vaccination that might reduce the need for outpatient care. Including these costs
or the costs of hospitalization for non-bacteremic pmeumococcal pneumonia would make pneumococcal
vaccination more economically attractive.
Several constraints on the availability of data from the different countries may have affected our
results. As in the U.S. study (23), we did not include the time lost for persons who were vaccinated. This
should be captured by measuring the reduction in quality of life (36), which, because of the few hours
involved, would be negligible. In addition, we applied the same age-dependent quality-of-life weights for
persons in all five European countries as were used in the U.S. study (23). There is little reason to
assume, however, that quality-of-life weights should differ appreciably between European countries and
the U.S., so the impact of this assumption on our results is likely to have been small. Finally, as in the
U.S. study, our base case analyses excluded costs in life-years gained because of vaccination. Whether to
include these costs is a topic of considerable debate (25); their inclusion has a large negative effect on
the cost-effectiveness of any life saving health care intervention.
In summary, we have shown that pneumococcal vaccination to prevent invasive pneumococcal
disease is acceptably to moderately cost-effective in five western European countries. Moreover, when
12
plausible assumptions on the incidence and mortality of invasive disease are applied to all five countries,
the cost-effectiveness ratios for pneumococcal vaccination are ~ ecu 12,000 or less per QALY. If
vaccination is also clinically effective in preventing pneumococcal pneumonia, it is always highly costeffective and usually cost saving. Because of the similarity of the results in all five countries,
pneumococcal vaccination is likely to be acceptably cost-effective in other western European countries
as well. On the basis of our findings, we believe public health authorities should consider policies for
encouraging pneumococcal vaccination for all persons  65 years in age.
ACKNOWLEDGEMENTS
The authors thank JE Sisk and WJ Whang for their advice on the methods used in the study.
Furthermore, we thank S Evers and J van Emmerik for their assistance in developing the cohort model.
The study was sponsored through an unrestricted grant from Pasteur-Mérieux MSD.
13
REFERENCES
1. Fedson DS. Pneumococcal vaccination. Four issues for western Europe. Biologicals 1997;25:215-9.
2. Sims RV, Steinmann WC, McConville JH, et al. The clinical effectiveness of pneumococcal vaccine
in the elderly. Ann Intern Med 1988;108:653-7.
3. Shapiro ED, Berg AT, Austrian R, et al. The protective efficacy of polyvalent pneumococcal
polysaccharide vaccine. N Engl J Med 1991;352:1453-8.
4. Butler JC, Breiman RF, Campbell JF, et al. Pneumococcal polysaccharide vaccine efficacy. An
evaluation of current recommendations. JAMA 1993;270:1826-31.
5. Farr BM, Johnston BL, Cobb DK, et al. Preventing pneumococcal bacteremia in patients at risk.
Arch Intern Med 1995;155:2336-40.
6. Fedson DS, Shapiro ED, LaForce FM, et al. Pneumococcal vaccine after 15 years of use: another
view. Arch Intern Med 1994;154:2531-5.
7. Austrian R, Douglas RM, Schiffmann G, et al. Prevention of pneumococcal pneumonia by
vaccination. Trans Assoc Amer Phys 1976;89:184-94.
8. Smit P, Oberholzer D, Hayden-Smith S, et al. Protective efficacy of pneumococcal polysaccharide
vaccines. JAMA 1977;238:2613-6.
9. Gaillat J, Zmirou D, Mallaret MR, et al. Essai clinique de vaccin antipneumococcique chez les
personnes âgées vivant en institution. Rev Epidemiol Sant Publ 1985;33:437-44.
10. Simberkoff MS, Cross AP, Al-Ibrahim M, et al. Efficacy of pneumococcal vaccine in high-risk
patients. N Engl J Med 1986;315:1318-31.
11. Fine MJ, Smith MA, Carson CA, et al. Efficacy of pneumococcal vaccination in adults. A metaanalysis of randomized controlled trials. Arch Intern Med 1994;154:2666-77.
12. Koivula I, Sten M, Leinonen M, et al. Clinical efficacy of pneumococcal vaccine in the elderly in a
randomised, single-blind population-based trial. Am J Med 1997;103:281-90.
13. Ortqvist A, Hedlund J, Burman LA, et al. Randomized trial of 23-valent pneumococcal capsular
polysaccharide vaccine in the prevention of pneumonia in middle-aged and elderly people. Lancet
1998;351:399-403.
14
14. Honkanen PO, Keistinen T, Miettinen L, et al. Incremental effectiveness of pneumococcal vaccine
on simultaneously administered influenza vaccine in preventing pneumonia and pneumococcal
pneumonia among persons aged 65 years or older. Vaccine 1999; 17:2493-2500.
15. Centers for Disease Control and Prevention. Influenza and pneumococcal vaccination levels among
adults aged  65 years. MMWR Morbid Mortal Weekly Report 1998;74:797-8.
16. Fedson DS. Pneumococcal vaccination in the United States and 20 other developed countries, 19811996. Clin Infect Dis 1998;26:1117-23.
17. Fedson DS, Musher DM, Eskola J. Pneumococcal vaccine. In: Plotkin SA, Orenstein WA (Eds).
Vaccines, 3rd Ed. Philadelphia: WB Saunders Company, 1999;553-607.
18. Plans-Rubio PP, Morales PG, Sanmarti LS. Coste-efectividad de la vacunacion neumococia en
Cataluña. Rev Esp Salud Publica 1995;69:409-17.
19. Jimenez FJ, Guallar P, Rubio C, et al. Cost-effectiveness analysis of pneumococcal vaccination in
the elderly Spanish population. Br J Med Econ 1996;10:193-202.
20. Baltussen RMPM, Ament AJHA, Leidl RM, et al. Cost-effectiveness of vaccination against
pneumococcal pneumonia in The Netherlands. Eur J Public Health 1997;7:153-61.
21. Simberkoff MS. Pneumococcal vaccine in the prevention of community-acquired pneumonia: a
sceptical view of cost-effectiveness. Sem Respir Infect 1993;8:294-9.
22. Fedson DS. Pneumococcal vaccination in the prevention of community-acquired pneumonia: an
optimistic view of cost-effectiveness. Sem Respir Infect 1994;8:285-93.
23. Sisk J, Moskowitz AJ, Whang W, et al. Cost-effectiveness of vaccination against pneumococcal
bacteremia among elderly people. JAMA 1997;278:1333-9.
24. Drummond MF, Bloom BS, Carrin G, et al. Issues in the cross-national assessment of health
technology. Int J Techn Asses Health Care 1992;10:467-78.
25. Weinstein MC, Siegel JE, Gold MR, et al. Recommendations of the panel on cost-effectiveness in
health and medicine. JAMA 1996;276;15:1253-8.
26. Hager HL, Woolley TW, Berk SL. Review of recent pneumococcal infections with attention to
vaccine and nonvaccine serotypes. Rev Infect Dis 1990;12:267-72.
27. Erickson P, Wilson R, Shannon I. Years of healthy life. Stat Notes No. 7, April 1995, 1-14.
15
28. Saitz EW. Pneumococcal pneumonia: a misleading diagnosis for audit studies. JAMA
1978;239:2372.
29. Finkler SA. On the distinction between costs and charges. Ann Intern Med 1982;96:102-9.
30. Honkanen PO, Keistinen T, Kivela SL. Reactions following administration of influenza vaccine
alone or with pneumococcal vaccine to the elderly. Arch Intern Med 1996;156:205-8.
31. Solano V, Rubio S, Hernandez Navarette M, et al. Costes de las inoculaciones accidentales en
personal sanitario de un Hospital. Gac Sanit 1998;12:29-34.
32. Laupacis A, Feeny D, Detsky AS, et al. Tentative guidelines for using clinical and economic
evaluations revisited. Can Med Assoc J 1992;146:473-81.
33. Fiebach N, Beckett W. Prevention of respiratory infections in adults. Influenza and pneumococcal
vaccines. Arch Intern Med 1994;154:2545-57.
34. Hirschmann JV, Lipsky BA. The pneumococcal vaccine after 15 years of use. Arch Intern Med
1994;154:373-7.
35. Nichol KL, Baken L, Wuorenma J, Nelson A. The health and economic benefits associated with
pneumococcal vaccination of elderly persons with chronic lung disease. Arch Intern Med
1999;159:2437-42
36. Johannesson M. Avoid double-counting in pharmaeconomic studies. Pharmacoeconomics 1997;
11:385-88.
16
APPENDIX. Sources of country-specific data:
For all countries: age specific population and age specific mortality rates: Eurostat
Belgium. B1: Ministerie van Sociale Zaken, Volksgezondheid en Leefmilieu, Bestuur voor
gezondheidszorg 1996; B2: Unpublished data, J Verhaegen; B3: Ministry of Economic Affairs, 1996;
B4: Official tariff, Ministry of Health 1997; B5: Rijksinstituut voor Ziekte en InvaliditeitsVerzekering.
France. F1, F2: Unpublished data, J Gaillat; F3, F4 Eco Santé France 1998; F5: Cram Rhone-Alpes,
Eco Santé France, 1998.
Scotland. Sc1: Information & Statistics Division, Common Services Agency, Scotland; Sc2: Scottish
Centre for Infection & Environmental Health; Sc3: Hospital & Community Health Services Pay & Prices
Index, NHS Management Executive, Leeds, 1996; Sc4: Netten A, Dennett J. Unit costs of health and
social care, 1997; Sc5: Information & Statistics Division, Common Services Agency, Scotland, 1995.
Spain. Sp1, Sp2: Valencian Health Service 1995. Minimum Data Set; Sp3: Spanish National Institute of
Statistics 1996; Sp4 Valencian regional health authority 1995; Sp5: Centre for Economics and Health
Management Research (CIEGS). Calculo del coste por proceso en los hospitales del Servicio Valenciano
de Salud, 1997.
Sweden. Sw1: Swedish National Board of Health; Sw2: Unpublished data, Kalin M, Ortqvist A, et al;
Sw3: National Accounts, Statistics Sweden. Sw4: Landstingsforbundst. Kostnader per intagen patient,
varddag, lakerbsok mm 1993, Stockholm, 1994; Sw5: SPRI, Halso - och sjukvardens utvecklingsinstitut.
17
Table 1. Epidemiologic variables for invasive pneumococcal disease and pneumonia
Variable
Age groups (yrs)
Pneumonia†
Invasive pneumococcal disease*
Belgium
France
Scotland
Spain
Sweden
Belgium
France
Scotland
Spain
Sweden
65-74
28.3
20.5
21.8
40.0
22.7
409.3
235.1
460.3
390.5
640
75-84
41.2
28.6
33.8
83.1
34.1
943.5
615.3
1,068.7
752.9
1,440
 85
65.4
67.7
62.9
74.5
49.2
1,947.1
1,051.2
2,510.3
1,168.1
2,630
 65
35.6
29.3
29.5
57.2
34.1
725.0
451.0
850.0
575.0
1,167
65-74
12.8§
18.9
24.5 §
8.0§
6.3
12.8
6.7
24.5
8.0
5.0
75-84
19.9§
20.6
40.0§
22.7§
12.9
19.9
8.7
40.0
22.7
9.6
 85
26.3§
42.4
52.1§
26.8§
20.7
26.3
13.5
52.1
26.8
13.0
 65
19.3§
25.8
37.9§
17.6§
11.7
19.3
9.5
37.9
17.6
9.2
65-74
16.5§
14.4
5.4§
11.6
11.0
16.5
10.9
5.4
12.2
6.7
75-84
19.1§
12.4
7.6§
12.1
9.9
19.1
11.6
7.6
11.9
7.3
 85
20.7§
11.8
9.4§
9.8
10.7
20.7
11.8
9.4
8.8
7.6
 65
18.7§
13.1
7.8§
11.5
10.5
18.7
12.0
7.8
11.2
7.2
Incidence
(per 100,000)
Mortality (%)
ALOS¶ (days)
*
Hospital admissions for invasive disease.
†
Hospital discharges for pneumonia (ICD-9 CM 480-486; first-listed diagnosis). Pneumococcal pneumonia was assumed to account for 40% of all pneumonias requiring
hospital admission.
§
In the absence of specific data for invasive pneumococcal disease, data for pneumonia were used.
¶
Average length-of-stay.
18
Table 2. Economic variables*†
Costs
Belgium
France
Scotland
Spain
Sweden
Hospital care (per day)
267.6
330.4
304.6
206.0
332.4
Vaccine purchase
19.0
13.8
14.3
11.5
10.2
Base case‡
3.0
3.0
3.0
3.0
3.0
Sensitivity analysis§
15.2
17.0
12.2
13.4
11.4
Medical price index (%)
3.0
6.5
4.0
6.0
3¶
Vaccine administration
*
Cost data are shown in European currency units (ecu) for 1995. See Appendix for details on sources for the data.
†
December 1995 exchange rates of national currencies to 1 ecu: Belgium: 38.14; France: 6.46; Scotland: 0.82; Spain: 161.46; Sweden: 9.27; U.S.:1.19.
‡
The base case analyses assumed that pneumococcal vaccine would be adminstered during the same visit as influenza vaccine.
§
The sensitivity analyses assumed that pneumococcal vaccine would be adminstered by itself, not at the same time as influenza vaccine.
¶
General price index (Appendix:Sw3).
19
Table 3. Results of the base case analyses
Age group
Invasive pneumococcal disease
Pneumococcal pneumonia
(yrs)
Belgium
France
Scotland
Spain
Sweden
Belgium
France
Scotland
Spain
Sweden
65-74
21,685†
15,552
13,253
11001
28,812
cost saving
2,067
429
cost saving
cost saving
75-84
30,979
28,294
15,769
7,871
32,999
cost saving
1,094
11
cost saving
cost saving
 85
68,339
27,864
27,464
40,657
68,744
cost saving
6,450
44
394
cost saving
 65
25,907
19,182
14,892
10,511
32,675
cost saving
2,126
242
cost saving
cost saving
† Cost-effectiveness ratio in ecu (1995) per QALY gained. See text for details.
20
Table 4. Results of the sensitivity analyses for invasive pneumococcal disease for persons  65 years in age*
Belgium
France
Scotland
Spain
Sweden
Base case†
25,907‡
19,182
14,892
10,511
32,675
Without influenza vaccination
43,106
37,886
23,422
20,013
57,314
Incidence of disease
(cases per 100,000  65 years)
30 / 100,000
29,593
15,217
13,645
23,774
31,123
40 / 100,000
21,003
10,633
9,975
17,101
21,821
50 / 100,000
15,849
7,883
7,773
13,097
16,240
20
19,796
19,904
23,135
7,342
15,024
30
13,197
13,269
15,423
4,895
10,016
40
9,898
9,951
11,566
3,671
7,512
current price – 3.00 ecus
21,678
15,174
12,110
7,770
23,875
current price + 3.00 ecus
30,136
23,190
17,673
13,252
41,474
best case
17,328
12,473
10,127
6,622
20,991
worst case
72,868
53,285
42,055
37,276
104,209
No quality-of-life adjustment
16,307
11,887
9,233
6,444
20,017
Discount rate, 0%
18,565
13,764
11,258
7,419
23,713
Discount rate, 5%
31,220
23,109
17,511
12,747
39,111
Mortality (%)
Vaccine price
Vaccination effectiveness§
*
In the sensitivity analyses only a single variable was changed. All other values were base case values.
†
Base case analyses assumed pneumococcal vaccine would be administered during the same visit as influenza vaccine.
‡
Cost-effectiveness ratio; the cost in ecu(1995) per QALY gained. See text for details.
§
Best case and worst case are high and low values of the 95% confidence interval of vaccination effectiveness, respectively. See reference 23 for details.
21
Table 5.Comparison of base case and two-way sensitivity analyses for invasive pneumococcal disease*
Belgium
France
Scotland
Spain
Sweden
25,907
19,182
14,892
10,511
32,675
20%
11,559
7,779
11,277
7,626
6,836
30%
7,706
5,186
7,518
5,084
4,557
40%
5,779
3,889
5,638
3,813
3,418
Base case†
Two-way sensitivity analyses
(incidence 50 cases /100,000)
Mortality
*
In the two-way sensitivity analyses only incidence and mortality data were changed. All other values were base case values.
†
Cost-effectiveness ratio for persons  65 years in age; the cost in ecu (1995) per QALY gained. See text for details.
22