A cost-benefit analysis of typhoid fever immunization programmes in an Indian urban slum community.
Abstract: Many economic analyses of immunization programmes focus on the benefits in terms of public-sector cost savings, but do not incorporate estimates of the private cost savings that individuals receive from vaccination. This paper considers the implications of Bahl et al.'s cost-of-illness estimates for typhoid immunization policy by examining how community-level incidence estimates and information on distribution of costs of illness among patients and the public-health sector can be used in the economic analysis of vaccination-programme options. The findings illustrate why typhoid vaccination programmes may often appear to be unattractive to public-health officials who adopt a public budgetary perspective. Under many plausible sets of assumptions, public-sector expenditure on typhoid vaccination does not yield comparable public-sector cost savings. If public-health officials adopt a societal perspective on the economic benefits of vaccination, there are many situations in which different vaccination programmes will make economic sense. The findings show that this is especially true when public decision-makers recognize that (a) the incidence of typhoid fever is underestimated by blood culture-positive cases and (b) avoided costs of illness represent a significant underestimate of the actual economic benefits to individuals of vaccination.

Key words: Typhoid fever; Cost-benefit analysis; Costs and cost analysis; Cost of illness; Vaccination; Typhoid-paratyphoid vaccines; Slums; Urban health; India
Article Type: Report
Subject: Typhoid fever (Usage)
Typhoid fever (Social aspects)
Public sector (Usage)
Public sector (Social aspects)
Public health (Usage)
Public health (Social aspects)
Medical research (Usage)
Medical research (Social aspects)
Medicine, Experimental (Usage)
Medicine, Experimental (Social aspects)
Cost benefit analysis (Usage)
Cost benefit analysis (Social aspects)
Vaccination (Usage)
Vaccination (Social aspects)
Authors: Poulos, Christine
Bahl, Rajiv
Whittington, Dale
Bhan, Maharaj K.
Clemens, John D.
Acosta, Camilo J.
Pub Date: 09/01/2004
Publication: Name: Journal of Health Population and Nutrition Publisher: International Centre for Diarrhoeal Disease Research Bangladesh Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2004 International Centre for Diarrhoeal Disease Research Bangladesh ISSN: 1606-0997
Issue: Date: Sept, 2004 Source Volume: 22 Source Issue: 3
Topic: Event Code: 290 Public affairs Computer Subject: Cost benefit analysis
Product: Product Code: 8000120 Public Health Care; 9005200 Health Programs-Total Govt; 9105200 Health Programs; 8000200 Medical Research; 9105220 Health Research Programs; 8000240 Epilepsy & Muscle Disease R&D; 8000146 Vaccination & Immunization NAICS Code: 62 Health Care and Social Assistance; 923 Administration of Human Resource Programs; 92312 Administration of Public Health Programs; 54171 Research and Development in the Physical, Engineering, and Life Sciences; 621999 All Other Miscellaneous Ambulatory Health Care Services
Accession Number: 212107159
Full Text: INTRODUCTION

Globally, there are an estimated 16 million cases of typhoid fever annually, causing 600,000 deaths mainly in developing countries (1). The World Health Organization recommends establishing typhoid fever vaccination programmes in endemic areas, and the Vi polysaccharide vaccine against typhoid fever has been shown to be effective, inexpensive, and well-tolerated (1). However, resources for implementing new and existing immunization programmes in developing countries are scarce, and the addition of new vaccines to current immunization schedules has logistical and financial difficulties (2). Many countries have concluded that they cannot afford to add new vaccines (e.g. against hepatitis B) to their Expanded Programme on Immunization (EPI), and public-health resources are under constant pressure to be redirected to other interventions.

In their study of an urban slum (Govindpuri) in New Delhi, India, Sinha et al. found that the incidence of typhoid fever in children aged five years and below was much higher than expected (3). Using household survey data from the same population, Bahl et al. generated the first estimates of costs of illness associated with typhoid fever at the community level in a developing country (4). Results of their study showed that the costs of illness per episode of typhoid fever were the highest in 2-5-year old children and that both hospitalization and clinical resistance to ciprofloxacin dramatically increased the cost of illness. Finally, Bahl et al. showed that the distribution of costs between private individuals and the public sector varies by age group (4). The public sector in India is estimated to bear 70% of costs of illness in 2-5-year old children. In all other age groups, the majority of costs of illness are borne by private households.

Many economic analyses of immunization programmes focus on the benefits in terms of public-sector cost savings, but do not incorporate estimates of the private cost savings that individuals receive from vaccination. This paper considers the implications of Bahl et al.'s cost-of-illness estimates for typhoid immunization policy by examining how community-level incidence estimates and information on the distribution of costs of illness among patients and the public-health sector can be used in the economic analysis of vaccination-programme options.

Economic analysis of vaccination programmes

Policy analyses of immunization programmes typically seek to answer the question: which immunization programme generates the best health outcomes per dollar spent by the public-health system? Health outcomes are commonly measured in terms of avoided cases, avoided deaths, and/or avoided costs. These analyses help decision-makers select programmes that maximize health benefits with a given budget. Economic analyses of immunization programmes typically measure the benefits of those programmes using avoided costs of illness (5-7)

Drummond et al. emphasize that the results of economic analyses will vary depending on the perspective assumed by the analyst (8). Most policy analyses of vaccination programmes take a public-sector budgetary perspective, i.e. they only look at avoided public-sector budget costs, or costs borne by healthcare providers, to measure vaccination benefits (5-7). Only a few take a societal perspective and measure the total costs of illness as the sum of avoided public and private costs (9). To the best of our knowledge, none of the previous economic analyses of vaccination programmes in the literature has explicitly compared the public-sector budget perspective with the societal perspective (public plus private costs avoided) using actual field data from a developing country.

This study uses estimates of both public and private costs of illness due to typhoid fever to analyze the economic and public-health impacts of three kinds of publicly-financed immunization programmes (school-based vaccination, targeted vaccination of pre-school children aged 2-5 years, and vaccination of the general population). We illustrate the difference between looking at the results of the economic analysis from a public budgetary perspective only and a fuller accounting of the economic benefits that includes cost savings to private individuals. We report the sensitivity of cost-benefit results to changes in three main sources of uncertainty: (a) the incidence of typhoid fever in the community, (b) the average cost of vaccination, and (c) the proportion of the total economic benefits represented by avoided costs of illness.

Background

Vi polysaccharide vaccine

The Vi polysaccharide vaccine is administered in one dose, as an injection, with revaccination recommended after three years (1). In most countries, it is indicated for use in adults and children aged over 24 months, while its efficacy in very young children needs to be determined (1). Studies in endemic areas have found that the efficacy of the Vi vaccine ranges from 55% to 75% (1,10). The Vi vaccine has a well-established safety profile with low incidence of mild side-effects, whether it is administered alone or with other vaccines (1).

Study population

The study area is a poor, densely-populated urban slum located in Kalkaji, New Delhi, India. The socioeconomic characteristics of the sample are described in the paper of Bahl et al. (4). Sinha et al. reported that the annual incidence of typhoid fever among all 0-40-year old residents was 9.8 per 1,000 persons, and the annual incidence in children aged less than five years was 27.3 cases per 1,000 persons (3). These are among the highest incidence rates of typhoid fever reported anywhere in the world.

Research design and fieldwork

The population of the study area was divided into clusters with 70 households in each, and 26 such clusters were randomly selected for active twice-weekly surveillance for detection of cases with typhoid fever (3). The residents in the remaining household clusters were kept under only passive surveillance at health facilities and neighbouring practitioners for detection of fever cases.

Blood specimens for culture were obtained from all children, aged less than five years, who had fever (temperature >38[degrees]C), identified through active or passive surveillance. Older children and adults had to have had continuous fever for at least three days before a blood specimen was obtained for culture. These methods have been described in detail previously (3).

In the study population, a total of 98 cases of culture-positive typhoid fever and 31 of culture-positive paratyphoid were identified through both active and passive surveillance. Ninety-four culture-negative cases exhibiting a characteristic clinical syndrome were also treated as typhoid fever. This 'clinical' typhoid syndrome was defined as continuous high fever for seven days or more with no other obvious causes and no response to three days of anti-malarial therapy. A diagnosis of 'clinical' typhoid was made earlier, i.e. after five days of fever, if the patient additionally had bradycardia or splenomegaly. Bahl et al. estimated the costs of illness for blood culture-positive Salmonella enterica serotype Typhi, blood culture-positive S. enterica serotype Paratyphi, and 'clinical typhoid' cases (4).

Measurement and calculation of costs of illness

Data to calculate the private costs of illness were collected through weekly in-person interviews conducted in the patient's home (4). Private, or patient, costs of illness are the sum of direct medical, direct non-medical, and indirect costs. Direct medical costs included out-of-pocket expenditure on consultation fees, laboratory tests, and medicines. Non-medical direct costs measured out-of-pocket expenditure on transportation, special foods and drinks, and other items. Indirect costs were calculated as the product of days of work missed by all household members and a monetary value of lost productivity. [Over the course of the study period from November 1995 to October 1996, the exchange rate between the Indian rupee and the US dollar fluctuated between INR 34.32 and 36.59 per US$. For the costs of illness presented in this study, we have assumed an average exchange rate over the period of US$ 1=INR 35.5. All costs are expressed in 1996 US$.] Public costs of illness, or non-patient costs, were defined as the costs borne by institutions or the public sector. Public costs were the sum of costs of outpatient visits or hospitalization in government hospitals, and laboratory tests and medicines provided free of charge to the patient.

Ex-ante costs of illness

The estimates of ex-ante costs-of-illness are the product of the average cost of illness and the disease incidence; they represent the per-capita expected annual losses due to typhoid fever. Table 1 presents these estimates for the entire study population and by age group. These point estimates represent averages around which the population's ex-ante costs of illness will be distributed. The private ex-ante costs of illness are the costs that households should expect to incur; the public ex-ante costs of illness are the costs that the public healthcare system should expect to incur.

The total private and public ex-ante costs were higher for children aged 2-5 years (US$ 4.34) than those for children aged 0-2 year(s) (US$ 0.70), individuals aged 5-19 years (US$ 0.90), and adult (US$ 0.06) subjects. Given that the average daily wage rate for an adult is US$ 1.97, these estimates imply that households could expect to lose two days' wages in the next year because of typhoid fever in 2-5-year old children and less than one-half day's wage due to typhoid fever affecting household members in other age groups.

The distribution of ex-ante costs between private individuals and the public sector varies widely by age group. The private costs are greater than public costs for all age groups except 2-5-year old children. In the 2-5-year age group, however, public-sector costs (US$ 3.32) are over three times greater than patient costs (US$ 1.04). The ex-ante public costs of illness are so high for 2-5-year old children because of the high incidence of typhoid fever in this age group and the relatively high rates of hospitalization. The private ex-ante costs are also high for this age group because of lost wages due to adult caretaking. The ex-ante public cost for adults is zero, indicating that typhoid fever in adults imposes little burden on the public sector. Considering all age groups, however, expected public costs (US$ 0.48) are roughly comparable to expected patient costs (US$ 0.42). The ex-ante cost of illness increases significantly for patients who exhibit antibiotic resistance and patients who were hospitalized (4). These results underscore the benefits of typhoid prevention in areas with antibiotic resistance and poor case treatment [See the paper by Bahl et al. (4) for more details, including sensitivity analyses conducted by varying assumptions on daily wages, the non-patient costs of outpatient visits, and the daily hospitalization costs].

MATERIALS AND METHODS

Policy interventions

We estimate the costs and benefits of three different public-sector vaccination programmes in the study area. The first is a mass vaccination campaign that is assumed to reach 80% of the total population in the study area. The second is a school vaccination campaign that is assumed to reach 80% of the 6-19-year old individuals [We do not have data on what percentage of school-age children are enrolled in schools. We assume that the vaccine would be made available to any child in this target age group, whether they attend school or not, via the health posts established for the vaccination campaign]. The third is a targeted vaccination campaign for pre-school children aged 2-5 years, which is assumed to reach 80% of this age group. All three programmes are assumed to be fully paid for by government, i.e. no user-fees are charged.

Vaccine benefits

We measure the outcomes of each of the three publicly-funded vaccination programmes in three different ways: (a) The number of typhoid fever cases avoided, calculated as the product of the incidence rate, cohort population, vaccine effectiveness, and vaccine coverage, i.e. the percentage of the population in an age cohort that is vaccinated; (b) Public-sector cost savings, estimated as the costs of illness avoided by public healthcare providers; and (c) Total public and private costs of illness avoided. We assume that the Vi vaccine is 70% effective for three years.

The first outcome measure is simply a non-monetarized count of the number of typhoid fever cases avoided by a vaccination programme. It is presumably of interest to both public-sector decision-makers and private individuals. However, from an economic perspective the benefits must be compared with the costs, and it is difficult to make such comparisons when the costs are measured in monetary units and the benefits are not. The second outcome measure is of particular interest to public-sector decision-makers who focus on the financial cost savings to the public-health budget. The third outcome measure reflects a broader societal perspective, combining both public and private-sector cost savings. Increases in all three measures indicate better programme performance.

We measure the economic benefits of the Vi vaccine by the avoided ex-ante cost of illness. If the vaccine were 100% effective in preventing typhoid fever, the expected benefits of the vaccine could be estimated by the present value of the ex-ante cost of illness avoided during the three years of protection. We calculate the present value of total vaccine benefits over the assumed three-year period of vaccine effectiveness by:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where, [PV.sub.B]=present value of the stream of vaccine benefits to the vaccinated population; Public [COI.sub.j]=annual ex-ante public-sector costs of illness avoided by vaccinating an individual in age cohort j (assuming that the vaccine is 70% effective); Private [COI.sub.j] =annual ex-ante private costs of illness avoided by vaccinating an individual in age cohort j (assuming that the vaccine is 70% effective); [n.sub.j]=number of individuals vaccinated by the programme in age cohort j; m=number of age cohorts; i=number of years the vaccine is effective; and r=real discount. The vaccine benefits are reported in Table 1.

We assume the real (i.e. net of inflation) discount rate to be 10%.

Vaccine costs

Field data on the costs of running the three different kinds of Vi vaccination programmes in this study population are not available. However, recent estimates of the cost of typhoid vaccination in Viet Nam have been prepared for the International Vaccine Institute based on data from a vaccination trial conducted in Hue, Viet Nam, in 2003. These cost estimates reflect the financial recurrent cost of resources that would be required to conduct a new Vi vaccination programme in Viet Nam. The per-unit vaccination cost was calculated by dividing the total reported expenditure on several cost items (e.g. vaccine dose, syringe, safety box, labour) by the number of vaccinations given. The estimates represent the 'incremental' costs of providing the typhoid vaccine to the target population in the sense that they assume that the existing cold-chain and central administrative structure have sufficient capacity to add the programme without the requirement of investing in additional capacity. The direct recurrent costs are estimated to range from US$ 0.79 to 1.58 per vaccine depending upon the assumptions made about the amount and value of the labour input needed for the vaccination programme (Stewart J. Personal communication, 2004). These per-unit vaccine costs are assumed to reflect the full costs of vaccine acquisition and delivery, including the opportunity costs of resources used in vaccine provision. From a societal cost-benefit perspective, one should add to these recurrent cost estimates the private costs to an individual in terms of transportation and time spent to obtain the vaccine.

Although there are numerous reasons why vaccination costs in India and Viet Nam may differ, we believe that these cost estimates of vaccination for Viet Nam are likely to be generally representative of costs that one would expect in urban India. However, to address the uncertainties in the cost estimates, we conduct our cost-benefit analysis using five assumed values for the per-capita cost of vaccination (Vc): US$ 0.75, US$ 1.00, US$ 1.50, US$ 2.00, and US$ 3.00.

Comparing costs and benefits

We compare the costs and benefits of the three vaccination programmes using five different metrics. First, we present a standard cost-effectiveness ratio: the dollars spent by the public sector on vaccination per typhoid fever case avoided:

[m.summation over (f=1)][n.sub.j][V.sub.c]/Total typhoid fever cases avoided (2)

Second, we present a ratio of the public-sector treatment cost savings per dollar spent by the public sector on vaccination:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

Our third cost-effectiveness ratio combines the three types of data in the first two cost-effectiveness ratios (i.e. the number of typhoid cases avoided, the costs of the vaccination programme, and the public-sector cost savings).

It is the public dollars spent on vaccination minus the public-sector cost savings per typhoid fever case avoided:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

This cost-effectiveness ratio shows the net cost to the public sector of avoiding a typhoid fever case, taking into account the fact that the costs to the public sector of vaccination are partially offset by reduced public-sector treatment costs. The numerator is, thus, the net effect on the public-health budget of avoiding a typhoid case. If the public-sector cost savings are greater than the vaccination programme costs, this numerator will be negative, implying that vaccination both reduces the public-health budget and saves typhoid cases, and is a win-win policy intervention.

Fourth, we calculate the total net benefits in the study area of each vaccination programme from the public-sector budgetary perspective:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

Fifth, we calculate the net benefits of each vaccination programme from the societal perspective:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

Limitations of methods

Several limitations of these benefit-cost calculations need to be emphasized. First, it has been shown both theoretically and empirically that the ex-ante cost of illness underestimates the benefits of disease prevention (11). The ex-ante private cost of illness does not incorporate (a) the expected pain and suffering due to typhoid fever, (b) the risks of mortality, or (c) the costs individuals may incur for activities that prevent typhoid fever. [Willingness-to-pay (WTP) for the Vi vaccine is a more comprehensive measure of the benefits of reducing the risk of typhoid fever. Individuals' expressed WTP for the Vi vaccine would likely reflect the reduction in ex-ante cost of illness, avoided pain and suffering, avoided disability, and the avoided costs of activities to prevent typhoid fever]. To explore the sensitivity of our results to the magnitude of this underestimation of total economic benefits, we multiply ex-ante cost-of-illness (public + private) estimates by a 'COI correction factor'-[alpha], that we vary from 1 (no correction) to 4 (which implies that avoided costs of illness represent 25% of the total economic benefits).

Second, the actual incidence of typhoid fever in the study area is difficult to estimate. To show the sensitivity of these cost-benefit calculations to changes in the incidence of typhoid fever in the study population, we use two estimates of incidence. First, we use a conservative estimate based on blood culture-confirmed cases detected by active surveillance (9.8 cases per 1,000). Second, we use a higher estimate based on blood culture-confirmed cases detected by both active and passive surveillance, plus estimates of 'clinical typhoid' (17.5 cases per 1,000).

Third, this analysis only considers the efficiency of alternative Vi vaccine programmes, not the efficiency of typhoid fever immunization programmes relative to other public-health programmes, such as EPI, polio eradication, or prevention of malaria or HIV/AIDS. The magnitude of Vi vaccine benefits relative to other health interventions will determine the priority placed on typhoid fever immunization programmes relative to other public-health programmes. This study also does not compare the efficiency of Vi vaccine programmes with other health interventions for typhoid fever, including water and sanitation improvements, changes in case treatment, or other vaccines for typhoid fever.

RESULTS

Panel A in Table 2 presents the results for the five different metrics for comparing costs and benefits of the mass vaccination programme for different per-unit vaccine costs, assuming an incidence rate in the study area based on blood culture-positive cases detected from active surveillance. Panels B and C show the same set of results for a school-based immunization programme and a pre-school vaccination programme respectively. There are a number of interesting findings from these calculations.

If one looks only at the public cost of avoiding a typhoid fever case (first cost-effectiveness ratio, metric 1), mass vaccination (Panel A) and school-based vaccination (Panel B) look similar. However, pre-school vaccination (Panel C) looks much more attractive. At a per-unit vaccine cost of US$ 1, the public vaccine cost per typhoid case avoided is about US$ 50 for mass vaccination and US$ 41 for school vaccination, but only about US$ 14 for a targeted vaccination programme for pre-school children.

An examination of the second metric (public treatment cost savings per dollar spent on vaccination) shows that from a public-sector budgetary perspective, the mass vaccination and school vaccination programmes do not look attractive at most per-unit vaccine costs, i.e. a dollar spent on vaccination yields less than a dollar reduction in public-sector treatment cost savings. However, a dollar spent on vaccinating 2-5-year old children yields more than a dollar reduction in public-sector treatment cost savings at all vaccine costs shown.

From the perspective of the net public cost per typhoid case avoided (metric 3), both mass vaccination and school vaccination look attractive at low to moderate vaccine costs. For example, at a per-unit vaccine cost of US$ 1, the net public cost of avoiding a case of typhoid is US$ 6 for mass vaccination and US$ 19 for school-based vaccination. Pre-school vaccination is much more attractive. At all per-unit vaccine costs shown, this intervention saves money and typhoid fever cases; in effect, the reduction in typhoid fever cases in this age cohort can make money for the public sector.

If one looks at the net benefits to the public sector (metric 4), the distinction between the three vaccination programmes is most stark. Mass vaccination and school-based vaccination have negative net benefits at almost all vaccine costs. The pre-school vaccination programme has high net public benefits at all vaccine costs shown in Table 2.

From an economic perspective, the net societal benefits (metric 5) are the most important results. In terms of net societal benefits, mass vaccination looks attractive at low to moderate vaccine costs (net societal benefits are positive for mass vaccination at a per-unit vaccine cost of US$ 1.50, but become negative at higher per-unit vaccine costs reported in Table 2). School-based vaccination has slightly lower net societal benefits than mass vaccination. On the other hand, the pre-school vaccination programme has positive net societal benefits at all the per-unit vaccine costs shown in Table 2.

As shown in Table 3, the cost-benefit results for all of these five metrics are sensitive to the assumptions one makes about the incidence of typhoid fever in the study population. Experienced epidemiologists are confident that estimating the incidence using the number of blood culture-positive cases detected even from active surveillance will result in a significant underestimate of typhoid fever in the population. If one includes the blood culture-positive cases from active plus passive surveillance and the 'clinical' typhoid cases in the calculation of the incidence rate, the cost-benefit results for all three vaccination programmes look much better. The net societal benefits (metric 5) of all three vaccination programmes are positive even at a high per-unit vaccine cost of US$ 2. Importantly, the net public benefits (metric 4) are still negative for mass vaccination if the vaccine costs are US$ 1.50 or higher and are negative for school-based vaccination if vaccine costs are US$ 1 or higher. Preschool vaccination looks even more attractive from both public and societal perspectives at all vaccine costs.

Just as epidemiologists are confident that the incidence of typhoid fever is significantly higher than indicated by blood culture-confirmed cases, economists are confident that the avoided private costs of illness are a significant underestimate of the actual economic value of risk reduction that a typhoid vaccine provides to individuals. Figure 1 shows how the net societal benefits of the three vaccination programmes change for different 'correction factors' for the magnitude of this underestimation of economic benefits, assuming an incidence rate based on blood culture-positive cases detected with active surveillance. Figure 2 shows the same results for an incidence rate based on blood culture-confirmed and 'clinical' typhoid fever. Note that the lines representing the mass immunization and school-based programmes are so close that they cannot be distinguished from one another.

In both the figures, the area lying above the lines representing the mass immunization and school-based programmes represents combinations of COI correction factors and vaccine prices for which programme benefits are greater than programme costs. In this region, all programmes make economic sense. The area lying between the two lines shown on each graph represents combinations of COI correction factors and vaccine prices for which the pre-school programme benefits are greater than the pre-school programme costs--costs exceed benefits for the other two programmes. The area lying below the line representing the pre-school programme represents combinations of COI correction factors and vaccine prices for which programme benefits are less than programme costs for all three vaccination programmes.

These sensitivity analyses show that, for moderate values of this economic correction factor (1.5-2.0), all three vaccination programmes easily pass a societal cost-benefit test if incidence is based on active surveillance of blood culture-positive and clinical typhoid fever, and the per-unit vaccine cost is less than US$ 3. Even if incidence is based only on blood culture-confirmed cases detected with active surveillance (Fig. 1), all three vaccination programmes still pass a societal cost-benefit test at a per-unit vaccine cost of US$ 3 for a correction factor of 2. The pre-school vaccination programme easily passes a societal cost-benefit test at all per-unit vaccine costs shown even if the avoided public and private cost of illness overestimates the total economic benefits (implying ?<1).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

DISCUSSION

Immunizations not only prevent mortality and morbidity, they also reduce the expenditure of public and private resources, permitting increases in consumption or investments that improve individuals' standards of living. Despite these benefits, public immunization programmes must compete with other health interventions and other sectors for resources. Economic studies of immunization programmes often demonstrate the magnitude of benefits of immunization programmes in terms of public-sector budgetary cost savings, but only a few include the private cost benefits to individuals as well.

When attempting to conduct cost-benefit analyses of typhoid vaccination programmes, analysts confront major uncertainties in a number of key input parameters. Three of the most important parameters are: (a) the incidence of typhoid fever in the study population, (b) the per-unit vaccine cost, and (c) the magnitude by which avoided costs of illness underestimate the actual economic benefits of risk reduction that vaccination provides. The results presented in this paper illustrate the sensitivity of cost-benefit calculations to changes in these key assumptions.

The findings illustrate why typhoid vaccination programmes may often appear to be unattractive to public-health officials who adopt a public budgetary perspective. Under many plausible sets of assumptions, public-sector expenditure on typhoid vaccination does not yield comparable public-sector cost savings. Of course, public-health officials need not adopt this decision criterion. If public-health officials have the financial resources to spend on typhoid vaccination and adopt a societal perspective on the economic benefits of vaccination, including not only public cost savings but private cost savings as well, there are many situations in which different vaccination programmes will make economic sense. Our findings show that this is especially true when public decision-makers recognize that (a) the incidence of typhoid fever is significantly underestimated by blood culture-positive cases and (b) avoided costs of illness represent a significant underestimate of actual economic benefits to individuals of vaccination.

On the other hand, the results presented in this paper are illustrative for slum areas with very high incidence of typhoid fever. Typhoid vaccination programmes will look much less attractive from an economic point of view in locations with lower incidence of typhoid fever. For example, our results for a slum community in New Delhi are different from those of Vollaard et al., who speculated that mass immunization against typhoid fever would not be appropriate in Jakarta, Indonesia (12).

Our cost-benefit calculations also illustrate the important differences between a public budgetary and a societal perspective on typhoid vaccination. From a public-sector budgetary perspective, under a wide array of conditions, mass vaccination and school-based vaccination programmes may appear to be unattractive even in a slum with very high incidence rates. This is, in large part, because standard policy interventions are not well-targeted from a public budgetary perspective. School vaccination misses the 2-5-year old children who impose high costs on the public-sector budget. Mass vaccinations would cover these 2-5-year old children, but spend resources vaccinating adults, who impose very low costs on the public-health budget.

ACKNOWLEDGEMENTS

This work was supported by the Disease of the Most Impoverished Programme, funded by the Bill and Melinda Gates Foundation and coordinated by the International Vaccine Institute, Seoul, Republic of Korea. The authors would like to thank Dohyeong Kim and Brian Maskery for research assistance.

REFERENCES

(1.) Hessel L, Debois H, Fletcher M, Dumas R. Experience with Salmonella typhi Vi capsular polysaccharide vaccine. Eur J Clin Microbiol Infect Dis 1999;18:609-20.

(2.) Griffin GE. Typhoid fever and childhood vaccination strategies. Lancet 1999;354:698-9.

(3.) Sinha A, Sazawal S, Kumar R, Sood S, Reddaiah VP, Singh B et al. Typhoid fever in children aged less than 5 years. Lancet 1999;354:734-7.

(4.) Bahl R, Sinha A, Poulos C, Whittington D, Sazawal S, Kumar R et al. Costs of illness due to typhoid fever in Indian urban slum community: implications for vaccination policy. J Health Popul Nutr 2004;22:304-10.

(5.) Miller MA. An assessment of the value of Haemophilus influenzae type b conjugate vaccine in Asia. Pediatr Infect Dis J 1998;17(Suppl):S152-9.

(6.) Miller MA, Redd S, Hadler S, Hinman A. Amodel to estimate the potential economic benefits of measles eradication for the United States. Vaccine 1998;16:1917-22.

(7.) Miller MA, Wenger J, Rosenstein N, Perkins B. Evaluation of meningococcal meningitis vaccination strategies for the meningitis belt in Africa. Pediatr Infect Dis J 1999;18:1051-9.

(8.) Drummond MF, O'Brien B, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programs. 2d ed. New York: Oxford University Press, 1997. 316 p.

(9.) Lieu TA, Ray GT, Black SB, Butler JC, Klein JO, Breiman RF et al. Projected cost-effectiveness of pneumococcal conjugate vaccination of healthy infants and young children. J Am Med Assoc 2000;283:1460-8.

(10.) Yang HH, Wu CG, Xie GZ, Gu QW, Wang BR, Wang LY et al. Efficacy trial of Vi polysaccharide vaccine against typhoid fever in south-western China. Bull World Health Organ 2001;79:625-31.

(11.) Harrington W, Portney PR. Valuing the benefits of health and safety regulation. J Urban Econ 1987; 22:101-12.

(12.) Vollaard AM, Ali S, van Asten HAGH, Widjaja S, Visser LG, Surjadi C et al. Risk factors for typhoid and paratyphoid fever in Jakarta, Indonesia. JAMA 2004;291:2607-15.

Christine Poulos [1], Rajiv Bahl [2], Dale Whittington [3], Maharaj K. Bhan [2], John D. Clemens [4], and Camilo J. Acosta [4]

[1] Research Triangle Institute, Research Triangle Park, North Carolina, NC 27709,

[2] Centre for Diarrhoeal Disease and Nutrition Research, Department of Paediatrics, All India Institute of Medical Sciences, New Delhi 110 029, India,

[3] Departments of Environmental Sciences & Engineering, City & Regional Planning, and Public Policy, University of North Carolina at Chapel Hill, USA, and

[4] International Vaccine Institute, Kwanak PO Box 14, Seoul 151-600, Republic of Korea

Correspondence and reprint requests should be addressed to:

Dr. Christine Poulos

Research Triangle Institute

3040 Cornwallis Road, PO Box 12194

Research Triangle Park

North Carolina, NC 27709

USA

Email: cpoulos@rti.org
Table 1. Incidence of typhoid fever, ex-ante costs of illness,
and vaccine benefits, by age group (70% effectiveness)

                           Annual
                           incidence (per
                           1,000 person-
                           years)

Age        Population of   Blood +ve
(years)    Govindpuri      typhoid from
group                      active
                           surveillance

0-2        1,123           13.6
2-5        2,011           34.9
5-19       8,370           11.7
>19        8,081           1.1
All ages   19,585          9.8

                           Blood +ve
                           typhoid from
                           active plus
                           passive
Age                        surveillance
(years)                    and 'clinical'
group                      diagnoses

0-2        1,123           24.5
2-5        2,011           53.1
5-19       8,370           19.3
>19        8,081           6
All ages   19,585          17.5

Age        Annual ex-ante costs
(years)    of illness (US$)
group
           Private         Public           Total

0-2        0.45            0.25             0.70
2-5        1.04            3.32             4.34
5-19       0.59            0.31             0.90
>19        0.06            0.00             0.06
All ages   0.42            0.48             0.90

Age
(years)
group

0-2        0.79            0.45             1.24
2-5        1.58            5.04             6.62
5-19       0.99            0.48             1.49
>19        0.31            0.03             0.34
All ages   0.79            0.45             1.24

           Per-capita vaccine benefits,
Age        discounted over 3-year
(years)    period (US$)
group
           Private         Public           Total

0-2        NA              NA               NA
2-5        1.81            5.79             7.55
5-19       1.03            0.54             1.57
>19        0.10            0.00             0.10
All ages   0.74            0.83             1.57

Age
(years)
group

0-2        NA              NA               NA
2-5        2.75            8.78             11.52
5-19       1.72            0.83             2.60
>19        0.54            0.05             0.59
All ages   1.27            1.52             2.80

           Expected
           number of
           cases of
Age        typhoid fever
(years)    prevented with
group      100% vaccine
           coverage

0-2        NA
2-5        147
5-19       206
>19        19
All ages   372

Age
(years)
group

0-2        NA
2-5        224
5-19       339
>19        102
All ages   665

NA=Indicates that there are no vaccine benefits in the 0-2-year
age group because the Vi polysaccharide vaccineis not indicated
for this age group

Table 2. Impacts and economic analyses of mass immunization
programme, school-based immunization programme, and pre-school
immunization programme by vaccine cost (based on incidence
of blood culture-positive typhoid fever from active
surveillance)

                                           Vaccine cost (US$)

Programmes and programme metrics
                                           0.75       1         1.5
Panel A: Mass immunization programme
Number of cases avoided                    297        297       297
Public cost/cases avoided (US$/case)       37.25      49.67     74.50
Public benefits (US$)/public cost (US$)    1.17       0.87      0.58
(Public cost--public benefit)/             -6.20      6.22      31.05
  cases avoided (US$/case)
Net public benefits (=public               1,844      -1,849    -9,233
  benefits--public costs) (US$)
Net societal benefits (=total              12,213     8,521     1,136
  benefits--public costs) (US$)

Panel B: School-based immunization programme

Number of cases avoided                    165        165       165
Public cost/cases avoided (US$/case)       30.53      40.70     61.05
Public benefits (US$)/public cost (US$)    0.72       0.54      0.36
(Public cost--public benefit)/             8.57       18.75     39.10
  cases avoided (US$/case)
Net public benefits (=public               -1,410     -3,084    -6,433
  benefits--public costs) (US$)
Net societal benefits (=total              5,485      3,811     463
  benefits--public costs) (US$)

Panel C: Pre-school immunization programme

Number of cases avoided                    118        118       118
Public cost/cases avoided (US$/case)       10.23      13.64     20.47
Public benefits (US$)/public cost (US$)    7.72       5.79      3.86
(Public cost--public benefit)/             -68.72     -65.31    -58.48
  cases avoided (US$/case)
Net public benefits (=public               8,102      7,700     6,896
  benefits--public costs) (US$)
Net societal benefits (=total              10,942     10,540    9,736
  benefits--public costs) (US$)

                                           Vaccine cost (US$)

Programmes and programme metrics
                                           2          3
Panel A: Mass immunization programme
Number of cases avoided                    297        297
Public cost/cases avoided (US$/case)       99.33      149.00
Public benefits (US$)/public cost (US$)    0.44       0.29
(Public cost--public benefit)/             55.88      105.55
  cases avoided (US$/case)
Net public benefits (=public               -16,618    -31,388
  benefits--public costs) (US$)
Net societal benefits (=total              -6,249     -21,018
  benefits--public costs) (US$)

Panel B: School-based immunization programme

Number of cases avoided                    165        165
Public cost/cases avoided (US$/case)       81.40      122.10
Public benefits (US$)/public cost (US$)    0.27       0.18
(Public cost_public benefit)/              59.45      100.15
  cases avoided (US$/case)
Net public benefits (=public               -9,781     -16,477
  benefits--public costs) (US$)
Net societal benefits (=total              -2,885     -9,581
  benefits--public costs) (US$)

Panel C: Pre-school immunization programme

Number of cases avoided                    118        118
Public cost/cases avoided (US$/case)       27.29      40.93
Public benefits (US$)/public cost (US$)    2.89       1.93
(Public cost_public benefit)/              -51.66     -38.02
  cases avoided (US$/case)
Net public benefits (=public               6,091      4,483
  benefits--public costs) (US$)
Net societal benefits (=total              8,931      7,323
  benefits--public costs) (US$)

Table 3. Impacts and economic analysis of mass immunization
programme, school-based immunization programme, and
pre-school immunization programme by vaccine cost
(based on incidence of blood culture-positive typhoid
fever from active plus passive surveillance
plus 'clinical' typhoid fever)

Programmes and programme benefits          Vaccine cost (US$)

                                           0.75     1         1.5

Panel A: Mass immunization programme
Number of cases avoided                    532      532       532
Public cost/cases avoided (US$/case)       20.81    27.75     41.62
Public benefits (US$)/public cost (US$)    1.81     1.36      0.90
(Public cost--public benefit)/             -16.80   -9.87     4.01
  cases avoided (US$/case)
Net public benefits (=public               8,943    5,251     -2,134
  benefits--public costs) (US$)
Net societal benefits (=total              28,669   24,977    17,592
  benefits--public costs) (US$)

Panel B: School-based immunization programme

Number of cases avoided                    271      271       271
Public cost/cases avoided (US$/case)       18.50    24.67     37.01
Public benefits (US$)/public cost (US$)    1.11     0.83      0.56
(Public cost--public benefit)/             -2.06    4.11      16.44
  cases avoided (US$/case)
Net public benefits (=public               560      -1,114    -4,462
  benefits--public costs) (US$)
Net societal benefits (=total              12,381   10,707    7,359
  benefits public costs) (US$)

Panel C: Pre-school immunization programme

Number of cases avoided                    179      179       179
Public cost/cases avoided (US$/case)       6.73     8.97      13.45
Public benefits (US$)/public cost (US$)    11.70    8.78      5.85
(Public cost--public benefit)/             -71.99   -69.75    -65.26
  cases avoided (US$/case)
Net public benefits (=public               12,915   12,512    11,708
  benefits--public costs) (US$)
Net societal benefits (=total              17,332   16,930    16,126
  benefits--public costs) (US$)

Programmes and programme benefits          Vaccine cost (US$)

                                           2        3

Panel A: Mass immunization programme
Number of cases avoided                    532      532
Public cost/cases avoided (US$/case)       55.50    83.25
Public benefits (US$)/public cost (US$)    0.68     0.45
(Public cost--public benefit)/             17.88    45.63
  cases avoided (US$/case)
Net public benefits (=public               -9,519   -24,288
  benefits--public costs) (US$)
Net societal benefits (=total              10,207   -4,562
  benefits--public costs) (US$)

Panel B: School-based immunization programme

Number of cases avoided                    271      271
Public cost/cases avoided (US$/case)       49.35    74.02
Public benefits (US$)/public cost (US$)    0.42     0.28
(Public cost--public benefit)/             28.78    53.45
  cases avoided (US$/case)
Net public benefits (=public               -7,811   -14,507
  benefits--public costs) (US$)
Net societal benefits (=total              4,011    -2,686
  benefits public costs) (US$)

Panel C: Pre-school immunization programme

Number of cases avoided                    179      179
Public cost/cases avoided (US$/case)       17.94    26.90
Public benefits (US$)/public cost (US$)    4.39     2.93
(Public cost--public benefit)/             -60.78   -51.81
  cases avoided (US$/case)
Net public benefits (=public               10,904   9,295
  benefits--public costs) (US$)
Net societal benefits (=total              15,321   13,713
  benefits--public costs) (US$)
Gale Copyright: Copyright 2004 Gale, Cengage Learning. All rights reserved.