Cost-effectiveness of an immunization programme in Indonesia

Bulletin of the World Health Organization, 58 (3): 499-503 (1980) Cost-effectiveness of an immunization programme in Indonesia H. N. BARNUM, 1 D. TARANTOLA,2 & I. F. SETIADY 3 The economic analysis reported below, based on hypothetical estimates of the programme impact, indicates that an expandedprogramme ofimmunizationfor diphtheria, pertussis, tetanus, and tuberculosis can be expected to be highly cost-effective in comparison with treatment. Sensitivity tests illustrate that this conclusion remains valid even when costs are increased by 20% and benefits reduced by 50%. A separate analysis was made of the DPT-tetanus toxoid and BCG components of the programme. The analysis revealed that although the BCG programme may not be justifiable when operated independently, its inclusion in a joint immunization programme is strongly justifiable on economic grounds (assuming a vaccine efficacy of 0.5). This result confirms one of the basic arguments advancedfor the WHO programmes of expanded immunization and illustrates that other immunizations, such as for poliomyelitis and measles, which may not be cost-efficient by themselves may be economically justifiable when included as part ofa larger immunization programme. The analysis presented here is intended to give a conservative assessment of the cost-effectiveness of the proposed expanded programme on immunization (EPI) in Indonesia. The objective of the programme is to reduce infant and child morbidity and mortality by expanding the present national programme of smallpox and BCG immunization to cover a larger proportion of the population and to include immu- nization for diphtheria, pertussis, and tetanus. The programme entails the delivery of BCG' and DPT vaccinations in two visits during the first year of life, and tetanus toxoid (T) vaccinations of pregnant women. In addition BCG vaccinations are to be given to children entering and leaving school.b The planning horizon for the project, and the period for which the economic analysis is conducted, is five years commencing in the fiscal year 1979-80. The estimates of morbidity and mortality impact are therefore conservative since, if the programme is continued more than five years, DT boosters for chil- dren entering school would give extended immunity that is not included in the calculations. In addition, the estimates include only the direct effects of immu- I Research Scientist, Center for Research on Economic Develop- ment, University of Michigan, Ann Arbor, MI 48109, USA. 2 Medical Officer, WHO Expanded Programme on Immuniz- ation, Indonesia. 3 Director of Epidemiology and Quarantine Services, Directorate General for Communicable Disease Control, Ministry of Health, Indonesia. a Referred to as BCG I. b Referred to as BCG II. nization and omit the effects of extended community protection as a result of reduced transmission. The estimates are also conservative since, to simplify our calculations, we have omitted the effect of tetanus immunization on maternal morbidity and mortality and have also omitted the impact of the BCG immunization of children when they leave school. The effect of these omissions is probably negligible since the incidence of tetanus in women of childbearing age is low except perhaps for postpartum tetanus associated with induced abortions, and the number of school leavers is also small. Both the economic analysis and the epidemiological framework are relatively simple and ignore a number of complications that might be included in a more extensive analysis. But the procedure used has the advantage of providing cost-effectiveness figures based on the data that are commonly available in developing countries. Given the paucity of data available, a number of assumptions have been made to allow the estimation of the hypothetical impact of the immunization pro- gramme. In some cases we have extrapolated from data available for specific regions of Java to the entire country. Also, hospital record surveys have been used in combination with other information in an attempt to reconstruct incidence and fatality rates. The diffi- culties in conducting this analysis with the available data are likely to be common to many countries and demonstrate the strong need for a careful surveillance system to be incorporated into the early stages of an immunization project. Early surveillance is imperative 3970 - 499 - 500 H. N. BARNUM ET AL. in order to allow a future retrospective assessment of the programme. COST-EFFECTIVENESS ANALYSIS We are attempting to ascertain whether at the levels of incidence in Indonesia it is cheaper to treat certain infections or to prevent infection through an immuni- zation programme. The estimates are obtained by tracking immunized cohorts over the period of efficacy of the immunization under consideration and accounting for survival and the number of cases and deaths that would have occurred in the absence of immunization. Because present disease surveillance is insufficient to provide accurate estimates of initial disease incidence (attack rates) and case-fatality rates, Table 1 is based on a variety of sources and assumptions including the use of rates observed in other tropical countries with similar population den- sities. An effort has been made to choose assumptions that will provide conservative estimates of the impact of the immunization programme.C Table 2 sum- c The methods used to calculate the number given in Table I and the sources of the basic data are given in an appendix which can be obtained from the authors. marizes the assumptions underlying the analysis. The estimates given in Table 1 indicate that a total of 255 878 deaths would be prevented by the five-year programme. Programme costs The programme costs are given in Table 3. The costs of operating the BCG and DPT + tetanus toxoid (DPTT) programmes independently were obtained by estimating the portion of total expenditures (taken from the financial projections for the EPI project made by the Indonesian Ministry of Health) that would be necessary to operate the programmes separately. For example, the total cost of the cold chain and transport categories is included in both programmes, but the costs of vaccine, vaccine produc- tion development, and syringes and other equipment are broken into components directly attributable to the separate programmes. The total costs of the programme have been adjusted by subtracting the value of capital remaining at the end of the five-year planning period. Straight- line depreciation is used. A life of ten years has been assumed for all capital equipment, except for vehicles, which are assumed to have a life of five years. Table 1. Hypothetical number of cases and deaths prevented by the immunization programme DPTT BCG Total Years Cases Deaths Cases Deaths Cases Deaths 1979-80 119738 14468 1 670 376 121 408 14844 1980 - 81 253 320 23 142 3 439 773 256 759 23 915 1981 - 82 421 914 36881 5 358 1 205 427 272 38 087 1982 - 83 598 020 47 668 7 442 1 675 605 462 49 343 1983 - 84 776 045 58 069 9 549 2 150 785 594 60 219 1984 - 85 508 457 161 817 11 368 2 556 519 825 19 373 1985-86 321 677 11 656 13742 3093 335419 14749 1986 - 87 224 376 7 273 16 371 3 682 240 747 10 955 1987 - 88 120 226 4 060 19 269 4 336 139 495 8 396 1988 - 89 22 312 5 019 22 312 5 019 1989 - 90 18 896 4 252 18 896 4 252 1990 - 91 14 688 3 305 14 688 3 305 1991 - 92 10 102 2 272 10 102 2 272 1992-93 5109 1 149 5109 1 149 Total 3 343 773 220 034 159 315 35 844 3 503 088 255 878 COST-EFFECTIVENESS OF IMMUNIZATION IN INDONESIA 501 Table 2. Summary of assumptions underlying the calculation of cases and deaths prevented 1. Assumptions related to year of programme Programme coverage in Programme effectivenessC total population b DPTT BCG BCG II Diphtheria Pertussis Tetanus BCG BCG 1I Number of Year of children 0-1 1-5 programme born a years years d 1979 - 80 5 800 000 0.20 0.59 0.47 0.190 0.05 0.10 0.190 0.295 0.235 1980 - 81 5 920 000 0.28 0.62 0.50 0.266 0.07 0.14 0.266 0.310 0.250 1981 - 82 6 040 000 0.42 0.67 0.54 0.399 0.105 0.21 0.399 0.335 0.270 1982 - 83 6 160 000 0.50 0.73 0.58 0.475 0.125 0.25 0.475 0.365 0.290 1983-84 6280000 0.57 0.73 0.59 0.542 0.1425 0.285 0.542 0.365 0.295 2. Assumptions related to agee Probability Attack rates Case fatality rates Age of surviving (years) one year Diphtheria Pertussis Tetanus Tuberculosis Diphtheria Pertussis Tetanus Tuberculosis 0 0.860 0.0006 0.40 0.0175 0.00055 0.222 0.0275 0.60 0.225 1 0.968 0.0020 0.16 0.0029 0.00055 0.204 0.019 0.20 0.225 2 0.968 0.0009 0.08 0.0051 0.00055 0.176 0.019 0.20 0.225 3 0.968 0.0010 0.08 0.0022 0.00055 0.143 0.019 0.20 0.225 4 0.968 0.0012 0.08 0.0032 0.00055 0.093 0.019 0.20 0.225 a Assumes a 2% population growth rate. b As planned by programme administrators, Indonesian Ministry of Health. c Programme effectiveness equals vaccine efficacy multiplied by programme coverage. Vaccine efficacy is assumed to be 0.95 for diphtheria, 0.50 for pertussis, 0.95 for tetanus, and 0.50 for BCG. d It is assumed that only 0.50 of all newborn children will be protected by the age of six months. The programme effectiveness for the first year is therefore one-half of the effectiveness in the 1-5-year age group. e The table is truncated at five years to save space. The actual table used was extended to age 15 years for tuberculosis. Rates are based on hospital records, morbidity surveys, and judgment. The attack rate for tuberculosis increased for age groups above 5 years from 0.00085 at age 5 to 0.0029 at age 15. The total cost ofthe complete programme after sub- tracting the residual value of capital at the end of the five-year planning period would be 13 808 x 106 Rp. The cost per contact, including fixed and recurrent costs, is 272 Rp (US$ 0.66) and the cost per fully vacci- natedd infant is 1532 Rp (US$ 3.69). The cost per fully vaccinated infant after the programme has reached full operating capacity in its last year is 1397 Rp (US$ 3.37). d The total number of contacts for DPT (2 visits), BCG, and T is 50 760 000. The US S costs given in parentheses were calculated assuming 415 Rp = I US$. This was the exchange rate in May 1978. Adjusting for children who do not return for the second immuniz- ation, the predicted number of children fully immunized with DPT and BCG and whose mothers have, if required, received T is 9 105 825. Treatment costs Treatment costs represent the costs of inpatient care for tetanus and diphtheria and outpatient care for tuberculosis and pertussis. The costs of inpatient care were obtained by multiplying the average length of stay for tetanus and diphtheria by the average daily coste of paediatric hospital care for sick children. The costs include drugs, food, nursing, laboratory tests e An average length of stay of 13.2 days for tetanus and 10.7 days for diphtheria was used. An initial estimate was obtained for the cost per patient day of operating sick child paediatric wards in a Class A and two Class B hospitals. Since the majority of hospitals are of class C and D and have substantially lower costs, the estimates were revised downwards by 30%. The final estimate of the average daily paediatric costs for all classes of hospital is S 600 Rp. 502 H. N. BARNUM ET AL. Table 3. Cost breakdown by programme and year of programme (rupiah x 106) Cost of Cost of Cost of Cost if Cost if adding adding total BCG DPTT DPTTto BCG to pro- Year only8 onlya BCGb DPTTb gramme (1) (2) (3) (4) (5) 1979 - 80 1388 2268 1153 273 2541 1980- 81 1457 2384 1197 270 2654 1981 - 82 1531 2641 1412 302 2943 1982 - 83 1785 3052 1594 327 3379 1983 - 84 1817 3267 1806 356 3623C Total 7978 13 612 7162 1528 15 140 Value of capital remaining at end of 5yearsd 1217 1311 115 21 1332 Total costs less value of capital at end of programme 6761 12 301 7047 1507 13 808 8 This represents the cost of operating a BCG programme without a DPTT programme (or conversely for column (2)). The costs of operating the two programmes separately do not add up to the costs of the total programme since many of the total costs are for shared expenditures. b This represents the added cost if a DPTT programme is considered as additional to a BCG programme (or conversely for column (4)). c Total 1983-84 costs, corrected for depreciation on capital acquired in previous years and the salvage value of capital acquired in 1983-84 are 3737 x 10' Rp. This figure is the basis for the calcu- lations of the cost-effectiveness of the programme in the final year of the plan. d This is the depreciated value of the cold chain, transport, invest- ment in vaccine production, and administrative equipment after deducting for maintenance. and doctors' salaries. The costs of outpatient care were the costs of specific treatment regimes for pertussis and tuberculosis! The resulting estimates of costs are 45 000 Rp per case for diphtheria, 6000 Rp per case for pertussis, 55 000 Rp per case for tetanus, and 49 000 Rp per case for tuberculosis. The cost of treating one case prevented by the DPTT programme is a weighted average of the costs of treatment for each of the three diseases involved in the immunization. Similarly the cost of treating one case prevented by the total programme is a weighted average of the costs of treatment for each of the four diseases. Results Results of the cost-effectiveness analysis are sum- marized in Table 4. It is apparent that, given the hypo- f The regimes and prices were outlined by Dr Guno Wiseso, Indonesian Ministry of Health. Table 4. Cost-effectiveness of the proposed immunization programme Cost Cost Cost per death per case of treating prevented prevented one casea (Rp) (Rp) (Rp) Total programme 53 963 3 942 13 166 (US$ 130) BCG programme only 188 623 42 438 49 000 (US$ 455) DPTT programme only 55905 3 679 11 459 (US$ 135) BCG considered as an 42 043 9 459 49 000 added programme (US$ 101) DPTT considered as an 32 027 2 107 11 459 added programme (US$ 77) Total programme in 50 750 3 620 13 166 last year of plan (US$ 122) (1983-84) 8 This is the weighted average of the costs of treatment of the diseases considered. The weights are the proportions of total pre- vented cases in each disease category. thetical impact based on the assumptions in Table 2, the total immunization programme would be highly cost-effective. The cost per death prevented is 53 963 Rp (US$ 130). The cost of preventing an aver- age case is approximately one-third of-the cost of treatment. The programme remains cost-effective even if we consider that programme costs have been underestimated by, say, 20%7o and treatment costs overestimated by, say, 330%o. In addition, the pro- gramme gains in efficiency over the five years of the project plan until the cost per death prevented falls to 50 750 Rp (US$ 122) and the cost per case prevented falls to 3620 Rp (US$ 8.72) when full operating capacity is reached in the final year (1983-84). Turning to the individual programme components, it can be seen that the DPTT programme would be cost-effective if operated alone. Further, if we con- sider it as an addition to an existing BCG programme already possessing an adequate cold chain and delivery system, the cost per additional case prevented is 2107 Rp. In contrast, the BCG programme is not as cost- effective when operated independently. In fact, if pro- gramme costs are increased by 20%o and treatment costs overestimated by 33%, the ratio of cost per case prevented to the cost of treatment is 1.6 and an independent BCG programme would not be cost- effective. However, if we consider the cost-effective- ness of adding BCG immunization to a DPTT pro- gramme, we see that the marginal cost per death pre- vented is only 42 043 Rp (US$ 101) and the cost per COST-EFFECTIVENESS OF IMMUNIZATION IN INDONESIA additional case prevented is substantially less than the cost of treatment. Thus, it is found that a programme that is inefficient by itself becomes cost-effective as one component of a larger immunization programme. This result suggests that once the cold-chain delivery system is in place, the addition of other immunization programmes, for example, against poliomyelitis and measles, which may not be cost justifiable when operated separately, may be warranted on a cost- effectiveness basis. ACKNOWLEDGEMENTS We wish to acknowledge the contributions of Dr Guno Wiseso, Dr Noto Abiprojo, Dr G. Ferster, Mr R. Hogan, Dr J. Keja, Dr H. Lopez, and Mr Sayuti. We also appreciate the support of the Government of Indonesia, the United States Agency for International Development, and the World Health Organization. Support does not imply agreement with the views and interpretations in this paper, which remain the sole responsibility of the authors. RESUMt RAPPORT COOT/EFFICACITL POUR UN PROGRAMME DE VACCINATION EN INDONtSIE L'analyse 6conomique presentee dans cet article avait pour objet une evaluation aussi prudente que possible du rapport coQt/efficacit6 du programme elargi de vaccination projete en Indonesie. Un tel programme aurait pour effet de reduire la morbidite et la mortalite infantiles et enfantines grace a l'extension du programme national actuel de vaccination anti-variolique et par le BCG de maniere A couvrir une plus forte proportion de la population, et A l'adjonction A ce programme de la vaccination triple contre la diphterie, le tetanos et la coqueluche (DTC). Le programme comporterait l'administration de vaccins BCG et DTC A la faveur de deux visites au cours de la premiere annee de vie, ainsi que la vaccination des femmes enceintes par l'anatoxine tetanique. Des vaccinations par le BCG seraient en outre pratiquees sur les enfants entrant a l'ecole ou quittant celle-ci. Les estimations auxquelles ont procede les analystes, fondees sur des hypotheses relatives aux effets du pro- gramme, permettent de conclure qu'un programme elargi de vaccination contre la diphterie, le tetanos, la coqueluche et la tuberculose peut etre considere comme presentant un rapport cofit/efficacit6 beaucoup plus favorable que ce ne serait le cas pour le traitement de ces maladies. Des tests de sensibilite montrent que cette conclusion demeure valable meme si le coOit s'accroissait de 20% et si les avantages etaient reduits de 50%. Une analyse separee a 6t6 faite pour les divers elements du programme: DTC, anatoxine tetanique et BCG. Bien que le programme de vaccination par le BCG puisse n'etre pas justifi6 s'il est execute isolement, son inclusion dans un programme de vaccination combine est fortement justifi6e sur le plan &conomique (en presumant une efficacite vaccinale de 0,5). Ce resultat confirme un des arguments principaux avances A l'appui des programmes elargis de vaccination de l'OMS et illustre le fait que d'autres vaccinations, contre la poliomyelite et la rougeole par exemple, dont le rapport coOt/efficacit6 peut paraitre defavorable, peuvent etre tres valables sur le plan economique lorsqu'elles font partie d'un programme plus vaste de vaccination. 503