Cost effectiveness of family planning services in Nepal

PHN Technical Notes RES 9 THE WORLD BANK COST EFFECTIVENESS OF FAMILY PLANNING SERVICES IN NEPAL December 1983 Population, Health and Nutrition Department * This paper is one of a series issued by the Population * * Health and Nutrition Department for the information and * * guidance of Bank staff working in these sectors. The * * views and opinions expressed in this paper do not neces- * * sarily reflect those of the Bank. *  RES 9 A B S T R A C T The cost effectiveness of the FP/MCH program in Nepal is analyzed using estimates of effect (births prevented) derived from an application of the population councils CONVERSE computer program. The analysis suggests that permanent sterilization compared to temporary methods is more cost effective but the analysis does not imply that high-cost methods, such as pills and condoms, should be dropped from the program. Instead the analysis reveals the need to take steps to overcome the high discontinuation rates of the various temporary methods. If continuation rates for temporary methods were substantially improved to potentially achieveable levels, the costs per birth prevented by condoms would fall from NRs 1,347 to NRs 230, and by pills from NRPs 919 to NRs 370. Reduction in costs would make it possible for an increased number of people to be reached by temporary methods with the same expenditure of funds. The analysis also provides a cost effectiveness comparison by terrain. Costs for hill and mountain districts per birth averted (range NRs 700 to NRs 1,800) are much higher than costs for the Terai (range NRs 300 Zeo NRs 900). These results suggest that economic efficiency would be best served by concentrating services on the more accessible parts of the country. Prepared by: Howard N. Barnum, PHND1 December 1983 Note: The anialysis in this paper was carried out in July 1982 as background material for Kingdom of Nepal: Report on Population Strategy (World Bank Report No. 4291-NEP). The author has benefited greatly from discussions with Graham Clarkson.  COST EFFECTIVENESS OF FAMILY PLANNING SERVICES IN NEPAL A. Introduction This paper provides estimates of the cost effectiveness of family planning services in Nepal with the objective of gaining information that can give guidance in program operations and aid in the formulation of a population strategy. To meet this objective the analysis estimates the relative cost of preventing birth through the use of alternative contraceptive methods offered by the FP/MCH project as well as estimating variations in program costs per birth prevented in different region$ of the country and in different terrains. Estimates are also given of the cost effectiveness of the distribution of contraceptives through the contraceptive retail sales. At the time of the 1981 census, the population of Nepal was 15 million and the population growth rate was 2.6%. Projections of Nepal's population predict between 23 and 29 million people by 2006 depending on whether or not a turning point in fertility can be achieved beginning in 1985. 1/ Given the already daunting problem of providing for basic needs, the achievement of a turning point in the trend of fertility has become a priority for the Nepalese Government. However, in spite of the existence of a family planning program since 1966, with significant commitments of donor and / Judith Banister and Shyam Thapa, The Population Dynamics of Nepal, (Papers of the East West Population Institute No. 78, December 1981), Honolulu, Hawaii.  -2- government resources continuing throughout the late seventies to the present time, the effectiveness of the program has remained limited. The proport on of married women of reproductive age currently using family planning is only 7%, and low continuation rates are a fundamental program deficiency. Cost effectiveness is only one element out of many that must be taken into consideration to increase the impact of the program and move closer to the critical turning point in fertility. But greater use of scarce program resources on the geographic areas and methods where the effect per unit of cost is greatest would be an important element in more rapid fertility reduction. Reduced to basic elements, cost effectiveness analysis is conceptually simple and involves division of a measure of program costs by a measure of program effects. However, data requirements for such an analysis are substantial and cannot be met without the use of a number of assumptions that reduce the precision of the results. Fortunately, substantial cost and service data are available for the fiscal year 1979/80 from the, FP/MCH project units in a fotm that permits disaggregation geographically and by contraceptive method. Nevertheless, two technical problems exist that respectively affect the denominator (births prevented) and the numerator (costs) of the cost effectiveness estimates. the first problem involves the estimation of births prevented per acceptor which requires assumptions concerning the rates of continuation with individual contraceptive methods as well as assumptions concerning duration of breast feeding,  fertility rates and other demographic parameters.. These assumptions and the procedures for estimating births prevented are given in. subsection B. The second problem, involving the disaggregation of costs among contraceptive methods, is attacked by allocating expenditures in proportion to the use of time by personnel in the delivery of different services. A discussion of time use-follows in.subsection C. A review of the findings and implications of the cost effectiveness analysis of the FP/MCH program is given in section D. The contraceptive retail sales program is considered in section E. B. Contraceptive User Data, Continuation Rates and Births Prevented Continuation rates for alternative contraceptive methods can be inferred from a comparison of annual acceptor data and the user prevalence data obtained in the 1976 NFS and the 1981 CPS. The inference is made by applying an exponential decay function to the 'Nepal acceptor data for the period 1966/67 through 1980/81 to .-. calculate the number of users for 1976 and 1981. Parameters of the decay function (which are, in effect, continuation rates) are chosen to provide rough consistency with the NFS and CPS user information. The function chosen is that used by the Population Council's family planning impact estimation programs, CONVERSE and  TABRAP. 2/ This function was chosen so that the estimated continuation rates can be used in the application of CONVERSE to estimate births prevented per acceptor of given family planning methods. The number of users, by method, estimated from the NFS and CPS data is given in columns 3 and 4 of Table 1. Continuation rates, i.e. values of a and r, that produce rough consistency with the NFS and CPS data are given in columns five and six. These implied continuation rates are extremely low for pills and condoms. In fact, the proportion of acceptors of condoms who do not use the method effectively from the start is .93 and is so high as tc render the method virtually valueless. The low continuation rates estimated by this process may be partially attributable to several sources of error. 3/ (a) undercounting of users due to use of male NFS and CPS interviewers;' (b) double counting of acceptors, especially condom acceptors because of surveillance problems; (c) failure to distinguish between old and new acceptor-s-in. FP/MCH and CHIP records; 2/ The decay function is Uts -r: As a e-r (t-s) where Uts is the number of users in time t attributable to As, the number of acceptors at time s. The parameter a is the proportion of acceptors that use the method effectively from the time of acceptance, and the parameter r is the annual rate of discontinuation. 3/ It is interesting to note that the acceptor data for sterilizations leads to an underestimate of users. This may be explained by private sources of sterilization as well as by sterilizations obtained in India by Terai residents.  -5- (d) intermittent supply of contraceptives to acceptors. However, these problems do not appear sufficient, of themselves, to explain the low implied continuation rates. The inescapable conclusion is that low continuation rates are a major problem with the Nepal FP program. Table 2 gives an estimate of the number of users calculated whern the continuation rates are improved and applied to Nepalese acceptor data. Using these rates, the number of.pill users is more than double the CPS estimates and the number of condom users is over eight times greater. Estimation of Births Prevented The continuation rates implied by consistency between acceptors and users were used in CONVERSE to estimate the number of births prevented over a ten year period by 10,000 initial acceptors of alternative methods. Considerable data, in addition to continuation rates, is used by CONVERSE in calculation of the program impact. 4/ This data includes age specific marital fertility rates, marriage rates, estimates of amenorrhea overlap, age distribution of acceptors by method and population age distrf.bution. 5/ 4/ An important omission of the CONVERSE program is an adjustment of sterilization impact by parity of acceptors. This omission may lead to an overestimate of the births prevented by sterilization. 5/ Data sources are Nepal Fertility Survey - First Report 1976 (World Fertility Survey); 1981 Nepal Contraceptive Prevalence Survey (unpublished, preliminary results); Coale-Demene, Model Life Life Tables (West); Nepal, Commission on Population, A Brief Review of Population Programmes and Activities in Nepal, 1981.  The number of births prevented by method, using the Nepalese continuation rates, are given in Table 3a. Under the present program the effect of temporary methods is very low with less than one tenth of a birth prevented for each acceptor of a temporary method. The number of births prevented by sterilizations is much higher, approximately one birth prevented per acceptor, and would be still higher if the average a6