Cost-effectiveness analysis of family planning programs in rural Bangladesh: evidence from Matlab.

The Family Planning Health Services Project in Matlab is often seen as more expensive than similar activities carried out by the government of Bangladesh. At the same time, it as been observed that the project is much more effective. The alleged high cost of the project is said to make it difficult to replicate throughout the nation. Previously, the true costs of the project had not been documented. This study systemically examines the cost of the project and assesses its cost-effectiveness. An experimental design framework is used as a basis for understanding the cost-effectiveness of the project, although a sensitivity analysis lends further support to the relative efficiency of the approach undertaken in Matlab. Although in the aggregate, the Matlab Project is more expensive than the government's family planning program, it is also more effective, generating enough output to offset the extra costs of the intensified delivery system.

PIP: The cost-effectiveness, in terms of cost per averted birth, of the Family Planning Health Services Project (FPHSP) in Matlab, Bangladesh was analyzed on 3 levels in comparison to that of the comparison area, which received the Government's family planning services. The Matlab project began in 1977, and is now renowned for its effectiveness, and quality and longevity of data. The experimental area received an intense, cafeteria-type family planning method mix (pills, tubectomy, condoms, foal, IUDs and injectables), with vaccinations, ORS and medicines by household delivery, structured horizontally, with in-depth, repeated surveys. The control area receives limited family planning methods (pills, tubectomy, vasectomy, condoms), ORS only, by vertical design. The project has yielded about 45% contraceptive prevalence of effective methods, and 6914 births averted from 1978-1985. The costing scheme is categorized into service rings denoted service delivery, supervision and administration, data management, research and international assistance and overhead. Research is included as a cost because it generates effects by a "Hawthorne effect." Costs have ranged from $133,000-164,000 yearly. Three models of cost analysis are presented and discussed from the viewpoint of sensitivity analysis. The estimated cost per birth prevented ranges from $150-220, figures that do not account for reduced mortality or improved reproductive health. The Matlab project generates about 3 times as many services as the government program. It costs more overall, but less per birth averted than the control government program. This suggests that the government program may benefit by offering a wider choice of contraceptive methods.

Transferring health and family planning service innovations to the public sector: an experiment in organization development in Bangladesh.

The International Centre for Diarrhoeal Disease Research, Bangladesh, has launched a field experiment in two rural thanas of Bangladesh to test the transferability of its successful health and family planning experiment in Matlab to the Ministry of Health and Population Control service system. This paper reviews the Matlab experiment with particular attention to its organization and identifies elements for transfer. The intervention strategy and operations research design of the new experiment are discussed. The proposed design follows an organization development strategy in which collaborative diagnostic research is used to foster institutional change.

PIP: The International Center for Diarrheal Disease Research, Bangladesh has launched a field experiment in 2 rural thanas of Bangladesh to test the transferability of its successful health and family planning experiment in Matlab to the Ministry of Health and Population Control service system. The 1st family planning experiment began in 1975 and concentrated on the household distribution of contraceptives. It was further designed to provide a broad range of contraceptive and immunization services and oral rehydration therapy. Strong management control is vested in the non-medical, male supervisory staff. Maternal and child health services were later combined with the family planning project. This extension project has 2 components: an intervention strategy and a research strategy to assess the efficacy of the program. The analysis of the effects of this extension program is achieved through the longitudinal observation of households with a sample registration system. The proposed design follows an organization development strategy in which collaborative diagnostic research is used to foster institutional change. The original Matlab experiment posited that a significant proportion of clients desire to limit or space childbearing, but lack contraceptive services for doing so. The experiment demonstrated that there is a set of conditions under which a significant proportion of a rural Bangladeshi population will use contraception, and thereby reduce fertility. In transferring the programs to the supervision of the Ministry of Health, a number of modes were used: 1) use of a team appraoch, especially on household distribution; 2) train more community workers, especially females, and use teams of both males and females; 3) develop a Tertiary Health Center referral system, especially for sterilization; 4) enlarge training of present workers to include better preventive and MCH care; and 5) arrange monthly meetings of union-level workers at the field centers. No change in salary structure or administrative structure was planned.

Post-neonatal mortality in rural India: implications of an economic model.

In this paper we develop and test a theory of childhood mortality after the first month of life. Parents are assumed to have well-defined family size and sex composition objectives and to face severe budget constraints. In this set of circumstances, it is understandable that they will make allocative decisions that will affect the survival probabilities of children. These decisions and the environmental influences on mortality are the basic forces which determine whether a child will survive through the post-neonatal period. The model is tested with survey data from rural Uttar Pradesh, India. The results are consistent with the hypothetical framework discussed above. The burden of this pattern of choice is felt particularly strongly by female births.

The educational status of parents, and infant and child mortality in rural North India.

PIP: The task of this discussion is to disentangle the various forces which influence the probability of death for children and, where possible, to identify the independent role of education. This discussion is one of a series of explorations of the mortality and fertility experiences of a sample of couples from North India. The analysis is based on interviews with a rural sample of husbands and wives living in 120 villages in Uttar Pradesh State in India. Information concerning fertility and mortality is based on the retrospective fertility histories collected in 1972. The population is heavily dependent upon agriculture, and it is poor and has little formal education. Per capita income at the time of the survey was less than $100/year, and 52% of the husbands and 87% of the wives had never been to school. The crude birth rate is over 50/1000 and the total fertility rate is 7.7. The overall level of infant mortality in the uncorrected data is on the order of 170/1000 births. For the 2nd and 3rd years of life the mortality rate is 74/1000 children surviving the 1st year. There are considerable differences in the mortality experience of children according to the educational attainment of their parents. In general, the mortality differentials between those with no education and those with a moderate amount of education are small, but for all 3 measures of education, children with the most educated parents experience substantially lower levels of mortality. Those households where the mother has had some education tend to have lower female than male mortality rates, i.e., children of both sexes born into those households experience lower levels of mortality than do children born into households where neither parent or only the father is educated, but female children do particularly well. A clear measure of association between education and mortality is shown, but these measures are suggestive. To examine further the role of education as a determinant of mortality, the multivariate correlates of mortality in the neonatal and postneonatal periods are examined separately. In each case the determinants of death are examined in the context of a complex model of behavior. The regressions show a reasonably high degree of explanation of mortality during the 1st month of life, but education is not a significant explanatory variable in any of the regressions. In the multivariate context the influence of education in the postneonatal period is restricted to female births. Those girls born to households where the mother has received some education are significantly less likely to die than are the counterparts born to families where neither parent is educated or just the father has attended school. In sum, education is not a statistically significant predictor for the mortality probability of males in either age group, but it is a significant and robust predictor of female mortality in the postneonatal period.^ieng

Neo-natal mortality in South Asia: The special role of tetanus.

Summary First-year mortality in rural Uttar Pradesh is characterized by a predominance (60 per cent) of deaths during the first month of life, of which 66 per cent are reported to be due to tetanus. This pattern is not typical of the historical experience of many developed countries and the current experience of some less developed countries where post-neo-natal mortality predominates. To examine this phenomenon, two causal models of neo-natal mortality (one for tetanus and one for all other diseases) are developed and tested using retrospective survey data from 2000 couples living in rural Uttar Pradesh. Neo-natal tetanus mortality is found to be primarily a function of opportunities for exposure to the disease (e.g. lack of antiseptic birth practices, ownership of large animals) rather than of socio-economic status or demographic variables. The importance of examining neo-natal mortality by cause, and the shortcomings inherent in making inferences from the historical experiences of Western nations are emphasized.