The village midwife program and infant mortality in Indonesia

Abstract

Indonesia introduced over 50,000 midwives into villages in the 1990s to provide primary care to women lacking easy access to health facilities. It seems plausible to argue that the significant reduction in infant mortality that occurred from about 1993–94 was a consequence of this. The paper estimates the village midwife program's impact on infant mortality, using data from the Indonesia Family Life Survey. Regressing mortality outcomes against choice of services would lead to biased estimates because of the correlation between service choice and unobserved individual characteristics. Furthermore, non-random placement of midwives could bias estimates of their impact on infant mortality. This study overcomes such endogeneity problems by aggregating mortality outcomes and program prevalence at district level and taking account of district fixed effects in estimating the program's impact. Surprisingly, the results do not support the hypothesis that the midwife program was responsible for the observed decline in infant mortality.

Notes

¹The infant mortality rate is defined as the number of deaths before the age of one per 1,000 live births. It is the sum of the neonatal mortality rate, defined as deaths during the first month per 1,000 live births, and the post-neonatal mortality rate, defined as the number of deaths per 1,000 live births after the first month and before the end of the first year.

²The terms ‘village’ and ‘community’ are used interchangeably here, although some of the communities in the sample were actually in urban areas.

³Indonesia Family Life Survey, <http://www.rand.org/labor/FLS/IFLS/>.

⁴The raw data, including number of live births and standard errors for each mortality rate in each year, are contained in appendix 1.

⁵A degree of caution is needed in interpreting the figures for some of the providers presented on the table, as they are based on a small sample of births. This leads to a lack of precision and high variability in the estimates of the mortality rates.

⁶For urban areas, the aggregation is at the level of the kota (municipality), the urban equivalent of the kabupaten. For simplicity, I refer to all such aggregations as being at the kabupaten level.

⁷The term ‘parity’ is used in demographic and medical literature to indicate the number of times a woman has given birth. Parity 1 refers to the first birth, parity 2 to the second, and so on. High-parity births of order 7 and over are associated with higher infant mortality risks (Hobcraft, McDonald and Rutstein 1985).