Content area
Abstract
Governments in developed and developing countries alike provide health care. Evaluation of these services is important, as resources for subsidizing health care have shriveled since the recession of the 1980s and the ensuing international debt crisis.
This dissertation examines the impact of access to health facilities, infrastructure, and personnel on infant mortality in Indonesia. Data are from the 1987 Indonesian Demographic and Health Survey, matched with two censuses of village infrastructure conducted in 1983 and 1986.
Developing a research design for evaluation of national health programs is problematic. Mortality levels or other community characteristics influence service locations. Estimates of the effects of interventions on health outcomes will be biased if statistical approaches do not control for the nonrandom allocation of health interventions. For example, if planners place government clinics in high mortality areas, the positive impact of access to services on health will be understated.
I use an approach developed by Chamberlain (1980). Within villages, I examine changes in children's survival probabilities over time as a function of changes in access to facilities. Fixed village characteristics cannot affect survival differences across infants born in the same village, so estimates are free of bias from unmeasured village effects.
The results demonstrate that the availability of health infrastructure lowers infant mortality risks. Maternity clinics exert the strongest effects: an additional maternity clinic decreases the odds of infant death by 20%. An additional doctor decreases the odds of death by about 1%. An additional health clinic decreases the odds of death by about 2%, but the effect is not statistically significant.
The parameter estimates can be used to estimate the impact of changes in health infrastructure on the infant mortality rate in Indonesia. The results of the decomposition indicate that positive changes in access to health services account for around 14% of the decline in infant mortality between the two periods. Positive changes in maternal education account for a similar amount of the decline.
