| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
No. 67, 1994 Mensch, Barbara, Mary Arends-Kuenning, and Anrudh Jain. "Assessing the impact of the quality of family planning services on contraceptive use in Peru: A case study linking situation analysis data to the DHS," Policy Research Division Working Paper no. 67. New York: Population Council. Abstract This study explored whether current contraceptive use in Peru is affected by the family planning (FP) service environment where the a woman resides. Service-delivery data from the 1992 Situation Analysis Survey (known by its Spanish name with the acronym EDIS) were linked by cluster number to behavioral and demographic data from the 1991–92 Demographic and Health Survey (DHS) individual survey in Peru to produce a woman-level file with individual, household, and cluster variables. The EDIS is a new type of survey of family planning (FP) service delivery points which involves an inventory of facilities, exit interviews with staff and clients, and observations of client/provider interactions. The EDIS provided the "supply" variables. For this analysis, DHS data were used for women who were married or in a consensual union (a weighted sample of 7,841 women). These data supplied the dependent variable, the current use of contraceptives, as well as a set of variables controlling for factors believed to affect a women's decision to control her fertility (demand variables). Analysis of these data revealed that, net of personal and household characteristics, a significant, although small, effect of service quality on contraceptive use does exist. Availability of services had no effect, but since quality is conditional on having services available, the two variables are highly correlated and difficult to distinguish statistically. As expected, quality of care is less important than standard socioeconomic variables in explaining differences in contraceptive use. It was noted that an upper limit exists on the percentage of variation that can be explained by variables measured at the cluster level, as is quality of care. This point must be remembered to avoid undervaluing services when evaluating their impact on reproductive behavior. Finally, it was predicted that contraceptive prevalence would increase 5 percent if all women lived in a cluster with the highest quality of care. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
