Family Planning and Women’s and Children’s Health: Long-Term Consequences of an Outreach Program in Matlab, Bangladesh

Maternal and child health activities were introduced in blocks A and C between 1981 and 1985. After 1985, they were also extended into blocks B and D. We were unable to find significant differences in infant or under-5 child mortality in A and C versus B and D villages in these years. Other research has found that measles declined more rapidly in the entire MCH-FP program areas than in the comparison areas, although perinatal mortality did not decline in the early period of 1979–1982 (Fauveau et al. 1990; Koenig et al. 1990; LeGrand and Phillips 1996). There may also have been a more rapid decline in maternal mortality in treatment villages, although significant variation in this more rare event was difficult to estimate until recently (Fauveau 1994; Koenig et al. 1988; Rahman et al. 2009).

The MHSS is a collaborative effort distributed by the Inter-University Consortium for Political and Social Research (ICPSR) at the University of Michigan and Rand (http://​rand.​org/​labor/​FLS/​MHSS.​html).

An important caveat here is that the original resident population in 1977 may not be represented in the 1996 MHSS. Female migration and mortality could change the composition of the population observed in 1996 in treatment and control villages. When dummy variables are added to the fertility or other outcome regressions that are equal to 1 only if the woman moved after marriage into the DSS area, or moved from program to comparison areas, or vice versa, these dummy variables are never statistically significant as explanatory variables in the fertility or family outcomes studied here. More people, however, migrated out of the comparison areas than the program areas after 1996, possibly because of the higher fertility in comparison areas (ICDDR,B 2004).

Many individuals in the 1996 MHSS in the representative strata 1 and 2 cannot be matched to a weight in the Rand data file: roughly one-fourth of the adults aged 15 and older and a somewhat larger share of the children aged 6–14. To see if the characteristics of those matched to a weight differed, we assigned the unweighted individuals the average weight in the matched sample, which was .53. The population means for the villages in the treatment and comparison areas did not change appreciably, and the differences were very similar to those reported in Table 2, panel B. In other words, of the variables examined in Table 2, only the child schooling and Muslim variables were significantly different between the treatment and comparison areas in 1996, whether the full “representative” sample or the sample matched to a Rand sample weight is included.

ADLEq0 = (1.0 – ADLscore). A woman’s self-reported ability to perform five activities of daily living, drawn from section GH2 of the MHSS, are aggregated into a score: (a) walk for 1 mile; (b) carry a heavy load for 20 m; (c) draw a pail of water from a tube well; (d) stand up from a sitting position without help; (e) use a ladder to climb to a storage place that is at least 5 feet high. Responses were coded either as “can perform the task easily” (a value of 1), “can do it with difficulty,” (a value of 2) and “unable to perform” (a value of 3). This ADL index is normalized following Stewart and Ware (1992).

The functional form that the diffusion of health knowledge takes is not well established in the empirical literature. Alternative specifications of this spillover effect were explored, but none we tried provided a better fit to the data on children ever born, child mortality to age 5, or other family outcomes or use of preventive health inputs. For example, the spillover effect might be an inverse function of the estimated distance between the control village and the nearest program village, a quadratic function of this intervillage distance, or a step function determined by the number of program villages adjacent to a specific control village, which ranges in Matlab from 0 to 4. One randomized study of spillovers (externalities resulting from reducing intestinal worms in a school cohort) estimated the logarithmic impacts on a school cohort as a function of the log of number of treated persons in that cohort in a geographic area and the log of the total cohort in that geographic area (Miguel and Kremer 2004). It is unclear why this specification was adopted, although it provides a basis to decompose the effects of the intervention operating through population density and the density of treatment in the area.

If program services complement or reinforce the fertility-reducing effect of women’s education, we would expect, other things being equal, fertility differences by women’s education to increase in program areas. Where program services substitute for women’s education, fertility differences by education are expected to diminish.

In an earlier version of this article, we included whether the village had a BRAC office, a microfinance NGO that also encourages family planning and female self-employment, in 1996. It was not significantly partially associated with fertility in these weighted regressions, but these institutions entered Matlab only toward the end of the study period and may have purposefully started offices in villages where the status of women was relatively low, violating our assumption that the village infrastructure was initially independent of household behavior.

The lower level of child mortality among women over age 65 could not have credibly been caused by the program. Other family interactions involving gender, birth spacing, mortality, and child schooling are explored in other studies of Matlab (e.g., Fauveau et al. 1991; Foster and Roy 2000; Joshi and Schultz 2007; Muhuri and Menken 1997; Muhuri and Preston 1991; Rahman et al. 2004; Sinha 2005). Their potential endogeneity at the family level makes their inclusion here as control variables more complicated.

When the samples of boys and girls are stacked, allowing all estimated coefficients to be sex-specific, the estimated program effects are not statistically significantly different for boys and girls when either the odds ratio or conditional marginal effects are used (not reported) because the estimated effects for boys are often of a similar magnitude to those of girls and are estimated less precisely than for girls.

Anthropometric data are undoubtedly reported with error (WHO 1995, 2006). In an effort to trim outliers, which are likely to embody errors, we initially dropped from the estimation sample women with BMI less than 10 or greater than 30. Adding back these 35 women with outlier values (to 5,273) has only a slight effect on the estimates of the program impact on BMI and no notable effect on the likelihood that BMI exceeds 18.

In previous work, we examined the difference in inoculations for neonatal tetanus, which was a serious health risk in Matlab at this time (Joshi and Schultz 2007). We omit this indicator here, however, because programs other than the MCH-FP prescribed tetanus toxoid inoculations, particularly as a placebo in cholera vaccine trials in the 1970s and later in government clinics in the 1990s (Fauveau 1994; LeGrand and Phillips 1996).

Barham (2009) explored the consequences of the program’s promotion of maternal and child health inputs in various blocks of villages from 1982 to 1986 for cognitive functioning of adolescents in the 1996 MHSS. We did not find significant differences in fertility or child mortality from 1981 to 1985 between these blocks (A and C; and B and D) when there were differences in program priorities in child and maternal health.

The estimate assumes that the differences in general fertility rates between the program and comparison areas are attributable to program expenditures (Simmons et al. 1991).

The cost per averted birth of the government program in the comparison areas was estimated to be $298 (Simmons et al. 1991). Obtaining this estimate is complicated by the lack of an obvious control population. Moreover, many features of the MCH-FP and the Bangladesh Government program differed, including not only the outreach design, but the systems of oversight and personnel tenure and compensation. It is possible that the government program was withdrawn from Matlab program villages (LeGrand and Phillips 1996). Fauveau (1994) revised the cost accounting of the program for the period 1986–1989 and concluded that the program expenditure per prevented birth was $60, substantially lower than previously estimated.