Effect of infant and child mortality on fertility in Kenya

Abstract

The specific objective of this study was to assess whether behavioural effects on fertility of replacement and insurance following an infant or child death exist in Kenya independent of biological effects. A further general objective was to find out whether the measured values of such effects depended on the methods utilized for analysis. To achieve the above objectives, several hypotheses to determine the magnitudes of these effects were tested by applying the Ordinary Least Square (OLS) and the Hazards models to data drawn from the Kenya Demographic and Health Survey (KDHS) which was conducted in 1989. The above hypotheses were tested within a conceptual framework developed from the Bongaart and Mosley and Chen frameworks for the analysis of fertility and infant and child mortality respectively. To operationalize this conceptual framework the birth interval was taken as the dependent variable and also a measure of fertility while seven variables were defined for the assessment of the effects. Several socio-economic, cultural and demographic variables were included in this framework as control variables. Results of the analysis are presented and discussed in Chapters 6 (bivariate) and 7 (multivariate) of the Thesis. The results in chapter 6 showed that infant and child deaths were associated with shorter birth intervals in case of the analysis based on the OLS and higher risks of birth interval termination in the case of the hazards models. For example, birth intervals were shorter on the average by about 4 months for deaths of female births taking place in those intervals where the birth opening the interval was alive at the time of the survey. In terms of the risk, it was 1: 5 higher in case of such deaths compared to when there were no such deaths. The assessment of the replacement and insurance effects undertaken in Chapter 7 showed that the shorter birth intervals associated with infant and child deaths were in part explained by differences in coital frequency, contraceptive use and breastfeeding among the two groups of mothers. These results suggest the existence of replacement and insurance effects. Coital frequency and contraceptive use were found to explain about 10 per cent of the differences while breastfeeding explained about 34 per cent. The results obtained from the two method were also found to be consistent. The main conclusion which is drawn is that the findings of the study are consistent with the existence of replacement and insurance effects. Further research is, however needed before firm conclusions can be drawn. The main policy implications of the findings of the study is that infant and child survival programmes should be integrated as part of the overall strategy for the increased acceptance and use of family planning in areas where infant and child mortality is high in Kenya. Emphasis of the role of breastfeeding may be an additional strategy for increasing infant and child survival and lowering fertility.