The Impact of Individual Characteristics, Community Resources and Maternal Preventive Health Behavior on Infant Mortality in Rural Kenya

Abstract

The determinants of infant mortality in rural Kenya were examined. Micro-level variables for maternal preventive health behavior and macro-level variables for individual and household characteristics and district level resources were used to explain infant mortality. The theoretical model was based on the studies of Mosley and Chen. Exogenous factors were identified as personal, household, and community resources, household environment, and biological or demographic factors. Endogenous factors were identified as tetanus immunization during pregnancy, modern prenatal care, and professional birth attendance at delivery. Data were obtained from the 1988-89 Kenya Demographic and Health Survey, the Inventory of Rural Health Facilities, and the Kenya Expanded Immunization Program. The contextual analysis was conducted in a single logistic regression model as suggested by Iverson and Boyd. Bivariate analysis revealed that women with more education are more likely to use health services. Reporting of preventive health behavior was more likely among women whose husbands have nonagricultural earnings. Tetanus immunization and professional birth attendants were less likely to be used where there was competition for maternal time, such as in obtaining water. Association membership and household wealth and quality of drinking water, but not marital status, were significantly positively related to the three health behaviors. Transportation availability was significantly related to maternal tetanus immunization. Type of toilet facilities, maternal education, and media awareness were significantly positively related to the three health behaviors. Housing type was only related to type of delivery. Bivariate associations with infant mortality indicated that maternal preventive health behavior was related to individual health behavior and to some district level characteristics, which varied in level of significance and consistency. The maximum likelihood estimates revealed a negative significant effect of maternal tetanus immunization on infant mortality but not on delivery by a birth attendant or use of prenatal care. Stunting and the number of health facilities as contextual variables had a significant impact on infant mortality. The findings suggest that individual and community resources are important in reducing infant mortality.