An analytical study of child survival using the Sudan, Egypt and Yemen PAP-CHILD surveys
The thesis is a comparative study of, Egypt, Sudan and Yemen, three countries with similar social and economic profiles, yet with a variable dynamic in achieving reductions in child mortality levels. The study begins with a consideration of the individual country backgrounds and then presents comparative findings on population health and child survival. Empirical results on the correlates of child survival are presented, together with a selective review of the related techniques of analysis. The analyses of survival to age five was based on data from the PAPCHILD surveys carried-out in Egypt (1991), Sudan (1992/93) and Yemen (1991/92). The aim was to investigate the determinants of child survival with the innovation of adjusting for the effect of a family's "child mortality background". Methods of analysis included life-table analysis, logistic (marginal and multilevel) and Cox regression models. The transition to better child survival could further benefit from the spacing of births, the avoidance of higher-order births, and the concentration of childbearing in the central reproductive ages. Unequivocally, deaths of older siblings prior to the birth of every index child were strong predictors of poor survival settings. Deaths of older siblings after the birth of the index child were rare, yet captured "immediate" risk spells. Events of conception, birth and death of a subsequent sibling entailed time-varying excess risks. Evidently, adjusting for measures of familial child losses explains much of the "between-households" variation in mortality risks and spell-out "within-households" inter-dependencies of survival. Households further correlate in risks to child survival when they belonged to the same geographical cluster. The novelty in representing the latter correlation with a "regional" component of unmeasured effects was in aid of pertinent policy recommendations. Further, the study makes recommendations on reducing reporting errors of demographic data collected from mothers. Critical findings and policy implications are: for Egypt, better child survival rates are achievable by narrowing "regional" socio-economic gaps and sustaining lower fertility rates; in Sudan, the slowing pace of declines in child mortality were not best explained by relations with observed correlates, and appears further underpinned by the country's economic crisis; in Yemen, child mortality levels can be reduced by a third if the timing between successive births could be extended to two years, net of key promotive socio-economic interventions.