A study of indirect methods of estimating mortality with reference to Kenyan data

Abstract

It is true that Kenya is one of the African countries where accurate, complete data on mortality are not available. Up to now, Civil Registration has not been developed to the extent that it can cover every birth and death which occurs. Even in the areas where the public have been exposed to this exercise, a low proportion of the events are registered. This method being the only reliable source of complete data on fertility and mortality, it is apparent that the mortality situation can only be studied by using estimates based on other less exhaustive methods of data collection. There are several indirect methods of mortality estimation from data of this nature. The question which can be posed here is whether or not these methods of estimation arrive at the same result. If they do not, can one be singled-out as the best performer on Kenyan data? This is the main aim of this study. The study begins by giving a brief description of each of three methods used in estimating infant and child mortality. Using data from 1969 and 1979 Population Censuses and some demographic surveys which were conducted between these two dates, mortality estimates have been calculated on provincial level by applying each of these methods. The estimates confirmed the fact that the level of mortality, particularly at the early ages of life is still high in most of the provinces of the country. In an attempt to find out whether or not there was any significant

difference between the resulting estimates, a comparison was made on these results by.method of estimation. The comparison showed that Trussell IS version of the methods of estimating infant and child mortality is more suitable to use under the conditions of varying fertility and mortality observed among the provinces. The estimation of adult mortality offered relatively little flexibility in that the relevant data were not available especially in the estimation of mortality among adult males. Where data were available, e.g. for the estimation of female adult mortality, there was hardly any choice in the method of estimation. The results obtained from the two methods used showed that female adult mortality is still high in most of the provinces. Finally, infant and child mortality estimates were patched together with female adult mortality estimates and the result was female life tables