Risk factors for measles mortality : Studies from Kenya and 19th century Stockholm

Abstract: Risk Factors for Measles Mortality Studies from Kenya and l9th century Stockholm by Bo Burström, Karolinska Institutet, Department of Public Health Sciences, Division of Social Medicine, SE-172 83 Sundbyberg, Sweden Measles kills about 1 million children per year in low-income developing countries. In many industrialized countries, the transmission of measles and its lethal effects have become controlled through immunization and social developments in general. The thesis addresses modifiable risk factors for measles mortality, particularly the effect of crowding and clustering of measles cases on the risk of death due to measles, using current data from rural Kenya 1986-88 and historical records from Stockholm in the years 1885, 1891 and 1910. Special emphasis is laid on trying to disentangle effects of co-varying factors, such as crowding, low age and low social class on the risk of death from measles. The thesis is based on six papers: the first two describe field studies of measles outbreaks in western Kenya, with respect to vaccine efficacy and risk factors for high case-fatality rates (CFR). Paper III compares the validity of hospital data and community data in assessing the magnitude of measles mortality in Kenya. Paper IV reviews studies concerning incidence and severity of measles, and measles vaccine efficacy in infants in East Africa. Paper V reports on the validity and relevance of Swedish historical data in the study of measles mortality. Paper VI addresses the impact of crowding and other social risk factors on the risk of death from measles in Stockholm. Measles CFR was high (8-9%) in Kenya. In one community in Kenya 33% of all child deaths were attributed to measles; in Stockholm 16%. A large proportion of deaths occurred in infants, 38% and 45% in Kenya; 31% in Stockholm. Vaccine efficacy was lower than optimal in Kenya. Hospital-based data on measles deaths gave lower estimates of CFR than community studies. The Swedish historical data were considered valid, and similar to recent studies in low-income developing countries with respect to major causes, levels and age- structure of death. In Kenya clustering of measles cases and low age at infection were associated with high CFR. In Stockholm proxies for crowding were associated with a high risk of measles death but not other causes of death, also when controlling for low age, low social class and being born out of wedlock. Low social class and low age were also associated with a high risk of measles death. Studies of historical data may contribute to the understanding of present-day child mortality in low-income developing countries. Key words: measles mortality, immunization, crowding, clustering, social factors, child mortality.