Urbanisation and health: Methodological difficulties in undertaking epidemiological research in developing countries

Rapid urbanisation has historically resulted in complex societal changes which have had both beneficial and adverse effects on the health of communities. Recently, it has been recognised that the urban poor in developing countries are at greatest risk for several adverse health effects. The epidemiologist has a key role to play in planning to meet current and future health needs of urban communities. Epidemiological research needs to focus on understanding the relationship between the changes that occur with urbanisation and their impact on health, in order to identify those at greatest risk, to identify the important risk factors that are amenable to intervention, and to evaluate the effectiveness of interventions aimed at improving the health of urban communities. This paper addresses several methodological difficulties in conducting epidemiological research on urbanisation and health in developing countries, with particular reference to Southern Africa. These relate to the definition of urban areas and residence thereof, the comparability across areas of exposure and outcome information, the identification of intra-urban variability, selection bias in cross-sectional studies, associating outcomes with specific urban exposures in analytic studies, and determining appropriate interventions and means of evaluation. Ways of overcoming these difficulties are suggested. Epidemiologists need to address these issues in order to make a valuable contribution to the improvement of health in urban communities.