Birth cohort studies offer the opportunity to study average risks, rates and occurrence times of disease longitudinally from birth. The effect of genetic and environmental factors and their interactions can be studied. Furthermore, quantity and duration of exposure to environmental agents can be evaluated prospectively. However, prospective birth cohort studies are expensive, labour intensive and take many years to complete. Loss of subjects over time as well as recall bias complicate the interpretation of observations. This paper summarises the potential pitfalls of such studies and discusses the experience of the German Multicentre Allergy Study (MAS), which began in 1990 in five German cities and included 1314 newborns for the study of the natural course of atopic diseases.
Copyright 2002 Elsevier Science Ltd.