The use of nutritional status as a second outcome measure in case-control studies of environmental risk factors for diarrhoeal diseases.

Case-control studies are typically used to study the effect of several factors on the risk/incidence rate of a single disease. This paper describes a particular situation in which it is of interest to study the effect of a factor--improved sanitation facilities--on the risk/incidence rate of two 'diseases'--the incidence rate of diarrhoea and the risk of undernutrition. The conditions under which it is valid to perform an analysis of the association between the risk factor (unimproved sanitation) and a second outcome variable (undernutrition) are examined. If the effect of exposure status (improved/unimproved sanitation facilities) on the propensity to report an episode of diarrhoea is independent of the effect of nutritional status it appears that such an analysis may be valid. There must also be no interaction between the risk factor (unimproved sanitation) and the second outcome (undernutrition) with respect to their effects as risk factors for the first outcome variable (diarrhoea incidence rate).