A total of 96 mothers of stillborn babies (cases) and 402 mothers of live-born babies (controls) were studied between the 1st November 1988 and the 31st July 1989 in an attempt to identify the risk factors for the high number of stillbirths in the Maternity wing of the Hôpital National Sourô SANON in Bobo-Dioulasso, Burkina Faso. From a stratified analysis, eight risk factors emerged as significant predictors of mortality: mother older than 35 years, high parity (five or more), previous history of stillbirth, lack of medical supervision of pregnancy, interval between last consultation and birth longer than 30 days, complications during delivery and birth weight less than 2,500 g. Polygamy was associated with a 51% reduction in the risk of stillbirth. The problem of selection bias which affects the validity of hospital based case-control studies is discussed. Techniques for reducing selection bias in hospital based case-control studies are proposed. Improvements in antenatal care in Bobo-Dioulasso are suggested.