Differences between blood donors and a population sample: Implications for case-control studies

Abstract

Background Selecting appropriate controls for studies of genetic variation in case series is important. The two major candidates involve the use of blood donors or a random sample of the population. Methods We compare and contrast the two different populations of controls for studies of genetic variation using data from parents enrolled in the Avon Longitudinal Study of Parents and Children (ALSPAC). In addition we compute different biases using a series of hypothetical assumptions. Results The study subjects who had been blood donors differed markedly from the general population in social, health-related, anthropometric, and personality-related variables. Using theoretical examples, we show that blood donors are a poor control group for non-genetic studies of diseases related to environmentally, behaviourally, or socially patterned exposures. However, we show that if blood donors are used as controls in genetic studies, these factors are unlikely to make a major difference in detecting true associations with relatively rare disorders (cumulative incidence through life of < 10%). Nevertheless, for more common disorders, the reduction in accuracy resulting from the inclusion in any control population of individuals who have or will develop the disease in question can create a greater bias than can socially patterned factors. Conclusions Information about the medical history of a control and the parents of the control (as a proxy for whether the control will develop the disease) is more important with regard to the choice of controls than whether the controls are a random population sample or blood donors. © The Author 2013; All rights reserved.