Who participates in a family study?

Potential subject participation biases in a family study of outpatients with mood disorders and personality disorders (PDs) were explored at three levels: (1) differences between probands who granted permission to contact all relatives, those who gave permission to contact only a subset of relatives, and those who denied permission to contact any relatives; (2) differences between relatives whom the proband granted permission to contact and those whom the proband denied permission to contact; and (3) for the relatives who could be contacted, differences between those who agreed to participate and those who declined. Subjects included 156 outpatients with mood disorders and PDs and 611 of their first-degree relatives. Axis I and II disorders in probands and relatives were evaluated using structured diagnostic interviews. In addition, informant reports on relatives were obtained from family history (FH) interviews. Results indicated that probands who gave and who withheld consent to contact their relatives did not differ significantly on most variables. However, relatives whom we were not permitted to contact were significantly more likely to have drug abuse and PDs. Finally, of the relatives we were permitted to contact, there were few differences between those who participated in the study and those who refused to participate. These findings indicate that the greatest risk of sampling bias in family studies stems from probands' reluctance to grant access to relatives with drug abuse and PDs.