Informed consent as a source of bias in clinical research

Clinical observation suggested that schizophrenic patients more often refused to participate in experiments. In two studies psychiatric patient volunteers and nonvolunteers were compared on discharge diagnoses (psychiatric and medical), medication given, demographic measures, and hospitalization characteristics. In each study, patients with paranoid schizophrenia were more likely to refuse participation. Also, in each study, volunteers received more benzodiazepines and more nonpsychiatric medication. The latter difference could not be explained on the bases of rated severity of or probability of drug use in diagnosed medical illnesses, or number of medical illnesses. Characteristics of schizophrenic patients such as suspiciousness, negativism, and ambivalence were suggested as possible sources of their refusal to participate. Nonvolunteers were suggested to be less agreeable and compliant in general, and perhaps less likable, possibly explaining their receiving fewer benzodiazepines and nonpsychiatric medications. Attempts to characterize and correct the volunteer bias should be made, including offering tangible reward.