The selection of healthy volunteers for clinical investigation: the case for volunteer pools.

Discussions on the healthy volunteer in clinical investigation do not attempt to define normality. In 34 healthy volunteers, 1653 biochemical and haematological tests were performed. The incidence of abnormal results was 11%. Only 4 subjects had all tests within normal limits and when these repeated only 1 subject still had all results within normal limits. The authors recommend that a stable volunteer pool, with regular physical examinations and blood tests, would provide a large, serial data base for use in subject selection for clinical trials. Attention should be given to human error. Protocols should be realistic, while ensuring the safety of volunteers.