college education. For the second tdal, educational level was also associated with enrollment, as were race/ ethnicity and household composition. These findings prompted refinements to recruitment techniques, patient education, and study protocol.
A103 A STUDY OF PATIENT ACCRUAL PAI"I'ERNS IN 23 VA MULTICENTER CLINICAL TRIALS
William G. Henderson VA Medical Center
VA Cooperative Studies Program Coordinating Center Hines, Illinois
Patient accrual patterns in 23 VA multicanter clinical trials from 9 different disease areas were examined to help plan and conduct future trials. Accrual rates ranged from a high of 4-6 patients per hospital-month in some hypertension tdals to a low of 0.3-0.5 in some tdals involving relatively rare diseases. In only three studies was actual accrual considerably below required accrual, although three other studies required ex- tensions to the intake period and one study required more hospitals. Observed/expected (O/E) accrual ratios were computed for hospitals, months of the year, and years of the intake period, with expected accrual calculated assuming a uniform distribution. Larger hospitals tended to have petter accrual, although the differences were not statistically significant (p = 0.20). Surprisingly, accrual was better in hospitals serving a population under 200,000 compared to those serving populations of over 1 million (p < 0.05). Accrual rates were highest in the spring months and lowest in the winter months. In 83% of the trials there was a fall-off in accrual in the later years of the tdals. On average 32% of the screened patients were randomized. About 8% of screened patients refused to participate. The number of patients randomized at a hospital was correlated with the number of patients screened (r = 0.33). Group meetings did not appear to affect patient accrual. PI interest in the study, as assessed by service on the Executive Committee, was positively associated with accrual. Implications of these findings on planning future tdals will be discussed.
A104 COMPARISON OF CUNICAL TRIALS REGULATORY REQUIREMENTS AND PRACTICE BETWEEN
EUROPEAN COMMUNITY AND UNITED STATES
Jay Herson and John Glasby Applied Logic Associates, Inc.
Houston, Texas International Research Consultants
The authors use published regulatory guidelines as well as their expedance to compare theory and practice of clinical trials in the EC and USA for regulatory approval of four classes of drugs: anti-inflammatory, anti- epileptic, anti-arrhythmic, anti-depressant. The authors cover designs and requirements of phase I trials, dosage response trials, and pivotal efficacy studies. For the latter, they compare number of individual trials, patients and investigators required, designs used, type of control group, end-points for safety and efficacy, statistical analyses performed, adverse event reporting, use of foreign data, special requirements for children and the elderly, the role of cost-effectiveness studies, quality of life studies and metaanalysis. The overall length of an R&D program from discovery to commencement of phase I trials, commencement of phase II/ III tdals, and regulatory approval are also compared. The authors explain the implications of similarities and differences in clinical trial requirements between the EC and USA for designing and conducting international clinical trials.
A105 PHASE I TRIALS: IS IT BETTER TO USE PATIENTS OR NORMAL VOLUNTEERS?
Rlcherd M. Kaplt National Institute on Drug Abuse
In the United States, the standard methodology for introducing a new compound into man is performing a rising single dose tolerance study, Usually such trials make use of normal volunteers. Until the late 1950s, there was no such standard practice, and new drugs were first tried in patients by practitioners. The author, a former clinical reviewer at the FDA, traces the history of clinical tdals, emphasizing the origins of phase I