How pilot studies improve large-scale clinical trials: lessons learned from the COMBINE Study.

Objective: The design of a clinical trial to evaluate a potential therapy requires decisions about issues that include safety, efficacy, measurement, feasibility and training. Experience from the COMBINE Study, which tests the combination of medications and behavioral therapies for alcohol dependence, is presented as an example of how pilot studies improve large-scale clinical trials.

Method: The COMBINE Pilot 1 inpatient study was designed to inform the main trial about the safety and tolerability of the doses of acamprosate (3 g/day) and naltrexone (100 mg/day) selected for study, alone and in combination. Pilot 2 was conducted as a feasibility study for the main trial, with the goals of (1) assessing the length of and compliance with research assessments, (2) developing methods for subject recruitment and staff training and (3) assessing the safety of the medications under less controlled outpatient conditions.

Results: Results from Pilot 1 provided safety information to support testing the medications in an outpatient study and contributed to the decision to incorporate dose reductions into the main trial protocol to manage adverse events. The results of Pilot 2 formed a basis for (1) reducing the length of the assessment battery, (2) having staff fully trained and recruitment procedures established for the main trial and (3) extending the drug safety results of Pilot 1 to outpatient conditions similar to those of the main trial.

Conclusions: The COMBINE Study provides several examples of the successful application of pilot studies to inform the design of a clinical trial.