Evaluation of the quality of investigative centers using clinical ratings and compliance data

In late phase clinical trials, a large number of investigative centers are needed for recruiting sufficient number of patients in a timely manner. The quality of investigative centers involved in multi-center clinical studies can be crucial for the success of the trials. If 'high quality' centers can be identified from completed studies, we may use these centers in future trials to improve the chance of success. We explored the possibility of evaluating clinical centers based on clinically rated scores, the extent of protocol violations and patient non-compliance from several double-blinded, multi-center, placebo-controlled studies of aprepitant. To evaluate patient non-compliance on protocol-scheduled visits, a generalized estimating equations (GEE) approach was used to estimate the visit compliance scores for the investigative centers. A new rank score that combines these factors was derived using principal component analysis. The rank score was also compared to the original clinical rated overall score.