Lotte Keikes, Linda Mol, Martijn G.H. van Oijen, Miriam Koopman, Cornelis J. A. Punt and Petronella B Ottevanger
Objective: High quality clinical trials are essential for further improvement of treatment strategies for prolonged survival and reliable evidence-based outcomes. However, there are no defined standards for the quality of clinical trial performance. The aim of this study is to examine and compare clinical trial performance with a composite score between (different types of) hospitals, to identify potentially predicting factors for a high trial performance and examine a learning curve in composite performance scores between early compared to subsequent included patients.
Methods: We evaluated trial performance in three large phase 3 randomized clinical trials in metastatic colorectal cancer (CAIRO studies of the Dutch Colorectal Cancer Group, total n=2131) with a newly introduced composite score, consisting of stratification errors, major protocol violations, number of included ineligible patients, and reporting of serious adverse events (SAE) on hospital and patient level. These data were supplemented with a hospital survey containing questions about number of beds, oncologists and research nurses. A logistic regression was performed to identify factors associated with better trial performance (3-4 points).
Results: We observed variation in trial performance between 84 participating hospitals. However, no differences in performance between hospital categories (university, teaching and regional hospitals) were identified and none of the examined variables could be linked to a high composite performance score. In top 10 ranking hospitals with highest inclusion rates, trial performance on patient level was significantly lower in the first three inclusions compared to subsequent patients..
Conclusions: Trial performance was comparable between different types of hospitals and no factors were able to predict a high composite trial performance score. In the highest including hospitals we identified a learning curve for trial performance. We therefore recommend increased support during the first patient inclusions in participating centres in order to improve trial performance. Our composite score could be used as a quality metric for trial performance for individually based hospital evaluation.
PDFShare this article
Journal of Cancer Clinical Trials received 95 citations as per Google Scholar report
