Monitoring clinical research. A report from the Childrens Cancer Study Group.

This report describes an evaluative approach in assessing study quality in clinical trials. The size of the data set required is defined using decision analysis methods. Assuming entry of eligible patients, the minimum data set required in determining study quality includes: the treatment agent, the protocol stipulated treatment time, and the actual time of administration. While toxicity, complication and response data are primary for the resolution of the scientific issues, they are secondary for study performance questions. Separation of study performance and scientific questions simplifies the design and execution of complex, multimodal, multidisciplinary clinical trials. Application of separate criteria for study performance and scientific questions enhances understanding of protocol requirements by data managers, physicians, and nurses. Once study performance issues are resolved, the scientific issues addressed by the protocol can be considered in detail. The usual practice of focusing on science and study performance simultaneously precludes efficient consideration of either data set. The distinction between analyses of study performance and of scientific questions is illustrated using data from Childrens Cancer Study Group protocol CCG-551. This study is a controlled trial of therapies for non-Hodgkin's lymphoma. The protocol provides an example of modern diagnostic and therapeutic management complexities.