The comprehensive cohort study: the use of registry data to confirm and extend a randomized trial

Abstract

There are distinct advantages to the comprehensive cohort design for randomized clinical trials. The design allows the researcher to systematically compare the trial participants and the trial results with the characteristics and results in the population of eligible patients who were not randomly assigned to treatment. There are some problems with this design. The added costs of enrollment and follow-up for the nonrandomized cohort can add substantially to the cost of a trial. As the CASS example illustrates, careful attention must be given to definition of the treatment groups in the nonrandomized cohort. For the randomized patients, group assignment by "intention to treat" is preferred in most cases. The definition of "intention to treat" in the nonrandomized cohort may be difficult, since the patient may receive recommendations from several sources and, in the case of chronic disease, the decision may be postponed for a long period. This problem is not specific to the comprehensive cohort design, however, and a comparison group of concurrently enrolled and followed patients is almost always preferred over a comparison group found elsewhere.