Can categorization of patient background problems be used to determine early termination in a clinical trial?

Many patients entering a long-term clinical study are unable to remain active in the trial (taking medication as prescribed and attending clinic as scheduled) for up to 6 years. We studied a group of 622 patients entering an antiepileptic drug study with maximum possible follow-up period of 6 years. A centralized case validation system was used in this VA Epilepsy Cooperative Study to note problems in patients entering the study. The categories established for pre-entry case status were associated with events occurring during the course of the study that led to noncompletion of the study. The data were reviewed to determine whether pre-entry status can be used to suggest the likelihood that an individual patient will remain in the study to completion or become a non-drug-related loss. Over one-half of all losses were related to failure to return to clinic. Medical problems (18%), medication noncompliance (14%), and psychiatric problems (12%) accounted for the remainder of losses. Pre-entry issues often were the same reason for early loss. However, many patients with problems known at entry were maintained in the study when targeted for extra support. We found some groups of patients had considerably more early terminations whereas others had less loss to the trial than we predicted originally. Thus, case categorization at entry can be used to estimate the likelihood of early termination from a clinical trial.