Selection of patients for randomised trials: a study based on the MACOP‐B vs CHOP in NHL study

Abstract

Selection of patients for a clinical trial is affected by awareness of the existence of the trial, interest in the study question and clinical practices and views of the clinicians. To investigate the selectivity that may have occurred at Peter MacCallum Cancer Institute (PMCI) during the ANZ Lymphoma Group trial of MACOP-B vs CHOP in non-Hodgkin's lymphoma (NHL). NHL patients at PMCI in the study period were assessed against the trial's eligibility criteria. Comparisons were made between eligible (except for consent) non-trial patients and all patients actually randomised into the trial. Of 497 patients presenting during the trial period, 320 (64%) did not meet the specified eligibility criteria, 102 (21%) were unsuitable on other grounds (age and medical) and 75 (15%) were eligible. Of those eligible, 43 (57%) were entered into the trial and 32 (43%) were not. Four non-trial patients had inappropriate application of eligibility criteria and 13 unknown reason. Eligible non-trial patients were similar to trial patients in most patient and tumour characteristics and overall survival. Significantly more non-trial patients had higher stage disease (p = 0.02). More non-trial patients had lower grade histology, but this was not significant. Physician selectivity occurred with respect to patient entry, but trial and non-trial patients were similar in most characteristics. Eligibility criteria should specify that patients can withstand all trial drugs and patient availability for treatment and follow-up. PMCI trial accural could have been up to 33% greater. These results suggest the trial accrual period could have been 25% shorter. Patient entry into this trial by PMCI clinicians compared favourably with other centres.