Abstract
BACKGROUND: Attrition is common among supportive care/palliative oncology clinical trials. However, to the authors' knowledge, few studies to date have documented the reasons and predictors for dropout. In the current study, the authors' objective was to determine the rate, reasons, and factors associated with attrition both before reaching the primary endpoint and at the end of the study. METHODS: A review of all prospective interventional supportive care/palliative oncology trials conducted in the Department of Palliative Care and Rehabilitation Medicine at The University of Texas MD Anderson Cancer Center in Houston between 1999 and 2011 was performed. Patient and study characteristics and attrition data were extracted. RESULTS: A total of 1214 patients were included in 18 clinical trials. The median age of the patients was 60 years. Approximately 41% had an Eastern Cooperative Oncology Group performance status of ≥ 3, a median Edmonton Symptom Assessment Scale (ESAS) for fatigue of 7 of 10, and a median ESAS for dyspnea of 2 of 10. The attrition rate was 26% (95% confidence interval [95% CI], 23%-28%) for the primary endpoint and 44% (95% CI, 41%-47%) for the end of the study. Common reasons for primary endpoint dropout were symptom burden (21%), patient preference (15%), hospitalization (10%), and death (6%). Primary endpoint attrition was associated with a higher baseline intensity of fatigue (odds ratio [OR], 1.10 per point; P = .01) and a longer study duration (P = .04). End-of-study attrition was associated with higher baseline levels of dyspnea (OR, 1.06; P = .01), fatigue (OR, 1.08; P = .01), Hispanic race (OR, 1.87; P = .002), higher level of education (P = .02), longer study duration (P = .01), and outpatient studies (P = 0.05). CONCLUSIONS: The attrition rate was high in supportive care/palliative oncology clinical trials, and was associated with various patient characteristics and a high baseline symptom burden. These findings have implications for future clinical trial design including eligibility criteria and sample size calculation.