Accurate and robust adverse event (AE) data collection is crucial in cancer clinical trials to ensure participant safety. Frameworks have been developed to facilitate the collection of AE data and now the traditional workflows are facing renewal to include patient-reported data, improving completeness of AE data. We explored one of these workflows in a cancer clinical trial unit. The study was a single-site study conducted at a tertiary hospital located in Australia. Patients consenting to a clinical trial were eligible for inclusion in this study. Participants used an electronic platform-My Health My Way (MHMW)-to report their symptomatic data weekly for 24 weeks. A symptom list was included within the platform, along with a free text field. Data reported via the platform was compared with data recorded in the patient's medical chart. Time taken to compile data from each source was recorded, along with missing data points. Agreement between patient-reported data and data recorded in the medical notes was assessed using Kappa and Gwet's AC1; time taken to compile data and missing data points were assessed using a Wilcoxon signed rank test. Low agreement was found between patient- and clinician-reported data (-0.482 and -0.159 by Kappa and Gwet's AC1 respectively). Only 127 (30%) of the total 428 AEs were reported by both MHMW and medical notes. Patients reported higher rates of symptoms from the symptom list, while clinicians reported higher rates of symptoms outside of the symptom list. Time taken to compile the data from MHMW was significantly less than that taken to review medical notes (2.19min versus 5.73min respectively; P < 0.001). There were significantly less missing data points from the MHMW data compared with the medical notes (1.4 versus 7.8; P<0.001). This study confirms previous reports that patient- and clinician-reported adverse event data show low agreement. This study also shows that clinical trial sites could significantly reduce the work performed by research staff in the collection of adverse event data by implementing an electronic, patient-reported platform.
Read full abstract