Factors influencing withdrawal without a defined reason in oncology trials.

Abstract

e23014 Background: Withdrawal without a defined reason (WWDR) undermines the reliability and interpretability of clinical trial results and poses a critical challenge for trial design. The underlying factors driving WWDR remain poorly understood. This study seeks to identify trial-level factors associated with WWDR that can be intervened upon to improve participant retention. Methods: We systematically reviewed therapeutic phase 3 oncology trials registered on ClinicalTrials.gov from August 20, 2014 to August 20, 2024. Supportive care trials and those lacking a participant flow diagram or equivalent were excluded. WWDR was defined as withdrawal from the study initiated by either patients or investigators without a specified cause recorded (e.g. nonadherence, patient requests, or loss to follow-up). Generalized Additive Models with bidirectional stepwise selection were used to perform multivariate regression. Continuous variables were standardized and associations evaluated with rate ratio (RR). Results: 300 studies met inclusion criteria, including 165,674 patients. The median follow-up was 35.6 months (Interquartile range [IQR]: 4.4–66.8), and median enrollment size was 480.5 participants (IQR: 259–702). Most trials focused on solid tumors (76.2%, n = 232), mainly metastatic (77.2%, n = 179), with fewer addressing earlier stage disease (22.8%, n = 53). Hematologic malignancies accounted for 23.2% (n = 70). The majority of trials were randomized (88.0%, n = 264), unblinded (61.7%, n = 185), industry-sponsored (90.7%, n = 272), conducted internationally (95.0%, n = 285), and multi-site (93.7%, n = 281). Key predictors of WWDR are detailed in Table 1. Industry sponsorship was linked to a 67% higher likelihood of WWDR (RR: 1.67, 95% CI: 1.18–2.57, p = 0.02), and larger trial enrollments showed increased likelihood of WWDR per standard deviation (SD) in enrollment (SD = 491; RR: 1.13, 95% CI: 1–1.28, p = 0.045). The model explained 14.3% of WWDR rate variance. Conclusions: Industry-sponsored trials and those with larger patient enrollment were associated with significantly higher rates of WWDR. In contrast, combination therapies and radiotherapy (RT) reduces these rates. Trials with higher enrollment often face logistical complexities and variability across sites, while industry-sponsored trials may impose stricter protocols, raising WWDR. Combination therapies and RT may improve retention by increasing patient interaction and monitoring inherent to the treatments. Future research should analyze patient-level factors to clarify WWDR and identify strategies to reduce these rates prior to enrollment. Trial-level factors associated with WWDR. Predictor OR (95% CI) p-value Combination therapy 0.73 (0.56, 0.93) 0.01 Solid tumor receiving definitive RT 0.21 (0.10, 0.37) <0.001 Trial enrollment size (per SD increase)* 1.13 (1, 1.28) 0.045 Industry-sponsored trial 1.67 (1.18, 2.57) 0.02 Standardized by SD = 491 patients.