Drivers of clinician engagement in oncology trial accrual within an IMC in Shanghai: A mixed-methods study.

Abstract

e23034 Background: Recruitment shortfalls frequently delay oncology clinical trials. Integrated medical consortia (IMCs) may enhance accrual by strengthening cross-institution collaboration and aligning incentives across core hospitals and community health centers, but real-world evidence on the key IMC-related drivers of recruitment engagement remains limited. Methods: We conducted an anonymous cross-sectional online survey of healthcare professionals within an IMC led by Fudan University Pudong Medical Center (Shanghai, China). Using 5-point Likert items, we measured IMC-related constructs—Inter-organizational Collaboration Strength (ICS), Incentive Performance Perception Scale (IPS), Non-Performance Incentive Scale (NPIS), Subject Protection Index (SPI), and Capability Support (CS)—as well as Proactiveness in Screening/Recruitment (PS). Participants also reported accrual indicators from the prior 12 months (enrollment, informed-consent time, and loss-to-follow-up rate) and provided open-ended responses. Internal consistency was assessed using Cronbach’s α. Associations with PS were examined using Pearson correlations and multivariable linear regression with HC3 robust standard errors; mediation by CS was evaluated using a regression-based path model. Results: We analyzed 187 valid questionnaires; 22 (11.76%) respondents were from the core hospital and 165 (88.24%) from community health centers. Most were physicians (140, 74.87%), and 128 (68.45%) reported contact with cancer patients in the past 12 months. Mean years in practice was 16.07 (SD 10.32; n = 186). Recruitment indicators were right-skewed: median annual enrollments were 3 (IQR 0–10; n = 167), median informed-consent time was 15 minutes (IQR 0–30; n = 167), and median loss-to-follow-up rate was 10% (IQR 0–30; n = 168). Mean construct scores ranged from 2.91±0.60 (CS) to 3.60±1.04 (SPI), with excellent reliability (Cronbach’s α = 0.93–0.98). All constructs were positively correlated (r = 0.60–0.89; all p < 0.01), and PS correlated most strongly with NPIS (r = 0.89) and SPI (r = 0.84). In multivariable regression, the model explained substantial variance in PS (R² = 0.83; adjusted R² = 0.82; overall p < 0.001). After mutual adjustment, NPIS remained the strongest independent correlate of PS (β = 0.60, p < 0.001), and SPI also remained significant (β = 0.26, p = 0.020); ICS, IPS, and CS were not significant. In the path model, CS was predicted by SPI (β = 0.60, p < 0.001) and IPS (β = 0.24, p = 0.030), while the CS→PS path was not significant (β = 0.03, p = 0.591), indicating no evidence of mediation. Conclusions: In this IMC, NPIS and SPI showed the most robust associations with healthcare professionals’ PS. Implementation strategies that pair transparent recognition and career-aligned incentives with strengthened participant-protection workflows may enhance recruitment performance in consortium-based oncology trials.