Abstract 4137986: Evaluation of an AI-Based Clinical Trial Screening Method Through a Randomized Controlled Implementation Study

Abstract

Background: Clinical trial screening is labor-intensive, time-consuming, and error prone. We have developed RECTIFIER, an AI-based clinical trial screening tool, to enhance the efficiency and accuracy of patient recruitment. This study aims to evaluate RECTIFIER's effectiveness compared to manual screening in a randomized implementation study. Methods: This study was designed as an implementation study as part of an active heart failure trial named COPILOT-HF (NCT05734690). Potential eligible patients were identified via a structured electronic medical record query and randomized to be screened for clinical trial eligibility either by RECTIFIER or manually by clinical staff. The outcome measures included the number of patients contacted, and the number of patients reached for clinical trial enrollment. Data was collected over a period of 3 months. Results: A total of 3834 patients were included in the study, with 1919 patients randomized to the RECTIFIER group and 1915 patients to the manual screening group (Figure). Study staff could manually screen only 1367 patients at the end of the 3-month period. RECTIFIER identified more eligible patients compared to manual screening (833[43.4%] vs. 284[14.8%], p<0.001). The most common reason for exclusion was absence of symptomatic heart failure in both groups (287 in RECTIFIER 304 in manual screening). Screening via RECTIFIER also led to a higher contact rate, with 455 patients being contacted compared to 216 in the manual group (23.7% vs. 11.3%, p<0.001). The percentage of patients who could be reached for an enrollment call was also higher in the RECTIFIER group, with 330 patients responding versus 164 in the manual group (17.2% vs. 8.6%, p<0.001). Conclusion: Screening via RECTIFIER significantly outperformed manual screening in identifying eligible patients, contacting them, and eliciting responses. These results suggest that integrating AI into clinical trial recruitment processes can enhance efficiency and effectiveness, potentially leading to increased patient participation in clinical trials. Future studies with sufficient power should assess recruitment endpoints and explore scalability.

Ähnliche Arbeiten