Utility of automated data transfer for cancer clinical trials.

e23011 Background: The burden of data collection on site staff for cancer clinical trials is steadily increasing, and inefficiencies in data entry into electronic data capture (EDC) systems lead to poor data quality and delays in reporting. Software that facilitates automated transfer of mapped structured data from the electronic health record (EHR) to the study EDC system can help to address these challenges. We examined the impact of multi-site usage of an embedded point and click EHR-to-EDC tool on study data capture across multiple cancer clinical trials. Methods: We conducted a retrospective analysis of usage of an EHR-to-EDC tool among research teams at 3 Academic Medical Centers (AMCs) for 4 industry-sponsored Phase 1 cancer clinical trials from the start of use of the tool at each AMC to a cutoff date of 12/31/23. We assessed the types and volume of structured data accessible via FHIR APIs and amenable for transfer (“mapped”) across protocols, the volume of data transferred per protocol, and the time for data transfer, defined as time from case report form (CRF) launch to submission in EDC and time spent in the software application overall. Descriptive statistics were generated for all study metrics. Results: During the 10 month observation period, the EHR-to-EDC tool was used by 8 users to transfer data for 25 patients enrolled across the 4 protocols. Table 1 shows the percentage of total CRF forms mapped and breakdown of proportion of mapped fields per protocol. Overall, 8,597 individual data points (87% labs, 10% vitals and 3% concomitant medications) were pushed for 718 unique CRF submissions. The average number of individual data points and unique CRF submissions per patient were as follows: 400 and 36 (Study A), 420 and 36 (Study B), 251 and 16 (Study C), 270 and 21 (Study D). The average total time spent in the tool per CRF submission was 2.16 minutes (range 1.17-7.00 minutes, inclusive of potential inactive time, application launch and data search). The average time to complete a CRF was 46 seconds (range 15-65 seconds). Conclusions: Use of EHR-to-EDC software facilitated efficient automated transfer of a high volume of study data due to EDC-required repeat use of mapped CRF forms. The impact appeared consistent across protocols, with dozens of unique CRF submissions and hundreds of individual data points pushed per patient for each study. Time for CRF completion was low and uniform across users. This study supports the growing adoption of EHR-to-EDC technology among sites and sponsors, and highlights the evolving potential to further improve and facilitate clinical trial data entry, as tool capabilities are developed beyond structured fields to enable unstructured data capture and transfer. Table: see text