Technology-enabled clinical trials: Intentional capture of source data (IDC) in the electronic-health record (EHR) and direct transfer to trial database (electronic data capture [EDC]) in a phase II multicenter trial.

Abstract

1568 Background: Traditional data collection for clinical trials can be cumbersome and inefficient. FDA guidance now promotes capturing clinical trial source data in electronic form and transmitting it to the EDC, recognizing that this should reduce double data entry and improve data quality and timeliness. Electronic data originators can include the EHR, which can be used during routine clinical care and for gathering data during a clinical investigation. We evaluated a pilot of structured EHR data transfer to EDC in a clinical trial setting. Methods: IDC capability in the EHR was developed and applied to a subset of patients for an open label prospective Phase II study. EHR variables, that include free text, standard dictionaries (eg CTCAEv5, CDISC, First Databank), and derived responses, were mapped to the electronic case report form (eCRF). Bulk transmission of IDC clinical data (adverse events, AEs; concomitant medications, CMs; past medical history, PMH) from the EHR to EDC was executed at a designated time point during study conduct. We assessed the quantity (records) and quality (queries) of data captured and transferred, and timeliness of entry. Data to be used for the production analysis of the trial was collected via eCRF in an EDC system in parallel per usual standards. Results: Data for 102 patients across 8 sites of care by 99 different users was captured in the EHR via IDC from 3/1/2021 to 9/9/2022 and transferred from EHR to EDC in bulk on 9/9/22. 3,958 data records were transferred (852 AEs; 1,581 CMs; 1,525 PMHs). Of the AEs, 99% had an associated grade. Of the CMs and PMHs, none were copy-forwarded (all were entered de novo). Regarding queries in the EDC for transcription error validation among matching records, EHR data produced 205 opened edit check queries from 12 edit checks across all domains (149 AEs, 0 CMs, and 56 PMHs). Regarding timeliness, 95% of matched records were entered in IDC prior to entry in the production EDC. Conclusions: For an open label Phase II study, clinical trial source data was captured electronically through IDC using the EHR as the data originator, and directly bulk-transmitted to EDC. This may enable the elimination of transcription of source data entry into eCRF and result in more timely access to the data in the EDC. Additional investigation into how IDC with EHR-to-EDC compares in quality to traditional EDC is warranted. Planned follow up research includes interviewing site users to evaluate their experiences with IDC data entry.