Abstract WMP60: Consent Deviations in an Acute Ischemic Stroke Clinical Trial

Abstract

Background: MOST (Multi-arm Optimization of Stroke Thrombolysis) was a multicenter, randomized controlled trial of adjuvant argatroban, eptifibatide, or placebo in moderate to severe acute ischemic stroke patients who received intravenous thrombolysis within three hours of stroke onset. We compared protocol deviations between participants consented using conventional paper informed consent documents (ICDs) versus electronic informed consent (eConsent). Methods: MOST began in 2019 and was completed in 2023. When obtaining informed consent, research teams could elect to use conventional paper ICDs or an eConsent platform managed centrally by the NIH StrokeNet. We studied informed consent-related data clarification requests (DCRs) and reportable unanticipated events (UE) (i.e. protocol deviations). Data for all modalities of consent (paper-in person, paper-remote, eConsent-in person, eConsent-remote) were collected in WebDCU™. REDCap was used as a repository for completed eConsents. We used Chi-squared and Kruskal Wallis tests to detect differences between consent modalities and stratified by whether they were remote or in-person. Two authors reviewed UEs to determine if they were consent-related and to categorize themes; in cases of disagreement a third reviewer adjudicated the final decision. Results: The rate of DCRs and UEs per 100 enrollments for each consent modality is shown in Table 1. Pairwise testing revealed that eConsent-in person had significantly less DCRs than both paper-in person and eConsent-remote (p<0.001), but there was no difference between paper-in person and eConsent-remote (p=0.32). Paper-in person had significantly more UEs than both eConsent-in person and eConsent-remote (p<0.001); the rate of UEs for eConsent-in person was also less than eConsent-remote (p=0.04). The themes that emerged in order of frequency were: HIPAA form missing/incorrect, incorrect version of the main consent, wrong person/wrong section signed and miscellaneous problems. Conclusion: The rate of DCRs and UEs were highest for paper in-person consent modality. These findings are most likely attributable to having typed vs handwritten data fields, required and forced-entry fields, a built-in reminder to complete HIPAA authorization, and centralized version control. This highlights the potential for eConsent to alleviate the research burden related to error mitigation and remittance in clinical trials; especially if scaled across research networks.

Ähnliche Arbeiten