Remote preenrollment checking of consent forms to reduce nonconformity

In biomedical research, the signed consent form must be checked for compliance with regulatory requirements. Checking usually is performed on site, most frequently after a participant's final enrollment.

We piloted a procedure for remote preenrollment consent forms checking. We applied it in five trials and assessed its efficiency to reduce form nonconformity before participant enrollment.

Our clinical trials unit (CTU) routinely uses a consent form with an additional copy that contains a pattern that partially masks the participant's name and signature. After completion and signatures by the participant and investigator, this masked copy is faxed to the CTU for checking. In case of detected nonconformity, the CTU suspends the participant's enrollment until the form is brought into compliance. We checked nonconformities of consent forms both remotely before enrollment and on site in five trials conducted in our CTU. We tabulated the number and nature of nonconformities by location of detection: at the CTU or on site. We used these data for a pseudo before-and-after analysis and estimated the efficiency of this remote checking procedure in terms of reduction of nonconformities before enrollment as compared to the standard on-site checking procedure. We searched for nonconformity determinants among characteristics of trials, consent forms, investigator sites, and participants through multivariate logistic regression so as to identify opportunities for improvement in our procedure.

Five trials, starting sequentially but running concurrently, with remote preenrollment and on-site checking of consent forms from 415 participants screened in 2006-2009 led to 518 consent forms checked; 94 nonconformities were detected in 75 forms, 75 (80%) remotely and 19 more (20%) on site. Nonconformities infrequently concerned dates of signatures (7%) and information about participants (12%). Most nonconformities dealt with investigator information (76%), primarily contact information (54%). The procedure reduced nonconformities by 81% (95% confidence interval (CI): 73%-89%) before enrollment. Nonconforming consent forms dropped from 25% to 0% over the period, indicating a rapid learning effect between trials. Fewer nonconformities were observed for participants screened later in a trial (odds ratio (95% CI): 0.5 (0.3-0.8); p = 0.004), indicating a learning effect within trials. Nonconformities were more common for participants enrolled after screening (2.4 (1.1-5.3); p = 0.03), indicating a stricter scrutiny by form checkers.

Although our study had a pseudo before-and-after design, no major bias was identified. Power and generalizability of our findings were sufficient to support implementation in future trials.

This procedure substantially limited nonconformity of consent forms with regulatory requirements before enrollment, thus proving a key component of a risk-based monitoring strategy that has been recommended to optimize resources for clinical research.