Testing of electronic case report forms (eCRFs): survey of UK academic clinical trial units

Abstract Background When collecting data for clinical trials, there must be checks and testing (validation) undertaken to ensure that the data collected/entered is captured exactly and in a robust and reproducible manner verified by the teams managing the trial. It is therefore important to ensure that electronic case report forms (eCRFs) are meticulously tested. Inadequate testing can compromise the reliability of trial data and reduce the impact of the research findings. Despite the central role of eCRF testing in clinical data management, there has been little systematic investigation into how academic clinical trial units (CTUs) within the UK conduct this process. The aim of this study was to examine how eCRF testing is conducted within these units, to then inform which aspects of testing could be potentially improved or addressed. Methods A survey link was sent to the information technology (IT) leads of the 50 UKCRC-registered CTUs to complete between April and June 2024 regarding the electronic data capture (EDC) systems used, testing processes undertaken and their thoughts on automated testing. Responses were analysed in Stata 18 using Fisher’s exact, Kruskal–Wallis and logistic regression tests. Results Responses from 27 of the 50 UKCRC-registered CTUs were received, representing 54% of units registered at the time of the survey. The most used data collection system was reported to be REDCap academic at 33.3% of CTUs using this as their main system. Regardless of the system being utilised, 22 (81.5%) used paper test scripts to validate the eCRFs that had been initially configured. Only 22.2% used a risk-proportionate approach to determine whether every eCRF required testing. Statistical analysis, using Fisher’s exact tests, showed both the testing method ( p = 0.025) and the team size ( p = 0.043) did have an association with the use of a risk-proportionate approach to determine what eCRFs are tested. Kruskal–Wallis tests showed no significant relationships between the continuous variables and categorical outcomes that were tested. Conclusions The survey showed that there are varying EDC systems used, various testing approaches, and the implementation of risk-proportionality with regard to eCRF testing across the UKCRC-registered CTU network differed. Significant associations between the team size or the testing method and the use of a risk-proportionate approach to determine what is tested suggest that organisational and procedural factors shape how risk-based testing is applied. However, the small sample size ( n = 27) limited statistical power. Sharing of eCRF testing processes including how to utilise risk-proportionality and generate testing paperwork by CTUs would be useful to generate consistency and minimum standards for CTUs to enable good practice to be shared by all and support the units with limited capacity to be able to extend/review and develop their testing processes.