Implementation of imitation modelling for quantative risk assessment in clinical trial data collection

Abstract

Listen

Translate article

Appropriate data collection is an essential aspect of clinical trial (CT) quality and includes accurate and precise data entered into a case report form (CRF). The paper presents the results of risk assessment and modelling for the data collection process in the trial site. In research 292 CRF were analysed and CRF corrections documentation for 12 CT carried out in Clinical and Diagnostic Centre of the National University of Pharmacy (CDC NPhaU). In order to evaluate the rate of errors made in CRF we proposed a key risk indicator – error coefficient (k n ) with specific intervals of values: optimal (0-3); threatening (3,1-5); critical (more than 5). The model of the error coefficient changes was created by Monte-Carlo method and probability of critical values of the error coefficient was 0.026. The economic consequences for these cases were predicted by modelling on the basis of expert estimates regarding CT project time and required costs. Probability of the maximal project time increase to 4 days was 0.019; probability of maximal additional costs up to 3% of the budget was 0.02. The results obtained indicate the high quality of the data collection process in the CDC NPhaU. The method proposed may be recommended for sponsors for risk-based monitoring implementation, for trial sites preparing to monitoring visits, audits or inspections, for regulatory bodies for planning inspections of trial sites.