Comparing risk-adjusted and unadjusted CUSUM monitoring in Dutch liver procurement quality control: Evaluating the potential in a retrospective setting

Background: Risk-adjusted CUSUM outcome monitoring adjusts for risk factors that may influence alarm signaling in heterogeneous datasets. This study aimed to compare risk-adjusted and unadjusted CUSUM graphs for iatrogenic injury and discard rates in liver organ procurement in the Netherlands. Methods: Data on iatrogenic injury (C-event) and discard rates post-injury (C2-events) from liver graft procurements (Sep 2010 - Oct 2018) were collected. Multivariate analysis identified predictive risk factors, and risk-adjusted CUSUM graphs were plotted and visually compared to unadjusted charts. Results: Ten out of the 12 CUSUM charts demonstrated a change after adjustment (9 showing a reduction and 1 showing an increase). Risk adjustment resulted in the elimination of alarms for one local team, while another local team experienced earlier and longer alarms, with no significant impact on alarm signaling in the remaining charts. Conclusion: Risk-adjusted CUSUM substantially corrects most unadjusted charts, but its monitoring value depends on the chosen alarm threshold and the quality of the risk model, which may require periodic updating.