Benefits of an automated postoperative delirium risk prediction tool combined with non-pharmacological delirium prevention on delirium incidence and length of stay: a before-after analysis based on a quality improvement project

Background: Postoperative delirium (POD) significantly impacts older surgical patients, necessitating effective prevention strategies. Objective: To assess the effectiveness of the Pre-Interventional Preventive Risk Assessment (PIPRA) automated delirium risk prediction tool alongside non-pharmacological prevention strategies on POD incidence, hospital length of stay (LOS) and nursing time. Methods: This quality improvement project, set in a 335-bed Swiss private hospital, employed a before–after design to evaluate the impact of PIPRA and preventive measures on POD, LOS and nursing time in non-cardiac and non-intracranial surgery inpatients aged 60 or older. The control phase focused on enhancing POD screening, whilst the intervention phase incorporated PIPRA for risk assessment and staff training to enable targeted non-pharmacological prevention in patients at risk. Results: A total of 866 patients were included; 299 control and 567 intervention. The odds ratio of POD, comparing the intervention group to the control, was 0.71 [95% confidence interval (CI) 0.44–1.16] when adjusting for baseline patient characteristics. The intervention was associated with an LOS 0.94 (95% CI 0.85–1.05) and nursing time 0.96 (95% CI 0.86–1.07) times that of the control, adjusted for baseline patient characteristics. Medium risk patients (21.6% of patients) had an LOS 0.74 (95% CI 0.59–0.92) and required nursing time 0.79 (95% CI from 0.62–1.00) times the control, adjusted for baseline patient characteristics, equivalent to an LOS reduction of 1.36 days and nursing time saving of 19.3 hours per patient. Conclusions: Medium risk patients in the intervention group had shorter LOS and nursing time compared to the control group, underscoring the importance of targeted prevention.