Feasibility and safety of a perioperative clinical decision support system: results of a multicenter randomized controlled trial
L Milz, C Milz, C Wagner, S Hottenrott, A Hennemuth, R Alpers, M Huellebrand, P Boerm, K Zacharowski, C Spies, L Sundmacher, S Nagel, P Meybohm, A GhanemAbstract
Background
Digital clinical decision support (CDS) systems are increasingly implemented to standardise perioperative risk assessment and improve guideline adherence. However, randomised evidence regarding their safety in routine clinical practice remains limited.
Purpose
To evaluate the feasibility and safety of a guideline-based perioperative CDS system with respect to bleeding complications requiring transfusion.
Methods
In a prospective multicenter randomized controlled trial, 229 adult patients (ASA II–IV) undergoing elective non-cardiothoracic surgery were randomized to standard preoperative assessment with or without CDS support. The CDS system provided structured risk stratification and guideline-based management recommendations. The primary endpoint for this analysis was the occurrence of at least one perioperative red blood cell transfusion (intra- or postoperative). Analyses were performed in patients who underwent surgery. Odds ratios (OR) with 95% confidence intervals (CI) were calculated.
Results
Of 229 randomized patients, 223 underwent surgery and were included in the bleeding analysis (CDS n=120; control n=103). Transfusions occurred in 11.7% of patients in the CDS group and 7.8% in the control group (OR 1.57; 95% CI 0.58–4.51; p=0.374). In low/intermediate-risk procedures, transfusion rates were 10.3% vs 4.8% (OR 2.24; 95% CI 0.51–13.67; p=0.346). In high-risk procedures, rates were 25.0% vs 29.4% (OR 0.80; 95% CI 0.16–4.16; p=1.000). No statistically significant differences were observed. Confidence intervals were wide due to limited event numbers.
Conclusion
Implementation of a guideline-based perioperative CDS system was feasible and not associated with an increased risk of bleeding requiring transfusion. These findings provide randomized evidence supporting the safety of digital decision support integration into perioperative clinical workflows. Larger adequately powered studies are required to assess potential effects on clinical outcomes.Flow-ChartForrest-Plot