Implementation of a software-based decision support tool for guideline-appropriate preoperative evaluation: a prospective agreement study.

Background: Guideline adherence in the medical field leaves room for improvement. Digitalised decision support helps improve compliance. However, the complex nature of the guidelines makes implementation in clinical practice difficult.

Methods: This single-centre prospective study included 204 adult ASA physical status 3-4 patients undergoing elective noncardiac surgery at a German university hospital. Agreement of clearance for surgery between a guideline expert and a digital guideline support tool was investigated. The decision made by the on-duty anaesthetists (standard approach) was assessed for agreement with the expert in a cross-over design. The main outcome was the level of agreement between digital guideline support and the expert.

Results: The digital guideline support approach cleared 18.1% of the patients for surgery, the standard approach cleared 74.0%, and the expert approach cleared 47.5%. Agreement of the expert decision with digital guideline support (66.7%) and the standard approach (67.6%) was fair (Cohen's kappa 0.37 [interquartile range 0.26-0.48] vs 0.31 [0.21-0.42], P=0.6). Taking the expert decision as a benchmark, correct clearance using digital guideline support was 50.5%, and correct clearance using the standard approach was 44.6%. Digital guideline support incorrectly asked for additional examinations in 31.4% of the patients, whereas the standard approach did not consider conditions that would have justified additional examinations before surgery in 29.4%.

Conclusions: Strict guideline adherence for clearance for surgery through digitalised decision support inadequately considered patients, clinical context. Vague formulations, weak recommendations, and low-quality evidence complicate guideline translation into explicit rules.

Clinical Trial Registration: NCT04058769.