The use of computerized physician order entry with clinical decision support reduces practice variance in ordering preoperative investigations: A retrospective cohort study.

BACKGROUND AND GOAL OF STUDY: Over-ordering of routine preoperative investigations is prevalent. Adherence to institutional guidelines differs among physicians. Our institution integrated a Clinical Decision Support (CDS) model into our hospital's Computerized Physician Ordering Entry (CPOE) system to guide physician orders. We investigate if the implementation of CDS into CPOE increases physician adherence to our institutional guidelines.

METHODS: A retrospective cohort study in a tertiary academic hospital over 18 months. The CDS model incorporated into the hospital's CPOE system prompts preoperative investigations based on the patient's age, gender, American Society of Anesthesiologists (ASA) score and complexity of the surgery when physicians use the Electronic Admission Form (EAF). These investigations include: Full Blood Count (FBC), Chest Radiography (CXR), Coagulation Panel (CP), Renal Panel (RP) and Electrocardiogram (ECG). Orders are 'concordant' if they followed guidelines; 'over' if they were not required by guidelines, and 'under' if they were required by guidelines but not ordered.

RESULTS AND DISCUSSION: 11,792 patients - 7977 patients in the pre-implementation group, and 3815 patients in the post-implementation group. After implementation of CDS, overall guideline-concordant ordering rate increased by 3.9%, over orders decreased by 0.6% and under-orders decreased by 3.3% (P<0.001). CP showed the greatest increase in concordant orders by 12.0% and greatest decrease in under-orders by 11.1%. RP, ECG and CXR also showed modest increases in concordance rate. No significant change in ordering of FBC was found, due to the high pre-implementation concordance frequency of 96.3%. ECG and CXR have the lowest rates of concordant orders and highest rates of 'over' orders in both groups. Concordant orders were lowest in ASA 1 patients, and better in patients with higher ASA. Concordant orders across all ASA scores improved significantly after CDS was implemented.

CONCLUSION: Implementation of CDS model into the CPOE system has improved physician adherence to guidelines for certain preoperative investigations.