Radiology order decision support: examination-indication appropriateness assessed using 2 electronic systems

Area of Science:

• Radiology Informatics
• Health Information Technology
• Clinical Decision Support Systems

Background:

• Clinical decision support (CDS) systems are increasingly used to optimize medical imaging orders.
• Evaluating the effectiveness of different CDS systems is crucial for improving healthcare efficiency and patient outcomes.
• Guideline implementation strategies within CDS systems can significantly impact their performance.

Purpose of the Study:

• To compare the performance of two commercial radiology clinical decision support (CDS) systems, Medicalis SmartReq and Nuance RadPort.
• To assess the impact of different guideline implementation strategies on the appropriateness and insurance disposition of MRI and CT orders.

Main Methods:

• A comparative study evaluating approximately 2,000 outpatient MRI and CT orders over two 3-month periods.
• Utilized Medicalis SmartReq and Nuance RadPort CDS systems.
• Logistic regression analysis was employed to compare outcomes, adjusting for patient-mix differences.

Main Results:

• Nuance RadPort significantly outperformed Medicalis SmartReq in scoring exams as appropriate (80% vs. 51%) and predicting insurance outcomes (76% vs. 58%).
• Medicalis SmartReq classified a substantial portion of exams (60%) as "indeterminate" or "not validated".
• Reimbursement rates varied significantly based on utility ratings, with low-utility exams being reimbursed more often than high- or moderate-utility exams.

Conclusions:

• The primary difference influencing outcomes was how systems handled exams lacking guidelines or sufficient information.
• Nuance RadPort's approach of augmenting guidelines with best practices proved more effective than Medicalis SmartReq's request for additional information.
• User interface design and guideline implementation choices critically influenced CDS system interaction and user adoption.