Leveraging real-world data for continuous evaluation of computational clinical practice guidelines.

Area of Science:

• Oncology
• Health Informatics
• Clinical Practice Guidelines

Background:

• Clinical practice guidelines (CPGs) and clinical care have a bidirectional relationship, necessitating methods for monitoring guideline adherence.
• Advancements in computable CPGs enable automated evaluation using real-world registry data.
• Assessing adherence to Dutch endometrial cancer (EC) guidelines is crucial for improving care.

Purpose of the Study:

• To evaluate the feasibility of using real-world registry data to assess adherence to Dutch EC guidelines.
• To establish a framework for continuous monitoring of guideline adherence and trends.
• To identify factors influencing adherence to EC guidelines.

Main Methods:

• A retrospective cohort study utilizing real-world data from the Netherlands Cancer Registry (NCR) (January 2010 - May 2022).
• The Dutch EC guideline was translated into clinical decision trees (CDTs).
• Guideline adherence was computed for various patient and disease subgroups.

Main Results:

• The Dutch EC guideline was parsed into 10 CDTs, encompassing 22 patient/disease characteristics and 46 interventions.
• Data from 21,602 EC cases were analyzed, with four CDTs successfully populated.
• Mean adherence was 82.7% (range 44-100%), with three significant adherence trends identified.

Conclusions:

• A novel framework for continuous evaluation of adherence to cancer guidelines was introduced.
• Integration of real-world data with computer-interpretable guidelines effectively calculated adherence facets for EC.
• Future research should incorporate health outcome measurements into adherence evaluations.