Importance of health system context for evaluating utilization patterns across systems.

Area of Science:

• Health Services Research
• Health Informatics
• Data Science

Background:

• Measuring patient healthcare utilization across diverse health systems is complex.
• Data from single systems omit out-of-system care, leading to incomplete utilization patterns.
• Varying data-generating processes (DGPs) across systems complicate cross-system data aggregation and accuracy.

Purpose of the Study:

• To develop and validate a hybrid methodology for aggregating health-care utilization data across multiple health systems.
• To address challenges in cross-system data aggregation and accuracy due to differing DGPs.
• To enable valid cross-system assessments of patient healthcare utilization.

Main Methods:

• Developed a hybrid methodology integrating strengths from different system-specific DGPs.
• Utilized positive and negative predictive probabilities to assess the hybrid methodology's impact.
• Compared public sector (Department of Veterans Affairs) and private sector (Medicare) patient-level data.

Main Results:

• The hybrid methodology demonstrates validity for cross-system health-care utilization assessments.
• Illustrates challenges and solutions in combining administrative and billing records.
• Highlights the importance of understanding system context and its effect on DGPs.

Conclusions:

• A hybrid approach can effectively aggregate data across disparate health systems for comprehensive utilization analysis.
• Understanding system-specific DGPs is crucial for accurate cross-system health-care evaluations.
• The proposed methodology enhances the accuracy of patient-level health-care utilization studies.