Predicting Fragmented Care: Beneficiary, Physician, Practice, and Market Characteristics.

Area of Science:

• Health Services Research
• Health Economics
• Medical Informatics

Background:

• Fragmented ambulatory care, defined as care spread across multiple providers without a dominant one, presents challenges in healthcare delivery.
• Understanding the drivers of this fragmentation is crucial for designing effective interventions to improve care coordination.

Purpose of the Study:

• To identify patient, primary care physician (PCP), practice site, and geographic market characteristics associated with highly fragmented ambulatory care within the US Medicare population.
• To quantify the proportion of variation in care fragmentation explained by observed factors.

Main Methods:

• Cross-sectional analysis of Medicare claims data from 2018, including over 3.5 million beneficiaries.
• Hierarchical linear models with random intercepts were used to predict high care fragmentation, defined as a reversed Bice-Boxerman Index score above 0.85.
• A comprehensive set of explanatory variables related to beneficiaries, physicians, practice sites, and markets was included.

Main Results:

• Observed characteristics of beneficiaries, PCPs, practice sites, and markets explained only 6% of the variation in highly fragmented care.
• Unobserved differences (physician, practice, market) accounted for an additional 4%, leaving 90% as unobserved residual variance.
• Patient demographics, PCP comprehensiveness, and practice characteristics like size or system affiliation showed only small associations with fragmentation.

Conclusions:

• Observed patient, provider, and practice-level factors do not adequately explain the variation in ambulatory care fragmentation.
• Unmeasured patient preferences and behaviors, alongside the aggregate actions of healthcare providers beyond primary care, appear to be significant predictors of fragmentation.