Dissertation Title:   "Health Systems, Quality and Variation"

Integrated health systems are becoming more important due to increasing trends toward consolidation in the midst of rapidly expanding health care reform efforts enacted by the Affordable Care Act. Consolidated health systems have advantages which include the ability to coordinate care across providers and care settings. Conceptually, health systems might be better equipped to achieve the breadth of alignment (i.e., social, functional and clinical) required to redesign and sustain care processes to achieve quality improvement due to greater capacity for monitoring and managing performance. However, enormous variation exists across health systems in terms of their organizational characteristics and operations. Yet, little knowledge exists regarding how or why quality of care varies at this level due to a lack of systematically collected data regarding health system affiliations, system characteristics or their behaviors to achieve true clinical integration to further quality goals.

In Chapter 1, we assess variation in quality performance across and within health systems to understand the extent to which health systems differ in their overall quality and the care they provide to vulnerable populations. I find that there is small variability in overall care quality across health systems; system-level intraclass correlation coefficients (ICCs) for quality outcomes are generally low, ranging from 0.05% to 1.79%. Although ICCs were small, system-level means ranged from 75%-80%, 73%-81%, and 67%-77% for the A1c, LDL and breast cancer screening measures, respectively. We find that there are systematic differences in observable quality performance for certain vulnerable patient groups compared to non-vulnerable patient groups and there is large variation in the extent of health disparities across health systems. Health systems are uniquely positioned to improve care delivery and care quality, making a strong case deployment of an incentive-driven system directed at health systems aimed at standardizing high quality care delivery for patients in certain demographic groups. Study findings support the notion that monitoring quality of care at the health system-level is valuable for advancing health care reform.

In Chapter 2, we examine the empirical relationships between individual structural features of the organization and quality. We document statistically significant relationships among the constructed structural characteristics studied and quality performance. For example, systems with greater primary care orientation are associated with better access to timely care, a greater perception on behalf of the patient of care coordination activities and fewer 30-day readmissions. Systems with more specialists are also associated with more 30-day readmissions. Patients attributed to larger and more dispersed systems report more positive care ratings and receive more guideline-concordant screenings for A1c and breast cancer but are found to experience more 30-day readmissions. Patients attributed to more dispersed systems also report higher perceptions of care coordination activities. Our findings indicate that structural characteristics of health systems are associated with quality and patient utilization of health care services. Health system leaders should be aware of their influence and should consider strategies to overcome or capitalize on the impacts of their structural characteristics to deliver care that is more integrated and of better quality.

In Chapter 3, we investigate the association between survey-derived measures of social and functional integration and care quality in a sample of Medicare beneficiaries attributed to fifty-nine practice sites in health systems across four states. Our analyses uncovered a strong positive relationship between patient integration, a form of social integration, and the quality of care a beneficiary receives. Further, we document a significant moderation by beneficiary clinical risk for aspects of both social integration and functional integration which suggest that certain types of integration may matter more for quality among certain patients. These findings indicate the importance of functional and social forms of integration, which have to date received less attention in research on integration than ownership and other structural integration forms. Policymakers and practitioners considering structural integrations of health systems should direct attention beyond structure to consider the potential for social and functional integration to impact outcomes and how they might be achieved.