Comparative Health System Performance Initiative: National Bureau of Economic Research Center of Excellence
David Cutler, Ph.D.
- Harvard Medical School (HMS) Department of Health Care Policy.
- Harvard Business School.
- The Network for Regional Healthcare Improvement.
- Massachusetts Health Quality Partners.
- Comagine Health (formerly HealthInsight).
- Center for Improving Value in Health Care.
- HMS Faculty at Dana Farber Cancer Institute.
- HMS Faculty at Boston Children’s Hospital.
The National Bureau of Economic Research (NBER) Center of Excellence and its team of collaborators are working to identify relationships among the organization of healthcare providers, delivery of evidence-based care, clinical and economic outcomes, and patient experiences. To this end, Center has mapped health systems across the entire United States over many years. The NBER team has developed some new performance measures and has calculated new and existing performance measures to study variation in performance across organizational types, health systems, and geographic areas. Finally, the team combines data on healthcare organization and performance to study topics such as the consequences of corporate consolidation and the diffusion of best practices.
The aims of the NBER Center of Excellence are to:
- Characterize the current delivery system and track changes in delivery system structure across space and over time.
- Gather data on the use of evidence from patient-centered outcomes research (PCOR) and related clinical and economic outcomes and merge those data with delivery system information.
- Use these data sources to examine associations between differences in delivery system structure and use of PCOR-based evidence and related clinical and economic outcomes.
Research from the NBER Center of Excellence is conducted at both the national and State levels. Analyses measure health system organization and performance at a point in time, changes over time, and across geographic areas. Under this initiative, five research projects are underway to examine delivery system performance in several distinct areas by looking across a variety of delivery systems (Projects 1 and 2); considering specific populations where evidence-based care involves tailoring care to the needs of the population, such as oncology, children, and post-acute care (Projects 3, 4, and 5); and considering the outcome of corporate integration in a number of settings (Projects 1 and 5):
- Delivery System Structure and Outcomes: A National Look. Project 1 focuses on how economic outcomes are affected by delivery system organization and ownership. The team has been working to describe these relationships by characterizing the organization of care and variation in clinical and economic outcomes across patients and groups of providers. Examples of studies from this project include:
- An examination of organizational features of health systems in the United States (e.g., size, composition, service scope, geographic scope, governance, mission) and the extent to which these features are associated with PCOR-based clinical, patient experience, and spending measures of performance (ongoing).
- An examination of the effects of the 340B Drug Pricing Program on hospital–physician consolidation and on the outpatient administration of parenteral drugs by hospital-owned facilities in three specialties in which parenteral drugs are frequently used.
- An examination of the relationship between physician practices’ commercial prices for office visits and the quality and use of care among fee-for-service Medicare patients served by the practices.
- An examination of the association between horizontal integration in the hospital industry and several dimensions of healthcare quality (ongoing; preliminary findings to be presented at the 2019 congress of the International Health Economics Association).
- Delivery Systems and Outcomes in Four States. Project 2 focuses on how care varies within and across four States (Colorado, Massachusetts, Oregon, and Utah). These States have different health systems and all-payer claims databases (APCDs), which have facilitated comparisons in care delivery across settings and tracking of the use of PCOR-based evidence and clinical and economic outcomes over time. Examples of studies from this project include:
- A descriptive account of the process followed to produce healthcare quality and cost measures across and within APCDs from four States to describe how APCDs can be used for multistate analysis.
- An examination of how care differs for Medicaid and commercially insured patients in four States, focusing on the extent to which these patient populations go to the same or different providers and characterizing the degree of segregation in provider choices (ongoing; presented at ASHEcon 2019, the eighth annual conference of the American Society of Health Economists).
- Characteristics of High-Performing Delivery Systems of Cancer Care. Project 3 focuses on how the organization of oncology care affects the use of PCOR-based processes and related clinical and economic outcomes in adult patients with cancer. Examples of studies from this project include:
- A descriptive study of the number and specialties of oncology physicians across the United States, including their prevalence in National Cancer Institute cancer centers and academic medical centers (ongoing).
- An examination of variation in the quality of end-of-life cancer care within and across integrated delivery system types and the persistence over time and consistency across measures of high performance on select quality measures (ongoing; presented at AcademyHealth 2019).
- Accelerating the Performance of Pediatric Health Systems. Project 4 focuses on pediatric features of health system structures and examines the degree to which these features relate to PCOR-based care quality. Examples of studies from this project include:
- An examination of the prevalence and scope of pediatric services in healthcare systems, including the extent to which pediatric expertise is present in systems; the extent to which pediatric-serving entities can provide services for low-, medium-, high-, and highest risk pediatric patients; (ongoing; presented at Pediatric Academic Societies 2017 and AcademyHealth 2017).
- An examination of trends in hospital-based pediatric services between 2012 and 2016, including the degree to which hospital characteristics and geographic location explain variation in inpatient pediatric services over time, as well as whether service availability corresponds to greater or lesser healthcare accessibility (ongoing; presented at Pediatric Academic Societies 2019 and AcademyHealth 2019).
- Post-Acute Care and Dialysis. Project 5 focuses on consolidation and organizational change in two understudied healthcare industries that account for a significant share of total medical spending: post-acute care facilities and dialysis facilities. The project examines the impact of these facilities on the use of PCOR-based evidence, spending, and clinical outcomes. Examples of studies from this project include:
- An examination of common investor ownership linkages across the acute care, post-acute care, and hospice sectors within the same geographic markets.
- An examination of the impact of vertical integration between hospitals and skilled nursing facilities on cost and quality of healthcare, focusing on self-referrals to hospital-owned skilled nursing facilities (ongoing).
Critical to the projects described above are the data to describe healthcare organization and to measure performance:
- An "enhanced system database." This relational database contains micro data on health systems, hospitals, physicians, and other healthcare providers (e.g., post-acute care facilities) for 2010–2016. Later years will be added as new data become available. The NBER team uses a network algorithm that combines data from a wide variety of sources (e.g., Medicare Provider Enrollment, Chain, and Ownership System [PECOS], Medicare and commercial claims, SK&A, Internal Revenue Service form 990 filings, American Hospital Association survey data) to identify health systems as groups of commonly owned or managed providers. The team identifies health systems in claims data using tax identification numbers and Medicare provider numbers (e.g., National Provider Identifiers and CMS Certification Numbers). They link these providers to the markets in which they participate (e.g., hospital referral regions, primary care service areas) and characterize the local populations.
- Administrative and claims data that can be used to measure the use of PCOR-based evidence, related clinical outcomes, and cost of care. These include Medicare claims data, national commercial insurance claims data, cancer registry data, dialysis facility and cost reports, Medicare cost reports, the Healthcare Cost and Utilization Project's Kids' Inpatient Database, Centers for Medicare & Medicaid Services Hospital Compare, and APCD data for four States.
- Patient and family self-reports about the quality of care received. These data include the Consumer Assessment of Healthcare Providers and Systems, Hospital Consumer Assessment of Healthcare Providers and Systems, and other sources for family surveys.