Internal and External Factors Conducive to Hospital Participation in CMS Accountable Care Organizations (ACO) Askar S. Chukmaitov, M.D., Ph.D. David W. Harless, Ph.D. Gloria J. Bazzoli, Ph.D. Yangyang Deng, M.S. Virginia Commonwealth University (VCU) Agency For Healthcare Research and Quality, Grant #R01 HS023332
Objectives To identify internal and external factors conducive to hospital participation in the Centers for Medicare and Medicaid Services’ (CMS) Medicare Shared Savings Program (MSSP) and Pioneer Accountable Care Organizations (ACOs) Internal: hospital structures and processes-of-care External: environment and market characteristics
Rationale for Study ACOs may seek hospital and health system participation as they can provide: Organizational and legal framework for comprehensive care delivery Effective care coordination and transitions of patients across the care continuum A broad array of health services addressing needs of Medicare beneficiaries Participating ACO hospitals may provide resources and invest in: Health Information Technology Performance measurement, data sharing, and reporting systems for managing population health
Conceptual Framework Favorable Structures: • More centralized health systems • Experience managing risk-based contracts • Tight physician-hospital linkages • More advanced HIT • Focus on preventative services • High structural quality score • Links with providers on care continuum MSSP: One-sided, Shared Savings Processes of care: • Evidence-based care processes in place Pioneer ACO: Two-sided , Shared Savings/Losses Environmental Munificence: • Availability of primary care physicians • Specialists • Additional Resources • Geographic location: rural vs. urban
Methods (1) Data: Participating ACO hospitals were identified through CMS announcements and lists, case-by-case evaluation, and reaching out to ACO coordinators 32 Pioneer ACO (2012) and 355 MSSP ACO (2012, 2013) Secondary data - AHA, CMS demonstration pilots, HIMSS, CMS Hospital Compare, and AHRF Sampling: All nonfederal, short-term, general medical-surgical hospitals with complete data from 50 states for 2011, i.e. prior to ACO formation: 105 (out of 158) hospitals in Pioneer ACOs 340 (out of 521) hospitals in MSSP 3,296 non-participating hospitals (2012)
Methods (2) Analytic Approach: The multinomial probit for models with three outcomes: Not participating in an ACO (referent) Participating in MSSP ACO Participating in Pioneer ACO Analyzed whether favorable structures, processes-of-care, or environmental munificence were more important for hospital participation in CMS ACOs Compared composite measures of more favorable vs. less favorable internal and external characteristics for hospitals in MSSP, Pioneer ACOs, and non-participating hospitals Supplemental analysis for hospital ACO participation, excluding 28 hospitals in Pioneer and 8 hospitals in MSSP ACOs that ceased participation in 2014.
Methods: Key Measures (1) Favorable Structures Hospital’s health (1) Centralized Health Systems (CHS); (2) Moderately system affiliation Centralized Health Systems (MCHS); (3) Decentralized Health Systems (DHS); (4) Freestanding Hospital (referent) PGP transition or (1) Yes; (2) No (referent) for participation in CMS pilots for HCBDP demonstration establishing ACO-like risk-based payments and care management programs Physicians in tightly integrated POAs – Management Service Number of Physicians in Tight POA Organization (MSO), Integrated Salary, Equity, and Foundation measured in hundreds Count of More (1) Nursing documentation and (2) Electronic medication Advanced HIT administration. Count of Preventative (1) Breast cancer screening; (2) Community outreach; (3) Crisis Services prevention; (4) Community health education; (5) Screenings; (6) Immunizations; (7) Indigent care; (8) Patient education center; (9) Patient representatives; (10) Social work; (11) Transportation to services Structural Quality (1) Highest quartile for inpatient admissions; (2) Joint Commission’s accreditation; (3) Commission of Cancer’s Score accreditation; (4) transplant services; (5) level I trauma center; (6) highest quartile for nurse-to-bed ratio; (7) teaching status
Methods: Key Measures (2) Favorable Structures (cont.) Linkages with (1) Freestanding outpatient center, (2) Hospital-based Ambulatory Facilities outpatient care center, (3) Primary care department, (4) Home (Count of five linkages) care, and (5) Urgent care. Processes of Care Hospital’s performance on 22 accountability measures for Total Joint Commission (TJC) Composite Score evidence-based processes-of-care for heart attack, pneumonia, surgical care were developed. If a hospital’s scored high on care for all four conditions, then it was rank in (1) Low composite score tier; if scored high on three conditions, then (2) Medium tier; if scored high on two condition, then (3) High tier Environmental Munificence Primary Care Sum of physicians in internal medicine, family practice, and Physicians Supply pediatrics per 1,000 residents in a county Specialists Supply (Total number of physicians - physicians in internal medicine, family practice, and pediatrics) per 1,000 population in a county Median Income Median household income in a county (in thousands of dollars) Hospital location (1) Rural; (2) Urban (referent)
Table1: Variable Means (Standard Deviations) in 2011, by 2012-2013 CMS Accountable Care Organization Participation. p -value Variable Nonparticipant MSSP Pioneer Centralized Health System 0.096 (0.295) 0.259 0.229 p<0.001 (0.439) (0.422) Moderately Centralized Health 0.146 (0.353) 0.282 0.352 p<0.001 System (0.451) (0.480) Decentralized Health System 0.242 (0.428) 0.238 0.190 0.48 (0.427) (0.395) PGP Transition or HCBDP 0.005 (0.070) 0.056 0.086 p<0.001 Demonstration (0.230) (0.281) Preventative Services Scope and Mix 6.37 (2.47) 7.39 (2.16) 7.35 (2.21) p<0.001 Number Physicians in Tight POAs 0.30 (1.23) 0.66 (1.84) 1.13 (4.00) p<0.001 (in hundreds) Count among More Advanced HIT 1.33 (0.85) 1.58 (0.75) 1.61 (0.70) p<0.001 Applications Count of Linkages with Ambulatory 2.12 (1.33) 2.43 (1.30) 2.21 (1.23) p<0.001 Facilities Structural Quality Score 1.73 (1.60) 2.51 (1.63) 2.54 (1.64) p<0.001 TJC Composite First Tier 0.128 (0.335) 0.132 0.190 0.18 (0.339) (0.395) TJC Composite Second Tier 0.111 (0.314) 0.171 0.143 0.003 (0.377) (0.352) TJC Composite Third Tier 0.096 (0.295) 0.115 0.190 0.004 (0.319) (0.395) PCPs per 1,000 Population 0.68 (0.39) 0.88 (0.56) 0.93 (0.48) p<0.001 Specialists per 1,000 Population 1.04 (1.19) 1.62 (1.69) 1.67 (1.55) p<0.001 Median Income (in thousands of 46.7 (12.0) 51.3 (12.1) 55.2 (12.9) p<0.001 dollars) Rural 0.266 (0.442) 0.121 0.095 p<0.001 (0.326) (0.295) �
Table 2: Estimates of Marginal Effects from Multinomial Probit Model of CMS ACO Participation. a Excluding Hospitals Full Estimation Ceasing Participation Sample by the End of 2014 Variable MSSP Pioneer MSSP Pioneer 0.1273 *** 0.0216 * 0.1303 *** Centralized Health System 0.0059 (0.0184) (0.0093) (0.0186) (0.0076) 0.1002 *** 0.0276 ** 0.0952 *** 0.0186 * Moderately Centralized Health System (0.0151) (0.0088) (0.0149) (0.0080) 0.0540 *** 0.0533 *** Decentralized Health System 0.0024 -0.0059 (0.0114) (0.0062) (0.0114) (0.0054) 0.2686 *** 0.0969 * 0.2813 *** 0.0873 * PGP Transition or HCBDP Demonstration (0.0649) (0.0444) (0.0656) (0.0401) 0.0028 * 0.0025 * Number Physicians Tight POAs (in hundreds) -0.0000 0.0004 (0.0028) (0.0013) (0.0028) (0.0011) 0.0363 *** 0.0349 *** 0.0092 * Count of Advanced HIT Applications 0.0042 (0.0077) (0.0045) (0.0078) (0.0044) 0.0320 * 0.0414 *** PCPs per 1,000 Population 0.0376 0.0193 (0.0229) (0.0145) (0.0239) (0.0122) 0.0576 * 0.0300 * 0.0597 ** 0.0383 ** ln (Median Income) (0.0230) (0.0136) (0.0227) (0.0120) 0.0793 *** 0.0259 *** 0.0770 *** 0.0144 * Not-for-profit Ownership (0.0106) (0.0065) (0.0108) (0.0063) -0.0383 * -0.0347 ** -0.0445 * -0.0295 ** Herfindahl-Hirschman Index (0.0192) (0.0120) (0.0192) (0.0105) a Standard errors of marginal effects in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001
Results – Main Effects (1) Hospital favorable structures were much more important then environmental factors for hospital participation in both MSSP and Pioneer ACOs Several factors were important for hospital participation that were estimated in both MSSP and Pioneer ACO models: System affiliation Prior experience with risk-based payments More advance HIT Non-profit ownership Location in counties with higher median income and hospital competition
Results – Main Effects (2) Hospitals in decentralized health systems were likely participants in MSSP ACO Hospital participating in Pioneer ACO were more likely to be in: Centralized and moderately centralized health systems Tightly integrated hospital – physician arrangements Located in areas with sufficient supply of primary care physicians
Recommend
More recommend
