Evidence, Incentives and Uncertainty: the Evolution of Pay-for-Performance in Health Care Glen P. Mays, PhD Department of Health Policy and Management nt Fay W. Boozman College of Public Health University of Arkansas for Medical Sciences
Getting what we pay for? WHO 2005
Getting what we pay for? Preventable Deaths per 100,000 population Source: Commonwealth Fund 2008
Getting what we pay for?
Getting what we pay for? McGlynn et al. New England Journal of Medicine 2003
McGlynn et al. 2007 Drivers of P4P Under-use of effective care
Drivers of P4P Over-use of costly, equivocal care
Drivers of P4P Missed opportunities for prevention Less than 50% of the population at risk is reached by: Smoking cessation Aspirin use Colorectal cancer screening Influenza vaccines Nutrition and physical activity programming Family yplanning
Evolution of Incentives There are many ways to pay a physician fee for service, salary, capitation and they are all bad. James Robinson, UC Berkeley, 2000
Evolution of Incentives Pay for infrastructure (IT, EMR) Pay for reporting Public reporting Performance incentives Delivery and payment reform: ACOs, bundled payments
How are P4P Programs Structured? Physicians (medical groups) about twice as likely as hospitals to be target Average of 5 performance measures Maximum bonus 5-10% of pay for physicians, 1-2% for hospitals Rewards for reaching fixed threshold dominate; only 23% reward improvement Rosenthal, 2007
Increasing inclusion of specialists and hospitals in P4P Proportion of Private P4P Programs with Selected Providers Rosenthal, 2007
Increasing emphasis on outcomes, IT, efficiency Proportion of Private P4P Programs with Selected Measures Rosenthal, 2007
Evidence of impact to date Rigorous studies of pay-for-performance in health care are few (17 since 1980) Overall findings are mixed: many null results even for large dollar amounts But in many cases negative findings may be due to short-term nature, small incentives Evidence suggests pay-for-performance performance can work but also can fail Rosenthal, 2007
Evidence of impact to date Pacificare P4P Program Results: Cervical Screening Schaefer, 2009
Lingering concerns and issues Accuracy and completeness of measures Small numbers, unstable measures Risk adjustment & case mix Unintended consequences Effects on non-measured domains of performance Disparities in care Costs vs. benefits
On the horizon: new perinatal quality measures National Quality Forum Endorsed Measures Elective delivery prior to 39 weeks Incidence of episiotomy Cesarean rate for low-risk first birth Prophylactic antibiotics in c-section DVT prophylaxis pop ya in csec c-section Birth trauma rate Appropriate use of antenatal steroids Infants<1500g delivered at NICU Exclusive breastfeeding at discharge Hep B vaccine at discharge Neonatal bloodstream infections for infants <1500g or <30wks First NICU temperature <36C for infants <1500g Neonatal retinopathy screening for infants 22-29wks 29wks Surfactant within 2 hours of birth for infants 22-29wks NQF 2010
Drivers of perinatal performance in Arkansas Highest-risk infants residing >80 miles from NICU-equipped hospitals were more likely to be delivered at these hospitals after ANGELS implementation (26% increase) 40% 35% 30% 25% 20% 15% 10% 5% 0% <1000 grams 1000-1499 grams 1500-1999 grams 2000-2499 grams 2001 2002 2003 2004 ANGELS Regression-adjusted estimates controlling for maternal risks, insurance source, socioeconomic characteristics, and race/ethnicity. p<0.05
Disruptive Innovation: PPACA Overview Expansion in health insurance coverage Incentives and demonstrations for delivery system reform Expanded investments in prevention and wellness Expanded workforce development investments Emphasis on accountability and transparency Emphasis on evidence-based approaches Imperative for efficiency & cost containment Public health and health system integration Combined impact of HITECH Act: EMR and HIE
Delivery System Incentives & Demos Accountable care organization models: bundled payments, global payments, shared savings models (2011-13) Medicaid Health Home & medical home programs (2011) Medicaid home visiting programs (2011) Community Health Worker programs (2010) Reduced Medicare & Medicaid payments for HAIs and preventable readmissions i (2012) Reduced DSH payments, MA payments (2011) Comparative Effectiveness Research: Patient Centered Outcome Research Institute
Accountable Care Organization Models Integrate inpatient and outpatient care providers in a defined geographic area Bundled, global payments for defined episodes of care: inpatient and outpatient Performance measures used to adjust payments Shared savings provisions to incentivize efficiency
Accountable Care Organization Models Prometheus ACO Model: Geisinger Health System (PA) De Brantes, 2009
Conclusions Health reform and HITECH will fuel measurement, P4P, and ACO models Arkansas has unique opportunities and challenges Small practices, rural and remote locations Medically underserved areas Telemedicine ANGELS AHECs, CHCs, and public health clinics Time and resources are available NOW to develop and test prototypes prior to 2014