Medicare has steadily moved toward value-based and alternative payment models, in which clinicians and hospitals are held accountable for both quality and costs of care. In theory, this makes sense, as the goal of value-based payment models is to provide incentives for delivering high-quality care.
However, inherent limitations of these models as they now stand may incur unintended consequences. According to two new studies published in JAMA Internal Medicine, both clinicians and hospitals may be penalized for caring for sicker patients, those with mental health/cognitive impairments, and patients at lower socioeconomic status.
In an editorial accompanying both studies, Julie Bynum, MD, MPH, from the University of Michigan Medical School, Ann Arbor, and Valerie Lewis, PhD, from the University of North Carolina at Chapel Hill, point out that with value-based payment models, there may be benefits to avoiding treating high-risk populations.
This can be an “appealing option for physician organizations, hospitals, or payers concerned that they will need to expend more resources for certain patients than they will receive to care for them,” they write.
Known by many names, including adverse selection, cherry picking, cream skimming, and patient dumping, this phenomenon “has been found in a variety of contexts related to quality reporting or pay for performance.”
This type of “adverse selection” represents a serious threat to the success of value-based payment models, but the biggest potential harm is to the high-risk patient, who may have limited access to high-quality clinicians, the editorialists note.
High Cost of Mental Health
The Merit-Based Incentive Payment System (MIPS), which adjusts payment to clinicians in the outpatient setting by ±4% on the basis of cost and quality performance, could penalize providers “inappropriately” on the basis of factors they are unable to control, the authors note.
The broadest cost measure in MIPS is the total per capita costs (TPCC), which evaluates providers on costs for all attributed Medicare patients. However, the TPCC measure fails to adjust for patient functional status and common neuropsychological comorbidities or for “local area supply-side” and other variables related to economic conditions.
“These unaccounted for patient risk factors are known to be associated with higher health care utilization and costs and are largely outside of outpatient clinicians’ control,” write the authors.
The Medicare Current Beneficiary Survey was used to evaluate patient-reported neuropsychological and functional status. The cohort included beneficiaries served by both safety-net (federally qualified health centers and rural health clinics) and non-safety-net clinics. The analysis included 111,414 unique physicians and 30,058 unique patients (mean age, 71.84 years).
The mean total annual cost of care was similar for patients who used safety-net clinicians ($9160) and those who did not ($9112). But under the Centers for Medicare & Medicaid Services (CMS) hierarchical conditions category score, which includes conditions such as heart failure, diabetes, and kidney disease, the total cost of care was 9% greater than the expected costs. Conversely, the costs of non-safety-net providers were 1% lower than anticipated ($9112 vs $9199)..
The addition of neuropsychological and functional status variables was associated with higher Medicare spending. Those with higher total costs included patients with depression ($14,436) or dementia ($18,311) and those having difficulty with three or more activities of daily living (ADL) or instrumental ADL ($17 443), living in a mental health care shortage area ($9233), having a high
proportion of residents in poverty ($9569), or that were unemployed ($9658).
The addition of neuropsychological/functional factors and local residence area factors reduced outpatient safety-net clinicians’ underperformance on total cost of care by 5 percentage points, or 52%, as compared with non-safety-net providers.
“This implies that failing to account for these factors could penalize outpatient safety-net clinicians under the MIPS program who care for patients who have greater neuropsychological and functional health needs and are located in underresourced areas,” write Johnston and colleagues.
Hospital Readmission Rates
Medicare also ties payments to hospital readmission rates as a barometer for assessing quality of care in several of their pay-for-performance programs. One pressing concern is that CMS limits adjustments for a wide range of patient characteristics, which can affect readmission rates. Some evidence indicates that hospitals serving higher-risk patients have received larger penalties, and proposals have been made that additional patient characteristics, including social risk, be accounted for.
In this study, Eric T. Roberts, PhD, from the University of Pittsburgh, Pennsylvania, and colleagues compared readmission rates before and after adjusting for several sets of patient characteristics in addition to those CMS currently uses to adjust for readmission risk. The investigators used Medicare claims for admissions in 2013 through 2014. The study sample included 1,169,014 index admissions among 1,003,664 unique Medicare beneficiaries in 2215 hospitals.
The mean unadjusted readmission rate was 11.9%; rates ranged from 8.9% to 14.8% for hospitals in the lowest to the highest quintiles of readmission rates for all causes within 30 days of discharge.
When adjusted for additional clinical and social patient characteristics, the overall variation in hospital readmission was reduced by 9.6%. The rates changed upward or downward by 0.37 to 0.72 percentage points for the 10% of hospitals most affected by the additional adjustments (±30.3% to ±58.9% of the hospital-level standard deviation). These changes would reduce expected penalties by 52%, 46%, and 41% for hospitals with the largest 1%, 5%, and 10% of penalty reductions, respectively.
These additional adjustments also narrowed the mean difference in readmission rates between hospitals in the top and bottom quintiles of high-risk patients by 0.53 percentage points (95% CI, 0.50 – 0.55; P < .001), or 54%.
“Our results support policies to adjust readmission rates for a more comprehensive set of patient characteristics,” the authors conclude, “…as well as alternative strategies to improve quality and address disparities.”
Saint Louis University purchased and provided access to the data used in the Johnston study; coauthor Dr Joynt Maddox is supported by a grant from National Heart, Lung, and Blood Institute. The Roberts study was supported by grants from the National Institutes of Health, the Laura and John Arnold Foundation, and the Marshall J Seidman Center for Studies in Health Economics and Health Care Policy at Harvard Medical School. The authors and editorialists have disclosed no relevant financial relationships.