New research suggests that Medicare’s pay-for-performance measures, including the Merit-based Incentive Payment System, unfairly penalize providers who serve older, sicker, poorer patients.
A study recently found that Medicare's Value-based Payment Modifier program inadvertently shifted money away from physicians who treated sicker, poorer patients to pay for bonuses that rewarded practices treating richer, healthier populations.
Study lead author Eric Roberts, an assistant professor of health policy and management at the University of Pittsburgh Graduate School of Public Health, says that if changes aren’t made, value-based payment models will continue to shortchange the poor.
Roberts spoke with HealthLeaders Media about his findings. The following is a lightly edited transcript.
HLM: Why did you do this study?
Roberts: We wanted to take a critical look at the implications of this earlier iteration of pay for performance for physicians, quality of care and spending which were the key outcomes that CMS used to judge physicians, and implications for how practices were paid and whether payment adjustments in the Value-based Payment Modifier really reflected differences in the value of care delivered by different practices or unmeasured differences in the characteristics of patients served.
This has big implications for whether physicians who are treating sicker and poorer patients are going to be treated fairly in payment models like the Value-based Payment Modifier and MIPS, and whether they will have the appropriate resources to address the needs of disadvantaged patients as Medicare continues to implement value-based purchasing initiatives.
It’s very hard to distinguish what is variation in provider quality, who are good and bad providers, from differences in the patients that they serve. Value-based Payment Modifiers and MIPS are one example where the case mix adjustment is particularly inadequate and may contribute over time to healthcare disparities.
HLM: What did you conclude?
Roberts: Our article very clearly crystalizes this argument that, with value-based purchasing models, policymakers need to think carefully about incentives and unintended outcomes to ensure that the goals of the policy are being achieved while the policy itself is not being undercut by the fact that providers serving sicker poorer patients are more likely to be penalized provided that risk adjustment is inadequate.
Risk adjustment is usually inadequate in these programs, in part, because it is difficult to measure the differences in complexity of patients across providers. We need to take a careful look at how incentives in these programs are structured and how performance is assessed in order to create the right incentives to improve value and outcomes for the most vulnerable patients.
HLM: How does this imbalance in risk adjustment manifest itself?
Roberts: If you look in the measures specifications that CMS puts out to evaluate provider performance under the Value-based Payment Modifier, there is a glaring inconsistency across the patient characteristics used to risk adjust outcomes.
For example, one of the key performance measures to assess practice quality in the value modifier is hospital admissions for ambulatory care sensitive conditions. CMS only adjusts that measure for the age and sex of beneficiaries, whereas it is quite conceivable that disabled patients, poorer patients, patients with long-term clinical comorbidities are at higher risk of admission, but those are not taken into account.
Other measures in the program are just adjusted for different sets of patient characteristics. This inconsistency in how risk adjustment is applied across performance measures, and the parsimony of the adjustment which doesn’t account for many relevant patient factors, is concerning.
HLM: How do you fix it?
Roberts: We would suggest that adjusting for observable clinical and socioeconomic characteristics of patients is advisable to mitigate any unintended consequences of this program for providers who serve more of these patients.
It is impossible from a practical standpoint to capture all differences among patients that may be relevant predictors of healthcare use and spending that we would want to distinguish provider performance from differences in patient mix. No amount of risk adjustment is going to completely mitigate the problem we have identified in the value modifier, where practices that serve sicker poorer patients in ways that we don’t measure well are going to get penalized.
In addition to better risk adjustment for clinical and social factors that we can observe, policymakers may want to consider primary care case management fees akin to what have been used in the comprehensive primary care model that CMS has, which would be risk adjusted so that providers who serve sicker poorer patients would receive a higher per member per month case management fee to help address some of those complex needs of patients that are not being adequately accounted for.
HLM: Are value-based payments fundamentally flawed because they do not adequately reflect patient mix?
Roberts: I would agree that it is a fundamental flaw. The principle which is that big payers like Medicare should be pushing the system toward a value-based system and not a fee-for-service system is well intentioned, but the execution is wanting. We don’t think many of those limitations are corrected in the MIPS.
Any efforts to promote value-based payment needs to be accompanied with other correctives to ensure that providers are correctly incentivized to improve value and not disadvantaged for treating poorer and sicker patients.
HLM: What other corrective actions do you recommend?
Roberts: Promoting competition among healthcare providers. There is substantial literature that suggests that provider consolidation is not associated with better patient outcomes, may be associated with worse outcomes, and certainly drives up healthcare costs, the opposite of the goal that policy makers are intending to achieve with value-based purchasing. So there is a certain amount of market-based activity that is needed to achieve the goals of value-based payments.
HLM: What other problems do you see with MIPS?
Roberts: MIPS retains the budget neutrality provision but includes more bonuses and penalties. More provider groups will be eligible for bonuses and penalties because they use a more continuous scoring system as opposed to a system where practices have to be above or below discreet thresholds to get payment adjustments.
There is more potential for adjustments that may enhance incentives for quality improvement to some extent, but practices subjected to MIPS are given greater leeway to choose their performance measures so they can pick the test questions they are scored on. Ultimately, this opens the door for gaming the system and doesn’t necessarily produced a balanced set of measures across which all practices are assessed for quality.
That may weaken incentives or focus providers on the wrong quality measures rather than on the ones that are most salient for patients. On balance we don’t think there is a net improvement in the design of the MIPS vis-a-vis, the value-based payment modifier.
John Commins is the news editor for HealthLeaders.