Finding ways to curtail growing costs and increase the value of health care is of perennial interest to policymakers and consumers alike. Ten years after the creation of the Center for Medicare and Medicaid Innovation (CMMI), a recent commentary from Zirui Song, M.D., Ph.D., in the journal HSR offers a broad review of available evidence on Accountable Care Organizations (ACOs), summarizes three new studies in the August issue, and offers perspective on future policy options.
The Affordable Care Act created CMMI to test payment and delivery models aimed at improving the value of care. CMMI did this largely through ACOs, which attempt to move health care providers away from systems that pay them solely by the quantity of care they offer (known as fee for service) to systems that offer prospective or lump sum payments based on different incentives.
Dr. Song’s commentary notes that evidence of overall impact is complicated, and mixed. Some ACOs have produced net saving over time, but others have not. The focus on rewards for quality appears to have encouraged providers to try to improve performance on quality measures, and the focus on shared savings and risk may have contributed to the recent slowing of aggregate health care spending. Yet savings attributable to alternative payment models (APMs) overall are small. Providers who make less in an APM than they would have under a fee-for-service model tend to exit those APMs at higher rates. Additionally, quality improvement thus far has been largely confined to care that can be measured by traditional primary care metrics.
Song states that, “moving the massive and fragmented delivery system toward higher value is a marathon, in which the last decade marked a few steps forward.” Through that lens, he posits that the picture facing policymakers isn’t whether payment reform efforts have been a success or a failure overall, but rather that reform efforts have illustrated some savings and quality improvements that offer both promising policy directions and important tradeoffs.
Improved contract design and the streamlining of incentives are two areas of promise. Song also makes the case for strong clinical leadership to drive change towards higher value clinical decisions, which are the fundamental building blocks of improving the value of societal health care spending.
A Closer Look at Payment Models
Voluntary APMs - where providers choose whether to participate or not – complicate the ways ACOs use incentives to drive change.
- Incentives need to be strong enough to motivate providers to change behaviors in ways that save money, but generous enough to encourage them to participate in the first place.
- The spending benchmarks ACOs set to measure progress are also critically important. A higher, or less aggressive, benchmark may be easier to achieve and therefore encourage participation, but with the tradeoff of weaker incentives to save. Lower benchmarks increase the incentive to save money, but require more work on the part of providers to change practice patterns and may drive providers out of participation.
- Quality bonuses encourage provider participation and help ensure delivery of certain services, but offsets potential savings for the health care system.
- Modestly generous benchmarks based on a provider’s own historical spending, which are not adjusted as performance changes over time, could help voluntary APMs strike the right balance of incentives. Adjusting the amount of shared savings and risk can refine this balance, and tying them to quality or reductions in wasteful care can further improve overall value.
- Improving risk adjustment – that is considering the underlying health and other factors that predict a patient’s health care needs – and creating risk corridors that limit losses and gains within an allowable range may also encourage provider participation in voluntary programs.
Mandatory payment models face legal and administrative hurdles, and mandatory participation may not in itself be enough to ensure savings or quality improvement. Instead, policy makers could make the status quo options less appealing, for instance by lowering fee-for-service rates outside the ACO model.
Addressing Price and Market Power
Song notes that in both voluntary and mandatory models, the underlying prices of services are still an important factor, as distortions in prices are baked into commercial ACO benchmarks. Addressing prices and provider market power, whether through competition or policy, remain central to slowing the growth of health care spending.
Streamlining incentives across payers and the patient populations across health plans helps clarify the potential financial impact for providers in ACOs, and evidence suggests that having a critical mass of patients covered under APM incentives helps encourage provider participation. Similarly, having streamlined quality measures across payers helps align clinical goals and allows providers to focus on care redesign in situations that apply across insurer populations. From the insurer perspective, benefit designs that align patient incentives with provider incentives also help encourage higher value choices.
Additional evidence is needed to understand how APMs work among commercial insurers, given that much of the evidence today is driven by CMMI models. Further, APMs in Medicaid are relatively new and present unique questions given the increased financial risk among providers who serve populations who have both higher health risks on average and significant inequities in access to care.
Policy instruments like alternative payment models can incentivize change, but probably only to a point. Song emphasizes that cultural change in clinical practice is likely necessary. Volume incentives like fee-for-service are well entrenched in health care, and highly priced services are often fiercely protected in the federal process of valuing clinical services. Prices tend to increase after consolidation – an growing trend in health care – and lobbying against efforts to constrain prices remains robust. Nevertheless, high quality and high value care aligns closely with the professional values of clinicians. Clinical leaders who can communicate and demonstrates those values are necessary to drive change and inspire the future workforce.
Additional ACO Evidence in this Month's HSR
Three new studies in the August issue of HSR help illustrate the complexity of assessing the potential of APMs Song describes.
The first, by Acevedo et al. looked at a large Medicare ACO model known as the Medicare Shared Savings Program (MSSP) and found that patients in the ACO received fewer mental health and substance use services and less adequate care for depression, with an apparent widening of racial and ethnic disparities. Most prior studies evaluating MSSP ACOs have not examined their impact on mental health or substance use services, or in racial/ethnic disparities in these services.
The second, by Zhang et al. looked closely at savings generated from encouraging provider prescribing of generic drugs in a commercial ACO that shared risk and financial savings equally among a participating hospital system, an independent physician group and a commercial managed care organization. The authors concluded the model was insufficient to produce a tangible benefit on its own. This is an important finding given Medicare ACO contracts typically carve out prescription drug spending from shared savings and risk. The authors suggest that both increased provider financial risk-sharing and lower brand-name drug prices are needed to produce greater value.
A third study, by Coe et al., looked more closely at a value-based or risk-sharing accountable care network in Washington State to understand differences in how patients interacted with the health care system. In the first year, patients were less likely to use outpatient hospital visits across all age-groups, and young children experienced an increase in primary care and specialty care visits. However, there was little impact on health care intensity or quality metrics during the same period. The researchers note that results could be limited by the short time frame, and additional evaluation over a longer term would be useful.