Practice transformation and payment reform are defining features of contemporary health policy debates. The story goes like this: new provider organizations, such as Accountable Care Organizations (ACOs) are transforming care delivery from fragmented and volume driven to integrated and optimized for quality; meanwhile, innovative payment models, such as bundled payments and risk-based contracting, herald a national transition from fee-for-service (FFS) to value-based payments. But a new article in Milbank Quarterly from LDI Senior Fellows Lawton R. Burns and Mark V. Pauly poses an uncomfortable question: what if transformation is simply hype?

Burns and Pauly challenge the prevailing wisdom that payment and practice transitions are happening quickly, shifting financial risk to providers, leading to higher quality, lower cost care, and truly inevitable. They provide historical context and a sweeping overview of studies to date to conclude that “the transformation from ‘volume to value’ appears to be driven more by ideology and aspiration than by evidence.”

How Widespread Are Alternative Payment and Practice Models?

How prevalent are these new payment and practice models, anyway? Burns and Pauly note that the data seem to conflict: one study suggests that commercial payments based on alternative payment models (APMs) rose from 11% to 40% between 2013 and 2014. Another study from 2015 suggests that 62% of insurer payments are based on FFS, and only 23% are based on APMs. A 2016 physician survey found that 77% of physicians said only a fifth or less of their compensation is linked to an APM. How can a large percentage of payments be tied to quality, but a small amount of compensation based on value? The way providers are paid provides some insight.

The transformation from ‘volume to value’ appears to be driven more by ideology and aspiration than by evidence.

Physicians and hospitals are reimbursed by multiple insurers across several payment models. Thus, while a large percentage of providers may receive some payment linked to value, those payments represent a small share of total compensation, and thus have only a limited impact on care delivery. Secondly, risk-based contracts with physician groups and hospitals rarely trickle down to individual physicians. Thus, many physicians have little to no exposure to financial risk, even if their organization is in such a contract. Finally, even in cases where physicians have an APM contract, FFS remains the dominant form of revenue. A similar trend is observed among hospitals: a Modern Healthcare survey showed that two thirds of hospitals derived less than 1% of total revenue from APMs.

Similar definitional challenges arrive when classifying physician organizations. For example, a primary care physician can simultaneously be a lone practitioner, hospital employee, and attributed to a Medicare ACO. Furthermore, there are no longitudinal data on many forms of practice organization, so comprehensive trend analysis is impossible. However, limited evidence suggests that most physicians remain in small or solo practices, and the transition to employment by integrated delivery systems is progressing very slowly.

Do Alternative Payment and Practice Models Reduce Cost or Improve Quality?

OK, so the pace of transformation has been overblown. But when it happens, does it reduce cost or improve quality? Burns & Pauly review the evidence and find that most have serious methodological drawbacks, including a lack of randomization and voluntary inclusion for most risk-based contracts. A meta-analysis of early CMS Pay for Performance (P4P) programs found no positive impact on patient outcomes, but improved effects on process measures. Bundled payments have yielded mixed results as well: modest savings for certain care episodes, mostly related to reduced prices for input costs and use of post-acute care. In general, the evidence for value-based payment arrangements is still mixed, and in many cases, nonexistent. Researchers have difficulty capturing all of the necessary clinical and financial data, and the results of many studies are non-significant.

The literature on the impact of new models of care delivery is far more developed. Most delivery system innovations within the CMS Innovation Center (CMMI) have had little to no impact on total cost of care, though there is a wide variation, and proponents have seized on the few hopeful results. One troubling finding is that horizontal integration of both physician practices and hospitals are associated with higher prices, without any improvement on quality, spending, or patient experience.

The promise of ACOs remains hotly contested. Most data come from the Pioneer ACO program, which ended in 2016, and the Medicare Shared Savings Program (MSSP) ACOs. Roughly half of Pioneer contracts reduced spending, but savings were not correlated with improved quality. Physician-led Pioneer ACOs had the highest savings rates, but only 12 of 32 participants remained by the end of the program. Among hospitals, the relatively small size of the ACO population diluted incentives to pursue strategies that would have had spillover effects. Savings from Medicare Shared Savings Program (MSSP) ACOs have been lethargic, with no correlation between savings and quality improvement. Physician-led ACOs have reaped the most shared savings, while many managed care veterans dropped out of the program. Although spending has decreased for about half of ACOs, the program cost CMS $216 million after paying out bonuses. MedPAC commissioners have called the program unsatisfying, and only a small percentage of ACOs are on tracks with two-sided risk contracts.

Lessons from the Past

Burns and Pauly see the ghosts of transformations past in the rhetoric of the present. In the 1990s, payment models built on capitation and cost effectiveness reduced cost without harming clinical measures of quality (access suffered, however). The same could not be said for payer-centric integrated delivery networks, physician-hospital organizations, and hospital alliances which failed to impact either cost or quality. The value proposition of the 1990s is remarkable similar to the 2010s. The apparent repeat of history could be because managed care was too ahead of its time: electronic medical records, data analytics, and improved value metrics have made the implementation of narrow networks and risk-based contracting more economically viable and politically palatable. On the other hand, a new generation of managers may simply be repurposing old ideas.

The backlash against managed care holds some lessons for reformers. Lower payments to providers and denial of low-value services to patients lowered costs, as did narrow networks and HMOs, without reducing quality. But physician and patient protests undid the HMO experiment. Efforts to develop care across a larger health care continuum were confounded by the high cost of developing large care networks and low revenues from services outside of hospitals.

Will the “Iron Triangle” Overtake the “Triple Aim?”

Crucially, reform in the 1990s and early 2000s was conceptually rooted in the “Iron Triangle,” which argued that increased access, higher quality, and lower cost care are inherently in tension with each other. HMOs restricted access to reduce cost, and the backlash against them increased provider network size at the expense of higher prices. In contrast, contemporary health reform is built on the “Triple Aim:” better care experience, better population health, and lower cost. However, if the iron triangle is an accurate framework, then cost and quality cannot be mutually optimized, and the potential for achieving the Triple Aim must be called into question.

Research on the relationship between cost and quality suggests little to no correlation in either direction. On the one hand, this may indicate efforts to reduce cost are wholly independent from efforts to increase quality. Alternatively, while there may be a cost/quality tradeoff, highly efficient firms manage to optimize both, and vice versa. Or, it may also be the case that the relationship between cost and quality is contingent on specific interventions. For some services (overused, unnecessary care), higher cost could accompany lower quality and result in a negative correlation; for other services (under-used appropriate care), higher cost could accompany higher quality care and yield a positive correlation. The key takeaway is that, no matter the causal logic, it is unlikely that a single payment or practice model can simultaneously improve all quality metrics while reducing global cost.

The Future of Reform: A Slow Burn?

In lieu of top-down, revolutionary payment reform, which is unlikely to materialize, Burns and Pauly see a future in which alternative payment models stall at bundles for surgical procedures and FFS plus penalties and bonuses for quality. But reforms to reduce cost are by no means dead. Revisiting the efforts of HMOs in the 1990s may spur short term, but effective ways to reduce costs—such as reducing unnecessary admissions and shortening lengths of stays in hospitals.

A slow-going approach does not require resignation to ever-rising costs. Policymakers can keep an open door for insurers to develop new ways to allocate resources to physicians and allow for greater consumer choice in the insurance market. Burns and Pauly suggest that the current dogma of transformation tied to payment and practice reforms could be replaced with more effective paradigms. For example, focusing on physician behavior change and care for the chronically ill, rather than a wholesale replacement of the FFS behemoth, can improve cost and quality. The evidence suggests reforms to the physician-patient and physician-hospital relationship may be a more effective framework.

Burns and Pauly are unafraid to be seen as flies in the ointment, and they describe their work as not a “negative screed but instead a critical evaluation.”  Rather than fall into group think, policymakers might heed the call to “curb their enthusiasm” when it comes to the latest version of health system transformation.