Refining the framework for payment reform


There is broad consensus that physician and other provider payment methods need to evolve from rewarding “volume” to promoting “value.”  But what that means exactly, and what will be the recipe for success, is yet to be determined.  It’s a complicated and messy process, as no method is ever implemented in isolation – either one at a time or in a vacuum – and many payment methods do not yet have a sufficient track record to help us prioritize among them.  Therefore, we need to tread carefully and experiment while amassing evidence so that one day we can promote payment methods based on ample road testing.

The Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) consolidated an array of pay-for-performance mechanisms into a single Merit-Based Incentive Payment System (MIPS) and created strong incentives for providers to participate in alternative payment models (APMs).  In concordance with this change, the Department of Health and Human Services (HHS) announced the intent to move Medicare payment methods along a continuum of added value.

The Health Care Payment Learning and Action Network

To support its own work in developing and testing promising payment methods, the Centers for Medicare and Medicaid Services (CMS) established the Health Care Payment Learning and Action Network (LAN). One of the LAN’s initial objectives was to elaborate a framework for classifying current and potential payment methods, relying on an HHS categorization, with the direct purpose of tracking their implementation. The Framework classifies 28 different approaches into the four major categories HHS established, migrating from fee-for-service as the least value-based to population-based payment as the most.  This continuum represents an increase in accountability for both quality and cost and a greater focus on payment for population health management.

The LAN Framework provides a useful kick-off in establishing a common language and logical categorization to discuss and monitor the progress of payment reform. In general, payment methods that explicitly measure performance and involve provider risk-bearing offer greater promise of producing higher value, all else being equal.

But rarely in today’s health policy landscape is all else equal. We believe it is important to look beyond the Framework’s straightforward continuum (hyperlink to be provided once the issue brief is ready) to account for the reality that every payment method has strengths and weaknessesThese often vary with the specific, operational designs adopted and the organizational culture and other circumstances in which the payment method is implemented. Indeed, the strength of an organization’s management structure, its culture and the effectiveness with which incentives are transmitted from the organization to the individual clinician help determine the success or failure of any payment method. Because of this reality, a fixed classification along a continuum can mislead policy makers about the relative merits that proposed APMs will have when applied in real circumstances.

A More Complex Landscape

To a large extent, the LAN classification continuum relies on the power of the incentives embedded in each payment method to improve quality and to reduce spending in a measurable way. Little considered, however, are the operational challenges associated with implementing the payment methods, particularly with those found at the most value-based end of the LAN continuum. For example, condition-based payment as a prominent form of population-based payment has strong inherent appeal. Yet, the vagaries of current diagnosis coding, as evidenced by large regional coding variations, raise practical concerns about the reliability of diagnosis/condition coding, a problem that would be magnified if the diagnosis code becomes the basis for a stream of episode payments.

Similarly, full-fledged population-based payment, such as capitation, requires workable and effective risk adjustment for health status, a challenge that still has not been fully surmounted. In short, the elegance of incentives is not sufficient to justify placing payment methods in different tiers, given the major operational challenges. Their performance in action ultimately determines their worthiness for broad adoption.

Indeed, the highest value might be produced with hybrid payment models that combine the strengths of various methods while mitigating their weaknesses, rather than relying on the singular payment methods that mostly make up the LAN Framework’s continuum.  Even the LAN acknowledges that there are hybrids in use.  Yet, the effectiveness of individual payment reform approaches tends to be evaluated based on their isolated ability to produce better outcomes.  This ignores the inevitable interactions and the importance of examining whether they alone or in combination help to eliminate, or at least ameliorate, existing impediments to better care.  Those impediments should be the immediate target of payment reforms.  Improvement can be accomplished both through modifications to legacy payment methods and the combination of multiple alternative approaches (which may include those modified legacy systems).

ACA requirements mean that the CMS Office of the Actuary must determine that demonstrations reduced spending without reducing the quality of care, or improved the quality of patient care without increasing spending and would not produce any increase in net program spending.  Our view is that only by testing in a variety of settings, with design variations, can policy makers draw conclusions about which payment methods deserve broad adoption, although we acknowledge that evaluation challenges make definitive evidence elusive.