Risky business: Achieving value-based purchasing goals in New York State

April 14, 2017 / By Michael Keyes

I’ve spent much of the last few months traversing the State of New York speaking to our Medicaid MCOs about their strategies to reach the state’s value-based purchasing (VBP) goals. It’s been well publicized in NYS publications and our previous blogs that plans are required to have 80 percent of their payments in value-based contracts by 2020. By our estimates, most plans will need to contract every group with greater than 500 members in Total Cost of Care contracts to achieve that goal. It would be a gross understatement to call this a significant lift. But added to that is a requirement to move some of those VBP contracts to shared risk. More specifically, plans need to have 30 percent of payments in Level 2 arrangements by 2020.

Long ago, I cut my teeth in the industry with some early risk-based arrangements. Over ten years ago, many providers experienced success in shared risk arrangements for their assigned Medicare Advantage population. Medicaid however, was a different story. One or two large provider groups gave it a try but were not successful. Sure, we know a lot more about VBP today than we did back then but the Medicaid population is harder to manage. Lower premiums in a more transient population make shared risk arrangement riskier than most providers are willing to accept.

Our most recent conversations with health plans have revolved around how to design shared risk programs that measure providers on what they can control while protecting them from what they cannot. To be clear, it’s my opinion that plans will need to engage in Total Care for General Population (TCGP) arrangements to reach New York State’s requirements. Anything other than that amplifies the amount of contracting leg work, creates even smaller cell sizes to risk adjust and is likely unsustainable. Having said that, I understand why providers might be unwilling to take on even partial risk for an entire population.

Just this week, I spent some time with a colleague and partner MCO brainstorming program design elements that treat both the provider and plan fairly in a shared risk VBP arrangement.

Creating risk-adjusted targets for total cost of care and utilization based quality metrics will be critical. Having sat on both sides of the table, I just can’t see a contracting entity accepting downside risk in the absence of risk adjustment. Beyond that, the mechanics of risk adjustment are critical.  It’s my opinion that a concurrent approach that monitors and adjusts for the risk of a population each month is the best option. This protects the provider from high-cost acute conditions outside of the control of a PCP – think appendicitis or hip fracture.  At the same time, a concurrent approach incentivizes the well-visits and the engagement of non-users.

The basis of measurement is also an important consideration.  In other words, how is a target set? One option is to compare providers to their historical performance. This is generally the preferred approach for providers whose population, when risk-adjusted, spends more than expected. On the other hand, a lower-cost provider may prefer to be measured against a network average or external benchmark.

Determining which members will be included in the measurement will become important. Providers have expressed concern about assuming risk for the very ill population whose costs are known to be high.  While sound risk adjustment and stop-loss policies should account for this, a plan may opt to exclude individuals with high-cost cancers or other catastrophic conditions.  Incidentally, 3M’s Clinical Risk Groups (CRG) methodology works quite well at categorizing clinically similar individuals and can be used for inclusion/exclusion criteria. A plan must be conservative in the exclusion logic so as not to come up short in meeting the state’s VBP requirement.

Stop-loss is a trigger on a member or claim after which the provider is no longer at risk for the dollars. Here a plan and provider should work together to determine a stop-loss policy that works for both parties. 3M recommends applying stop-loss at the person level.

Regardless of the approach, health plans will need to prepare for an open and transparent conversation when contracting for shared risk VBP and throughout a performance year since providers often lack the necessary longitudinal data to model contracting scenarios and identify the opportunities to improve the quality and cost of care their patients receive. 

Michael Keyes is an engagement leader for Populations and Payment Solutions at 3M Health Information Systems.