Monday, April 22, 2019

The Implications Of Recalibrating Medicare Advantage Risk Adjustment Using Encounter Data


The Medicare Advantage (MA) program allows Medicare beneficiaries to receive their Part A and B benefits through private plans that are paid a fixed amount per enrollee (or capitated) by Medicare to provide all such benefits to their enrolled members. MA plans play an increasingly important role in the Medicare program, with enrollment growing from about a quarter of all Medicare beneficiaries (or 11 million) in 2010 to 34 percent of beneficiaries (approximately 20 million) in 2018. And, according to the Congressional Budget Office, enrollment in MA is projected to exceed 40 percent (32 million) of the total Medicare population by 2028.
MA’s popularity indicates that it works well, and one of the reasons the system functions as well as it does is that plan payments are adjusted for the health risks of enrollees. The approach to risk adjustment in MA has evolved over time and will likely continue to evolve. Until now, the Centers for Medicare and Medicaid Services (CMS) has estimated its MA risk-adjustment models using data from traditional Medicare, but the agency has signaled in the past its interest in using encounter data from MA to estimate (or recalibrate) the model. CMS reiterated this interest in its fiscal year 2020 performance budget.
Under such recalibration, CMS would recalculate the coefficients for the demographic characteristics and health conditions based on MA encounter data, instead of on traditional fee-for-service Medicare claims. Encounter data, unlike fee-for-service claims data, are not necessarily the basis for payment to providers but instead capture services received by MA beneficiaries, regardless of how the MA plan paid for those services. Such a move would likely change the risk scores of certain populations in MA to more closely reflect MA costs and treatment patterns, although the Medicare Payment Advisory Commission (MedPAC) has questioned whether encounter data are yet of sufficient quality to rely on them for payment purposes, including risk adjustment. Likewise Sean Creighton and colleagues, writing recently in the Health Affairs blog, also explored the challenges that persist with encounter data.
CMS has released neither a specific proposal for this recalibration nor a timeline, but development of this new model would require the agency to address many policy, technical, and operational issues prior to implementation. The Urban Institute and the American Action Forum convened a one-day roundtable summit with experts on MA, risk adjustment, and related areas to discuss the implications of calibrating the model using encounter data. Attendees included 21 experts from a variety of sectors, including the insurance and pharmaceutical industries, academia, consultants, actuaries, and former CMS officials. Here, we highlight some of the key points that emerged during the summit. The full participant list and further findings are provided in an issue brief and panel discussion video
Risk Adjustment In MA Needs To Balance Payment And Policy Goals
Risk adjustment in MA balances two governmental goals: paying plans accurately for the health status of their enrollees and creating appropriate incentives for plans to compete by offering value instead of avoiding risks. Payment accuracy (that is, the degree of correspondence between payments and costs) would almost certainly improve by recalibrating the MA risk model with encounter data because the risk model would be fit to the patterns of spending and use for particular diagnosis groups within MA, instead of the cost and use for those diagnosis groups in traditional Medicare. Prior research has shown lower use and spending in MA than in the traditional fee-for-service side of Medicare, perhaps reflecting MA network design, referral requirements, or use management.
However, improved payment accuracy due to recalibration with encounter data could have implications for various financial incentives faced by MA plans. For example, if MA plans are, on average, more efficient at treating diabetes than is the traditional Medicare program, the calculated relative risk of—and thus the payments for—beneficiaries with diabetes would likely be lower in a risk-adjustment model calibrated with MA encounter data than in the current model. Lowering the risk score for diabetes could affect MA plans’ incentives to coordinate care and control costs for those patients. On the other hand, if plans are more efficient than fee-for-service providers because of care coordination programs designed for enrollees with diabetes—the costs of which are not reflected in encounter data—payments for these enrollees would be artificially low, and payment accuracy would be undermined.
If payments for enrollees with particular health conditions (such as diabetes) became too low because of recalibration, plans could select against such enrollees or shift costs onto beneficiaries. Under significant payment pressure, both selection and cost control could be accomplished by relatively straightforward benefit design changes, such as narrowing provider networks for specialists or increasing cost sharing for insulin. It is unclear, however, how feasible condition-specific risk selection would be. A significant reduction in relative payment for diabetes could also meaningfully alter the market for chronic condition special needs plans, which are targeted MA plans that limit enrollment to Medicare-eligible beneficiaries who have particular chronic conditions, such as diabetes. This could potentially result in plan exits from that market.
Recalibrating the risk model would have other implications, too. It would eliminate the need for the coding intensity adjustment currently applied to MA risk scores and could induce the need for changes to other aspects of the MA system that are also based on traditional Medicare costs, such as the benchmarking system for plans submitting bids, as described in the brief
Issues With Assigning Costs To MA Encounters And Completeness Of Encounter Data Must Also Be Addressed
Encounter data do not necessarily provide comparable, accurate costs across the MA system, particularly for vertically integrated plans or plans with subcapitated arrangements with provider groups. Because MA plans largely base provider reimbursements on traditional Medicare prices, MA costs could be estimated by applying traditional Medicare prices to MA use data, instead of relying on plan-submitted costs. While this approach could fix potential problems with encounter data prices, it would not address the likely underreporting of use data resulting from subcapitated payment arrangements between providers and plans.
In addition, because MA still makes up only about a third of the Medicare population, it is unclear if encounter data will provide sufficient sample sizes to estimate models for all the diagnosis categories that are currently part of the risk-adjustment model, potentially requiring CMS to combine some categories. Also, the total cost of care for MA beneficiaries may be underestimated if it is calculated only from the encounter data. Plans do not report encounter data for all supplemental benefits, therefore underestimating the costs of serving the MA enrollee population. MA encounter data may also not fully capture postacute care received from skilled nursing facilities and home health agencies because diagnoses from postacute care are not used in risk adjustment. Postacute care services have been shown to be underreported in encounter data as recently as 2015. 
Research Should Inform A Transparent Process For Changing MA Risk Adjustment
Before making major changes, CMS can use MA encounter data it already has on hand to conduct and disseminate research to inform eventual recalibration. Such research could promote a deliberative, public process. For example, analyses exploring how MA plans provide care, how the provision of care in MA differs from traditional Medicare, and how outcomes differ between MA and traditional Medicare could help CMS and stakeholders more fully understand the implications of recalibrating the risk-adjustment model using encounter data. Such research would support policy making on risk adjustment and the broader MA and traditional Medicare payment systems.
As CMS considers moving to a risk-adjustment model calibrated with MA encounter data, conversations between CMS and stakeholders will help create buy-in and address concerns about potential negative effects on the MA system. If CMS plans to move forward with recalibrating the risk-adjustment model with encounter data, it should communicate the policy goals it expects to achieve and how progress toward those goals will be measured. It should also establish a transparent process for developing the revised system with a timeline that allows for stakeholder input along the way.
Overall, the MA experts at the risk-adjustment summit agreed that encounter data are not a cure-all for all potential problems with the MA risk-adjustment system. Experts were united in calling for a careful, deliberative, and transparent process as CMS works to address the policy, technical, and operational issues raised by recalibrating the risk-adjustment model with encounter data.
Support for this project was provided by Anthem Inc. and Eli Lilly and Company. For more information on the Urban Institute’s funding principles, click here.

No comments:

Post a Comment