Understand types of financial risk in VBP models

In Oregon and nationally, payers and providers are increasingly engaging in risk-based payment models. Risk-based contracts are proliferating in the commercial insurance market, and Congress and CMS have created strong financial incentives and requirements for Medicare providers to move to shared risk payment arrangements.

Risk–based VBP arrangements offer providers the potential for greater financial rewards and more flexibility and the opportunity to invest in clinical care transformation. Risk-based VBP models include one-sided arrangements, offering you rewards or “shared-savings” resulting from more efficient provision of quality care, or two-sided models, with a requirement that you assume some financial or “downside risk,” if medical expenditures are higher than expected within the VBP model for the attributed population and identified services.

In shared savings or upside-only VBP models, providers may earn a portion of savings if their costs are below a defined benchmark. In downside-risk models, providers are responsible for paying a portion of the costs above the benchmark if the benchmark is exceeded. Payment models with downside risk typically provide stronger incentives and more flexibility for care delivery transformation than upside-only models.

While provider entities and clinicians are increasingly adopting VBP, many are reluctant to participate in VBP models with downside risk (shared risk) even though these models may provide stronger incentives or flexibility for care delivery transformation. In upside-only (shared savings) VBP models, providers may earn a bonus and/or a portion of savings if their costs are below a predetermined benchmark. In downside-risk models, providers are responsible for paying a portion of the costs above the benchmark if the benchmark is exceeded. These benchmarks are sometimes referred to as Total Cost of Care (TCOC) targets, even though some services and populations may be excluded from the TCOC calculations.

Most VBPs are built upon the FFS architecture where payers continue to pay claims as they always have. Some VBP models then reconcile claims payments to TCOC budgets for shared savings or shared risk arrangements at the conclusion of a performance period. Determining how close you are to the TCOC targets and waiting to receive any shared savings can be challenging. In some VBP models, you may not have access to up-front dollars (such as supplemental or foundational payments) to cover expenses for care or services not traditionally reimbursed in FFS. In addition, if a payer reconciles the claims payment to the VBP target long after the conclusion of a performance period, it dilutes the motivational impact of the incentive.

Consider natural fluctuations in health care service use and associated medical expenditures in any given population occur randomly and should be considered. VBP models that do not appropriately account for random variation within and across small numbers of patients, could result in a payer not rewarding a provider entity when “true” savings occurred, and/or penalizing a provider entity when “true” losses did not occur.

While smaller provider entities and/or those with a small volume of patients from a given payer, may face unique challenges in advanced VBP models, regardless of size, your providers may have some trepidation about entering into a risk-based payment model.

The good news – is that there are options to reduce and manage financial risk as you enter advanced VBP arrangements. Several options are outlined below. These risk-based model options are not mutually exclusive and can help you assess, identify, and in some cases modify, VBP models that work best for your situation and size.

Options for payers and providers to manage financial risks in VBP arrangements:
Below are some strategies to support success for providers who are advancing to models that include sharing in losses, known as downside risk.

Set a minimum number of attributed lives for shared risk models
Both provider entities and payers may want to set a required minimum number of attributed lives before they will enter into a risk-bearing contract. The minimum number of attributed lives will not necessarily be the same across models, but reflect the populations and services included in the VBP model.

Reduce variation in data and financial risk
Reducing variance to address the risk of inaccurate shared savings or recoupments is important. With health care providers increasingly being rewarded based on changes in cost of care, it is critical that sufficient statistical safeguards are in place to ensure that payment arrangements fairly reflect provider performance rather than random variation in medical utilization.(1) Strategies to reduce variance include:

  • Truncating extreme or outlier values by setting a maximum annual aggregate claim level (such as $100,000) and excluding every claim above that level from consideration in the VBP model. This could also be done by truncating claims at a percentile (such as 99th) of all population expenditures.
  • Excluding all dollars associated with predefined exceptional circumstances and catastrophic cases (such as patients receiving organ transplants.).
  • Using the same patients to measure outcomes in baseline and performance periods, In this prospective attribution approach provider entities are at-risk during the performance year for patients they had previously treated in the baseline period.
  • Covering all patients served by the provider entity in the performance year but measuring effects among higher-risk patients. Restricting performance measurement or risk-based contracting to a smaller, more costly group of patients will reduce the variance and increase the likelihood of detecting a true effect. This type of prospective attribution approach typically involves the use of a predictive risk-score algorithm such as Adjusted Clinical Groups (ACGs) or Chronic Illness and Disability Payment System (CDPS) to identify those at highest risk of future expenditures.

VBP Arrangements that include a Risk Adjuster
Even within VBP arrangements that do not involve prospective attribution, providers and payers can utilize risk-adjustment software to understand and modify expected spending levels by age and gender to reflect the patient population served. Payments can also be adjusted for clinical risk when part of a total cost of care arrangement. By incorporating information on overall risks of attributed populations in VBP calculations, payers can reduce the likelihood of financially penalizing providers for serving patients with a higher-than-average burden of illness or reward providers for serving a healthier than average patient population. Case-mix can vary substantially across providers. In addition, risk-adjustors help account for changes in attributed patients’ health status over time that may increase expenditure variance in an attributed panel. Primary care only payments should not be adjusted for clinical risk because a commonly accepted methodology to estimate how much primary care someone needs based on their medical condition(s) does not yet exist.

Set a minimum savings/loss percentage within VBP models
Small gains and losses are most subject to random variation in utilization and medical expenditures. The risk of a payer inappropriately rewarding or penalizing a provider entity can be minimized by setting a minimum savings/loss percentage, such as 1-2 percent of the total cost of care target, in a VBP arrangement before any savings/losses are shared with the accountable provider entity.

Discount small gains or losses within VBP models
Discounting small gains or losses, and not enabling (or requiring) that an accountable provider share in those savings (or losses) to the full extent, also helps mitigate the risk of false positives and false negatives in VBP arrangements. For example, a payer using this approach might discount savings (or losses) between 2-5 percent before calculating savings the provider has earned or the portion of losses a provider must offset.

Aggregate providers for measurement of financial performance
Grouping smaller panels/provider entities together for purpose of financial performance measurement increases the sample size available to assess provider performance. Aggregating with other providers may help reduce financial risk and uncertainty in an advanced VBP arrangement. However, your provider entity still needs to understand and agree to the methodology for sharing savings or financial penalties across involved entities.

(1) McCall N., Peikes D. (2016). Tricky Problems with Small Numbers: Methodological Challenges and Possible Solutions for Measuring PCMH and ACO Performance, State Health and Value Strategies, a program of the Robert Wood Johnson Foundation. https://www.shvs.org/resource/tricky-problems-with-small-numbers-methodological-challenges-and-possible-solutions-for-measuring-pcmh-and-aco-performance/