Introduction
Each month, the Centers for Medicare & Medicaid Services (CMS) makes more than $25 billion in payments to Medicare Advantage organizations (MAOs). For the vast majority of MAOs, aggregate payment amounts represent the sum of risk-adjusted monthly payments determined for each enrollee. This makes it critical for both CMS and MAOs that the risk-adjusted per enrollee per month amounts are calculated correctly. Hence, it is not surprising that CMS and MAOs devote significant resources to reviewing and auditing the data used to determine payment amounts.
In this blog, we will review results from one of the auditing programs conducted by CMS, in which they estimate program-wide overpayments and underpayments due to inaccuracies in the risk score calculation. Historically, because of the impact to government liability, overpayments have been the primary focus of public scrutiny. However, because underpayments have the potential to impact an MAO’s ability to cover optimized healthcare for their enrollees, they could play a significant role in program effectiveness. For that reason, this blog will focus on the history, drivers, and types of plans at more or less risk of underpayments.
Extent of Payment Inaccuracies
The Department of Health and Human Services (HHS) releases an annual Agency Financial Report (AFR) that provides fiscal and high-level performance results for each of its programs, including Medicare Advantage (MA). Included in the report are estimates of the accuracy of risk-adjusted payments made in the MA program. CMS provides these estimates as part of its Improper Payment Measurement (IPM) Program. In order to conduct these estimates, CMS identifies payment errors for a sample of MA enrollees and extrapolates the sample payment error to the full MA population subject to risk adjustment.
HHS breaks down estimated payment errors by overpayments and underpayments. According to the federal government estimates, there were $15.2 billion in overpayments and $8.0 billion in underpayments during payment year 2019, which is the most recent year of data available in the AFR. This amounts to approximately 5.5% and 2.9% of total program payments, respectively.
The figure below shows the trend in underpayments from payment year 2014 through 2019. During this time, underpayments increased approximately 70% from $4.7 billion to $8.0 billion. This growth is largely explained by the growth in enrollment of the MA program. However, there was an over 20% difference in the minimum (2.89% in 2015) and maximum (3.42%) underpayment rate — i.e., underpayments as a proportion of total payments — during this time period. This type of volatility can be problematic for MAOs, as it can introduce uncertainty into planning for beneficiary benefit offerings.
Part C Underpayment Estimates (in $ Billions) and Underpayment Rates: Payment Years 2014 – 2019
Source: Centers for Medicare & Medicaid Services. Error Rate Findings and Results.
Reasons for Underpayments
In general, there are three scenarios in which an underpayment occurs. In the first case, an underpayment occurs when a diagnosis is documented in medical record documentation but has not been submitted to the Encounter Data System (EDS) for risk adjustment (see prior blog for more information about how data is reported for risk adjustment, and the changes to the reporting being made in 2022, including the shift from using Risk Adjustment Processing System (RAPS)). In this case, the plan is liable for the costs of the beneficiaries’ conditions without the associated incremental payment and is consequently underpaid.
In the second case, diagnoses for higher manifestations within CMS-Hierarchical Condition Categories (HCCs) are supported in medical records, but only lower-level diagnoses have been submitted to EDS. This is an issue because for certain related HCCs; only the HCC associated with the highest severity for that group would influence the risk adjustment. For example, suppose a member was to have diagnoses in medical records that mapped into CMS-HCC 8 for Metastatic Cancer and Acute Leukemia, as well as one of the other cancer HCCs (i.e., Lung and Other Severe Cancers (HCC 9), Lymphoma and Other Cancers (10), Colorectal, Bladder and Other Cancers (11) and Breast, Prostate and Other Cancers or Tumors (12)). In this case, only the diagnoses within HCC 8 would impact risk adjustment since it is considered the most severe by the HCC model. Accordingly, the adjustment for HCC 8 is substantially higher than for the other cancer HCCs. In this case, because only the lower manifestation has been reported, the plan is liable for the full costs of the higher manifestation without the associated incremental payment and is consequently underpaid.
The third case would include diagnoses that are reported on an encounter and documented in the medical record documentation in one year, but unreported in a subsequent year. The payment model only includes HCCs representing chronic diseases that would be expected to persist year-over-year. As a consequence, the failure to recapture an HCC each year for a beneficiary’s condition is a likely additional source of underpayment. In the absence of a year-over-year analysis of the encounters and medical record documentation, this scenario is largely unobservable. Given the IPM does a within-year snapshot of diagnosis reporting and medical record documentation support, it would miss this category of underpayment. Therefore, it is likely the reported measure understates the actual underpayment rate MAOs experience.
A major driver in underreporting diagnoses on encounters is the fact that reimbursement of the provider by the plan is not generally dependent on the complete reporting of diagnoses. Consequently, achieving complete reporting relies on robust provider education and medical record review programs. MAO review programs face significant challenges identifying the medical record documentation, collecting it, and performing a review to identify unreported diagnoses. These challenges include logistical issues, provider abrasion, and the cost of human review of medical record documentation that are particularly burdensome to plans that are smaller, not-for-profit, or focused on underserved populations.
MAOs at Greater Risk for Underpayments
Based on the recent historical trends in underpayments, RaLytics estimates underpayments will reach nearly $12 billion in 2022, with an average underpayment rate of 3.1% or over three million dollars per MA plan. The percentage is higher, approximately 3.7%, when limiting the average underpayments to just MA plans that accept risk-adjusted payments.
Moreover, based on prior medical record reviews, it appears that all plans experience underpayments. This is not surprising given the incredible burden to review all diagnosis documentation facing MAOs. The underpayment rate has generally ranged from 1% to 10% across MA plans for any given year.
PPO plans, which make up approximately a third of the MA enrollment, tend to have underpayment rates that are 1-2 percentage points higher than HMOs, which account for about half of MA enrollment. Special Need Plans (SNPs) are also associated with higher levels of underpayment relative to non-SNPs, with Dual Eligible SNPs (D-SNPs) having the highest rates of underpayments.
There is also evidence that smaller plans are at risk for higher underpayment rates. Plans under the umbrella of the five parent MAOs with the most enrollment (these five account for over two-thirds of all MA enrollment) are associated with lower underpayment rates. These findings suggest that large MAOs are in a better position to conduct a robust risk adjustment data validation process that ensures more complete coding.
Conclusion
The data above underscore the importance of risk adjustment validation processes. Underpayments directly result in reduced revenues to MAOs, negatively impacting funding available to support necessary care management for their enrollees.
In response to the Improper Payment Elimination and Recovery Improvement Act (IPERIA), CMS runs two audit programs for improper payments. As discussed, the IPM measures a program-level estimate of the overpayments and underpayments in Medicare Advantage. However, the IPM does not measure contract-level improper payments or include any action to reconcile the improper payment with actual payments to MAOs. CMS’ contract-level Risk Adjustment Data Validation (RADV) audits are responsible for measuring a plan-level estimate of improper payment and recovering overpayments from the MAO based on those estimates. While the RADV audits do reconcile the audit findings to the actual payments, they do not apply to all plans, and they focus primarily on overpayments. The RADV audits do make adjustments for underpayments but only to the degree they offset the plans overpayment liability. For this reason and others cited above, MAOs cannot rely on CMS audits to recover underpayments. Any effective approach to reducing the vulnerability of underpayments must alleviate the principal burden of comprehensively reconciling encounters submitted to EDS with HCCs supported in medical record documentation within and across years. There are many tools available to assist MAOs in these efforts. For more information on how RaLytics’ tools can help, please visit ralytics.com/solutions/.