Health Care Service Issues Public Comment on Centers for Medicare & Medicaid Services Proposed Rule
* * *
Thank you for the opportunity to provide comments on behalf of
CMS is to be commended for its open, transparent process of holding stakeholder listening sessions and publishing a white paper to gather feedback from all stakeholders as it developed the policies that are outlined in this proposed rule. The risk adjustment program is a critical component of today's guaranteed issue, community rated individual and small group markets, allowing issuers to price coverage to the average risk of their markets with the expectation that they will be reimbursed for the additional risk of sicker than average enrollees. In the process, risk adjustment helps reduce any incentives that issuers might have to seek out healthier enrollees at the expense of those with more serious medical needs. RADV audits are, in turn, an indispensable component of that risk adjustment program. For risk adjustment to work, it is critical that issuers' self-reported EDGE submissions accurately reflect the actual risk of their memberships and that the RADV methodology and adjustments provide adequate incentives for issuers' EDGE submissions to be submitted with integrity and quality while validating and assessing each carrier's risk.
HCSC believes that CMS was correct to establish in regulation that the RADV should be attempting to validate our members' health status through the review of all relevant medical record documentation. For this reason, the RADV methodology should ideally be based upon an assessment of differences in risk scores between an issuer's membership as represented in its EDGE submissions and its audited RADV sample. The current methodology, which is based upon counts of HCCs before and after the audit, treats all HCCs within a given failure rate group as interchangeable, despite any differences (or lack of differences) in coefficients for the HCCs. We encourage CMS to continue to evaluate revisions that would move the RADV methodology towards one that measures differences in risk scores between issuers' EDGE submissions and audited RADV samples, rather than HCC counts.
On balance, however, we agree that the current methodology works well and provides appropriate incentives for issuers to submit accurate data to EDGE servers and to verify those data during the audit process. However, there are some areas in which modifications could improve accuracy and reduce volatility. In evaluating any changes to the methodology, however, CMS should take care to avoid modifications that would reduce the incentives for accurate EDGE submissions in the name of reducing volatility and improving predictability. We strongly believe that well-designed changes can both reduce volatility and maintain appropriate incentives for EDGE submission accuracy.
Proposed Super HCC Grouping is a Good Refinement to the RADV Methodology
RADV errors that reflect miscoding of HCCs within a coefficient estimation group, where there is no difference between the coefficients of the added and non-validated HCCs, do not have any impact upon the issuer's risk score. HCSC believes that it is therefore appropriate that these errors should also not be reflected in RADV error rates and risk score adjustments. We agree that the "super HCC" methodology laid out in the proposed rule is an appropriate way to accomplish this change and urge its finalization. HCSC also agrees with CMS that excluding a priori stability constraints on the child model from the super HCC groupings is the correct approach. These HCCs are additive and are not likely to be mistakenly coded for each other and should therefore not be kept together in the same failure rate groups.
Proposed Sliding Scale Adjustment Strikes the Appropriate Balance Between Mitigating the Payment Cliff and Maintaining Incentives for Accurate Coding
For some time, stakeholders have identified the "payment cliff" problem as one of the most pressing shortcomings of the RADV methodology. Under the current methodology, a very small difference in RADV results--the ability to verify one or two additional diagnoses--can lead to very substantial differences in financial outcomes among issuers and introduce significant volatility into the program. This, in turn, may result in more conservative pricing by issuers in the future to hedge against this volatility. The sliding scale approach proposed by CMS in this rule is an appropriate means of addressing this problem, as it would reduce the marginal impact of each HCC for issuers with group failure rates close to the edges of the confidence intervals without reducing the adjustments for issuers that have larger group failure rates and the incentives for them to improve EDGE submissions. It is also important that the proposed adjustment reduces volatility without excessive reductions to the incentives for proper coding.
Other options that were explored in the 2019 RADV White Paper would have very significantly reduced overall RADV adjustments, negatively affecting the RADV audit's ability to maintain the integrity of the risk adjustment program. The White Paper's "Option 1," under which adjustments would have begun at the boundaries to the current methodology's 95% confidence intervals, for instance, would have reduced overall adjustments in benefit year 2017 audits by nearly 60% in the individual market. Some might argue that in order to maintain an appropriate level of adjustments similar to what is seen in the current methodology, the sliding scale's starting points should reflect an 80% confidence interval instead. We believe that the White Paper's "Option 2," which is set forth in this proposed rule and phases in adjustments beginning at a 90% confidence interval, strikes the appropriate balance between reducing volatility and maintaining incentives for accurate EDGE submissions. The gentler slope of this option and the fact that it does not as dramatically reduce the overall magnitude of adjustments make it preferable to Option 1. CMS should finalize the sliding scale adjustment as proposed.
Negative Failure Rate Outliers
The RADV methodology is aimed at validating and assessing an issuer's true risk through the verification of its EDGE submissions. Sometimes, for one reason or another, the process of verifying those submissions reveals that an HCC was unsupported by the medical records. Other times a new diagnosis that the issuer is unaware of is uncovered in medical records. Every issuer should get credit for the risk they carried on the membership they insured and the enrollee whose HCC is uncovered through the RADV process does not represent any less risk than one whose HCC was previously known. In both cases, the issuer bore the risk of treating the member's health care condition and it is appropriate that the issuer get full credit in risk adjustment for doing so.
The NPRM proposes to add a constraint to the group adjustment factors of negative error rate issuers with negative failure rates to eliminate adjustments attributable to newly found HCCs that increase the issuer's total HCC count in a failure rate group beyond the number of HCCs initially submitted through EDGE. As a result, two issuers that identified the same number of new HCCs during the RADV process would not receive the same credit for those HCCs when risk scores are adjusted. The issuer with higher validation rates for EDGE-submitted HCCs would receive less credit than its competitor that had submitted a larger number of HCCs through its EDGE server that could not be verified in the RADV audit. Because of this unfair penalty upon issuers that submitted fewer unverifiable HCCs, CMS should not finalize the proposed constraint on issuers with negative error rates.
Some have expressed concern that issuers might try to game the RADV process by deliberately underreporting their risk in EDGE submissions with the hope of later identifying HCCs in RADV. An issuer taking such an approach would be taking an extraordinary risk, assuming first that its RADV sample will contain individuals with HCCs that were previously unknown, second that it will be able to identify those HCCs in RADV, and third that the issuer will be an outlier and the new HCC will be reflected in a RADV adjustment. A miscalculation on any one of these assumptions would have significant financial consequences that already effectively deter this sort of behavior by issuers. Furthermore, the difference in timing between payment of risk adjustment transfers and the payment of RADV adjustments creates a loss of cash flow for issuers that also makes intentional underreporting of risk in EDGE submissions an unrealistic strategy.
Issuers understand that is always a better approach to get credit for all of their risk in the original risk adjustment process, rather than through RADV. For this reason the proposed rule correctly notes that "any issuer with a negative failure rate is likely to review their internal processes to better capture missing HCCs in future EDGE data submissions." Indeed, issuers already have sufficient incentives to capture all of their membership's diagnoses in EDGE submissions, and most work hard to do so and to improve their EDGE submissions each year. The addition of supplemental diagnoses on EDGE is one tool that issuers have at their disposal to improve their EDGE submissions and to ensure that their risk scores accurately reflect the true morbidity of their membership. However, claim matching rules and coding guidance governing the submission of supplemental diagnoses on EDGE are more restrictive than those for newly identified diagnoses in RADV.
For instance, the claim matching rules for supplemental diagnosis code captures require that the code must have an exact match to a paid claim, whereas in RADV, the medical record must have a paid claim from the audited benefit year within it. This becomes important, for example, if a member enrolled with an issuer midway through the benefit year after having received a diagnosis earlier in the benefit year. Because of this discrepancy, issuers may be limited in their ability to fully reflect their membership's risk in supplemental EDGE submissions and a newly added HCC in RADV may be the only avenue for them to receive credit for some diagnoses. CMS should consider aligning the more restrictive EDGE claim matching rules and coding guidance to those used for RADV as an alternative to constraining negative error rates.
Changes to the RADV Methodology Should Not Be Implemented for Past Benefit Years
Each of the foregoing modifications to the RADV methodology are proposed to be implemented for the 2019 benefit year RADV audit. Annual rate filings reflect issuers' estimates of risk adjustment transfers, including the impact of RADV adjustments for a given benefit year. While CMS has issued guidance permitting RADV adjustments to be reflected in premiums in the year that the adjustments are actually paid out, this guidance conflicts with actuarial standards of practice and state law, which prohibit the recoupment of past losses in rate filings. Although the 2019 benefit year RADV has not yet occurred, issuers assumed that the audits would be conducted under the rules in place when benefit year 2019 policies were priced in 2018. The application of these changes to the 2019 benefit year's RADV process amounts to a retroactive rule change. CMS should instead apply these changes prospectively to the RADV processes for future benefit years.
Application of RADV Adjustments to the Audited Benefit Year
While HCSC agrees that RADV audit results should ideally be aligned with the risk adjustment year being audited, we do not believe that the "average error rate approach" proposed in the NPRM is the best way to accomplish this transition. By averaging the results of RADV audits for the EDGE submissions underlying two years of risk adjustment calculations, this approach would essentially reduce by one half the RADV adjustment attributable to each year of audited EDGE data. An issuer that had overstated its risk on EDGE in both of those years would therefore receive an adjustment approximately equivalent to what it would have received from one year's overstatement in a non-transition year. An issuer with advance knowledge of this approach would have an incentive to take liberties with its EDGE submissions with the understanding that its penalty for doing so would be reduced by half.
We believe that the alternative "combined plan liability risk score option" laid out in the proposed rule is a preferable approach. While this will result in a single year in which the adjustments from two RADV cycles are applied simultaneously each risk adjustment cycle would be corrected by approximately the same amount it would be under the current methodology and the incentive for accurate coding would not be reduced under the combined approach.
In addition to the two approaches outlined in the proposed rule, CMS might wish to consider other approaches. The first of these would be to simply maintain the current application of RADV adjustments to the following benefit year's risk scores and risk adjustment transfers. Finally, CMS should consider the impact of the COVID-19 pandemic upon the RADV process as it considers these options. HCSC believes that the pandemic is likely to make effective RADV audits for the 2020 benefit year difficult, as business interruptions are likely to have affected provider coding practices and may also make it difficult for providers to respond to issuer requests for the medical documentation needed to complete the audit. These effects are likely to be unevenly distributed across the country, with the effects being more pronounced in the cities, states and regions that have been more heavily impacted by the pandemic. This will place those issuers serving these areas at a significant disadvantage, as all issuers are judged against national mean failure rates. For this reason, CMS may wish to consider cancelling the 2020 RADV process and using instead an average of issuers' results from the 2019 and 2021 benefit years' RADV audits.
Thank you again for the opportunity to comment on this proposed rule. Please do not hesitate to call upon us if there is anything we can do to be of assistance.
Sincerely,
Senior Vice President and Chief Actuary
* * *
The proposed rule can be viewed at: https://www.regulations.gov/document?D=CMS-2020-0059-0001
TARGETED NEWS SERVICE (founded 2004) features non-partisan 'edited journalism' news briefs and information for news organizations, public policy groups and individuals; as well as 'gathered' public policy information, including news releases, reports, speeches. For more information contact
Sens. Brown, Casey Urge Social Security Administration to Once Again Suspend All Continuing Disability Reviews During COVID-19
Rep. Sarbanes Secures Federal Funding to Support National Health Care Workforce Commission
Advisor News
- Americans’ saving, investment product savvy improves, study finds
- Bull or bear? What are financial advisors’ market expectations for 2025?
- Panel: Is the American retirement system ‘broken’?
- Goldman Sachs, others weigh in on what’s ahead for the market in 2025
- How is consumers’ investing confidence impacting the need for advisors?
More Advisor NewsAnnuity News
Health/Employee Benefits News
- The illusive sunny side
- West Virginians say medications have helped them lose weight, but coverage can be a problem
- Kitsap could lose last hospital in United negotiations
- Most Americans blame insurance profits and denials alongside the killer in UHC CEO death, poll finds
- IBX will no longer cover drugs like Ozempic when they are prescribed solely for weight loss
More Health/Employee Benefits NewsLife Insurance News