Highmark Issues Public Comment on Centers for Medicare & Medicaid Services Notice

WASHINGTON, March 14 -- Amy Sawyer, director health policy at Highmark Inc., Pittsburgh, Pennsylvania, has issued a public comment on the Centers for Medicare and Medicaid Services' notice entitled "Advance2021PartII". The comment was written on March 6, 2020, and posted on March 13, 2020:

* * *

On behalf of Highmark Inc., I am writing to offer comments on the Advance Notice of Methodological Changes for Calendar Year (CY) 2021 for Medicare Advantage (MA) Capitation Rates and Part C and Part D Payment Policies - parts I and II.

Headquartered in Pittsburgh, Pennsylvania, and as the fourth-largest overall Blue Cross Blue Shield-affiliated organization, Highmark Inc. and its Blue-branded affiliates proudly cover the insurance needs of more than 4.4 million members in Pennsylvania, Delaware, and West Virginia. One of America's leading health insurance organizations and an independent licensee of the Blue Cross Blue Shield Association, Highmark Inc. and its affiliated health plans work passionately to deliver high-quality, accessible, understandable, and affordable experiences, outcomes, and solutions to customers.

As part of our mission, Highmark strives to deliver affordable, high quality coverage to Medicare beneficiaries in the markets we serve and has a deep and longstanding commitment to the Medicare Advantage program. We have been offering Medicare Advantage plans to beneficiaries for over 20 years and currently offer MA plans in Pennsylvania and West Virginia. We have over 235,000 Medicare Advantage members, and are proud to have over 98% of those members enrolled in 4 or 4.5 Star Contracts.

We appreciate the opportunity to comment on the Advanced Notice and the continued partnership to deliver outstanding Medicare Advantage products. Our specific comments are below:

Issue 1: Star Measure Weighting of Statin Use in Persons with Diabetes (SUPD) for Star Year 2021

In Table 1: 2021 Star Ratings Improvement Measures, the Statin Use in Persons with Diabetes measure is listed as an Intermediate Outcomes measure, and its corresponding weight is 3. Although Intermediate Outcomes Measures appropriately maintain a triple-weighting, many stakeholders--including CMS in its Proposed Rule--have asserted this measure is not a measure of outcomes and belongs as a single-weighted process measure. The rule-making process has begun to change the classification to a process measure, but that will not go into effect until the 2023 Star Rating.

In the CY2020 Final Call Letter, CMS stated "For the 2020 Star Ratings, we will keep this measure's weight at 1 as a new measure while we further review this measure's category and consider if CMS should propose to change the technical specifications to reclassify it in the future." We believe this action is sufficient precedent to enable CMS to hold the weight of the SUPD measure at 1 until it is formally changed to a process measure through rule making. Furthermore, in the CY2019 Final Rule, there are no stipulations that would prohibit a measure to be held at a specific weight.

We recommend holding the SUPD measure at 1 weight for Star Ratings for Star Year 2021 and 2022. In last year's final call letter, there was recognition of commenters' feedback that SUPD be considered a process measure, ultimately resulting in SUPD being held at a single weight while further feedback from PQA and other subject matter experts was gathered. In this year's proposed rule, CMS cites PQA's assertion SUPD should be a process measure. We recognize that, in order to formally change the classification of a measure, it must go through the rule making process (which has begun). However, due to the relative newness of the process, and the overwhelming evidence to classify SUPD as a process measure, we recommend CMS hold the weighting of SUPD at 1 until it is formally reclassified.

Issue 2: The introduction of Physical Functioning Activities of Daily Living (PFADL) measure, and the reliability of the Health Outcomes Survey

CMS is proposing the introduction of an additional longitudinal measure from the Health Outcomes Survey, and has requested comment. In order to have an informed opinion on the addition of this new measure, we would need data that validates its reliability. We recommend that, until the reliability of this measure is shared publicly, CMS not introduce this measure.

In the 2019 Final Rule, CMS states guiding principles used in making enhancements and updates to Star Ratings. One of these principles is that "data are complete, accurate, and reliable."/1

Multiple studies validate the reliability of the core components of the two current Outcomes Measures (VR-12 PCS and MCS). However, there is less research on the Performance Measures used in Stars Program (Improving or Maintaining Physical Health, and Improving or Maintaining Mental Health). One such study/2 (focused on individuals with chronic conditions) was conducted by RAND, and its conclusion was that the VR-12 PCS and the VR-12 MCS "point-in-time" measures are reliable, but the measures used to gauge change in those components over time have very low reliability. While the study's focus was not on the entire beneficiary population, its conclusions are still concerning, and indicates possible reliability concerns with the two Star Outcomes Measures. We request CMS publish the reliability of the outcomes measures to respond to this concern.

Our concern for reliability is rooted in trends we have observed in industry-wide performance in the two Outcomes measures. Data indicate that plans lack the ability to influence the Outcomes measures, possibly due to the reliability concerns. While plans are able to maintain consistent levels of performance in Star measures from HEDIS, CAHPS, and other data sources, plans are less able to maintain performance levels in the Outcomes measures obtained from HOS.

To illustrate this point, the table below displays for each measure source the percentage of high-performing contracts/3 in a given year that go on to maintain high performance over a three year period. For example, of the contracts that scored an average HOS measure rating of 4.0 or more in Star Year 2015, only 20% of them were able to maintain that 4+ HOS rating into 2016 and 2017. For the other measure sets, this percentage consistently ranges from 60% to 80% which is substantially higher and indicative of an actual ability to control quality in those areas. Since an additional guiding principle of the Stars program is that "Improvement on measures is under the control of the health or drug plan"/4 this further bolsters the need for a demonstration that the Outcomes measures are reliable indicators of differentiation in plan performance.

See table at https://www.regulations.gov/contentStreamer?documentId=CMS-2020-0003-0850&attachmentNumber=1&contentType=pdf

We also analyzed this issue examining contract rate percentiles as opposed to Star ratings, and the same results persisted--namely that high performing contracts were unable to maintain HOS scores, but were able to maintain scores in measures from other data sources. We did so to test if our conclusions would be addressed by changes to the Star Rating methodology (e.g., cut point guardrails, mean resampling), and we found the changes would not have had an impact for the measures obtained from HOS.

In this year's Proposed Rule, CMS provides insight into the inter-unit reliability of the CAHPS measures, in an attempt to be transparent prior to finalizing changes. We request the same data points be shared for the Outcomes measures (Improving or Maintaining Physical Health, and Improving or Maintaining Mental Health) and the PFADL measure. We recommend CMS set standards for reliability, and publish the reliability scores of all Quality Outcomes measures. To be clear, we are not requesting that reliability for the Physical Component Score (PCS) and Mental Component score (MCS) be shared. Studies have demonstrated the validity and reliability of those scores, but the same has not yet been shared about the Star measures themselves, which take into account risk adjusted comparison in the change in PCS and MCS over time.

Issue 3: Risk Adjustment for Medication Adherence Measures

CMS has suggested all three adherence measures should be risk adjusted for SDS characteristics, at the beneficiary level. In order to give plans the ability to manage medication adherence in a manner which is tailored to unique differences in sociodemographic status, we request greater transparency into how the risk-adjustment methodology will impact final rates and Star scores, during the measurement year.

Since CAI is a means to adjust for plan-level differences in LIS and Disabled beneficiaries, and those differences would be accounted for in these modified adherence measures, we would also request plans understand the corresponding expected impact to the CAI value. Furthermore, we request assurance in Star Year 2024 (the first year with these new measures) the CAI value will be calculated using contract performance of the measures while on display, and not the "legacy" versions of the measures.

We recommend transparency on how this measure will be scored, and its expected impact on CAI.

Highmark supports the movement toward stratification and risk adjustment along SDS characteristics for Medication Adherence. In order to provide plans with better data on how to serve beneficiaries we request:

* Modifications to the Patient Safety Reports to show plans how they will be scored (including risk adjustment) in each of the three Star measures, beginning with the first Patient Safety report of the measurement year

* Detailed documentation of the risk adjustment model used in this new measure

* Clarity on the plan for the calculation of CAI after these measures are introduced, with the expectation that CAI will be calculated (especially in its first year of inclusion as a Star measure) using the risk-adjusted measures

Issue 4: Introduction of Net Promoter Score to the CAHPS Survey

Highmark is concerned about introducing NPS as a measure in the CAHPS survey. Our concerns fall into three categories: 1) reliability, 2) inclusion of employer groups, and 3) timing / implementation.

1. Given the nature of the score, the NPS measure may lack reliability. Nationally, scores in the Rating of the Health Plan and Rating of the Drug Plan measures average in the mid-80s, suggesting a likely mode of 8 or 9 in those responses. We assume that the NPS question would likely to garner similar responses from survey respondents--and thus, many 8s and 9s. Since the underlying NPS measure effectively excludes 7s and 8s from its calculation, this would likely reduce the number of scored responses, and likely greatly reduce reliability.

2. Since NPS is a measurement of the likelihood to recommend, we are concerned that plans higher levels of employer group beneficiaries would be at a disadvantage in this measure. We would assume that beneficiaries in employer groups may be less likely to recommend a plan because they did not have a choice in picking their plan.

3. The maturity of the CAHPS survey as a whole is an asset to plans, in that years of data allow plans to gauge member sentiment over time and intervene appropriately. For this reason, if CMS were to ever introduce NPS as a scored measure, we would request extra time for it to be on display, and for it to be on display in such a way where plans would know what their Star performance is in this measure. Given the limited frequency of the survey and the intricacy of CAHPS scoring, being able to gauge beneficiary response to this question would likely take multiple years.

Given these reasons, we do not support the addition of NPS to CAHPS, and recommend that NPS only be considered if there was a clear expectation that this measure would have to be on display for at least 4 years.

On behalf of Highmark, I thank you for considering our comments and look forward to working with CMS in providing quality MA plans for seniors.

Respectfully,

Amy Sawyer

Director Health Policy

Highmark Inc.

* * *

Footnotes:

1/ "Medicare Program; Contract Year 2019 Policy and Technical Changes to the Medicare Advantage, Medicare Cost Plan, Medicare Fee-for-Service, the Medicare Prescription Drug Benefit Programs, and the PACE Program," [CMS4182-F]; page 16521

2/ Kazis, Lewis E., William H. Rogers, James Rothendler, Shirley Qian, Alfredo Selim, Maria Orlando Edelen, Brian D. Stucky, Adam J. Rose, and Emily Butcher, Outcome Performance Measure Development for Persons with Multiple Chronic Conditions. Santa Monica, CA: RAND Corporation, 2017. https://www.rand.org/pubs/research_reports/RR1844.html.

3/ "High performing Contracts" are defined in this context as plans who score a weighted average across measures obtained from a data source of 4.00 raw stars.

4/ Ibid, page 16521.

* * *

The notice can be viewed at: https://www.regulations.gov/document?D=CMS-2020-0003-0002

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 MYRON STRUCK, editor, [email protected], Springfield, Virginia; 703/304-1897; https://targetednews.com