Featured Health Business Daily Story, Dec. 5, 2016

CMS Socioeconomic Adjustments Don’t Move Stars Much for Adherence Measures (with Table: Characteristics Contributing to the Observed Disparities in Stars; Chart: Medication Adherence Cut Points)

Reprinted from DRUG BENEFIT NEWS, biweekly news and proven cost management strategies for health plans, PBMs, pharma companies and employers. Sign up for an $86 two-month trial subscription today.

By Jane Anderson, Senior Writer
November 18, 2016Volume 17Issue 22

Differences in medication adherence measures did have an effect on CMS’s 2017 star quality ratings as part of an interim adjustment used to address socioeconomic status (SES) in the star quality ratings program. But the overall effect of the SES adjustment was quite small, affecting only a handful of Medicare Part D plans, according to an analysis from Inovalon Inc. and its health care consulting subsidiary, Avalere Health.

These results indicate there’s more work to be done for CMS and other stakeholders when learning how best to adjust star ratings for duals’ characteristics, said Christie Teigland, Ph.D., Avalere’s vice president of advanced analytics. Inovalon conducted its own analysis, the results of which show that additional adjustments for factors linked to members’ lower SES could shake up star ratings far more, and potentially more fairly.

“It’s not a stretch for us to assume that the low dual contracts are in low poverty areas and that the high dual contracts are in the high poverty areas,” Teigland said. “Do we really think that the differences in performance on these measures in the low poverty areas compared to the high poverty areas represents the differences in quality?” she asked during the Nov. 2 AIS webinar, “The 2017 Medicare Star Ratings: How to Translate New CMS Data Into Future Successes.”

Teigland said she is currently working with an expert panel of the Pharmacy Quality Alliance, which develops the three medication adherence measures, on how best those measures could be adjusted for SES. The panel “is making some significant progress…in trying to find the right socioeconomic data to use in its recommendations,” she added.

Drug Benefit News

SES Adjustment Had Small Effect

After two major studies supported the theory that socioeconomic factors such as income, education and social supports can have an adverse effect on star quality ratings, CMS late last year considered a pair of complex approaches for adjusting the 2017 star ratings to take into account the impact of SES and disability.

Ultimately, the agency decided to apply a categorical adjustment index (CAI) factor to a contract’s overall and/or summary rating to adjust for within-contract disparities based on an Medicare Advantage (MA) contract’s percentages of low-income subsidy (LIS)/dual-eligible members.

Out of the 44 star measures, only seven were chosen to address SES in MA star ratings. The only Part D measure included was Medication Adherence for Hypertension, which according to CMS data averages about 3% lower for members in LIS/dual-eligible populations.

According to Teigland, Inovalon’s analysis of the 2017 star ratings data found that only 4% of all MA and Prescription Drug Plan (PDP) contracts — a total of 15 contracts — ultimately were affected by this adjustment, and all moved in a positive direction.

Each of those affected gained half a star: One contract moved from 2 stars to 2.5 stars, one rose from 2.5 stars to 3, eight went from 3 stars to 3.5 stars, and five increased from 3.5 stars to 4 stars, said Patrick Donnelly, director of product execution for Inovalon’s Quality Spectrum solution, who also spoke at the AIS webinar. “That’s incredibly significant for those plans, especially when they’re dealing with a disadvantaged population.”

Nonetheless, 96% of contracts saw no change in star rankings, meaning the overall effect of the SES adjustment was minor. That’s in line with earlier CMS estimates that applying the CAI data to the 2016 star rankings would have moved 11 plans up half a star.

Inovalon’s study used a highly representative sample of the Medicare population and market indices data from Acxicom. The study showed that various member characteristics common among the LIS/duals population increased disparity in Part D’s three medication adherence measures (hypertension, diabetes and cholesterol) after adjusting for dual status, Teigland said.

For example, members who have a history of substance abuse or who live in a high poverty area tend to have worse scores on those three measures, along with all-cause readmissions. The health care data analytics firm also found a significant impact from chronic conditions, age, gender, living in a primary care shortage area and never-married status (see chart, p. 6).

Stratifying the hypertension medication adherence measure simply by LIS status produced results that were similar to those achieved by CMS with its SES adjustment, she said. However, applying a more comprehensive risk adjustment — one that included more factors related to non-compliance, such as gender, race/ethnicity and the percentage of the neighborhood population below the poverty level — led to many plans changing rank.

The results of this analysis were consistent for all three medication adherence measures, Teigland said. “Once you control for all the characteristics of members, you see plans that are doing a much better job than the star ratings are saying they are.” Higher poverty and other indications of lower socio-economic status lead to worse performance on all three medication adherence measures, she said.

Contracts with a higher proportion of non-duals are much larger and have significantly lower risk scores on average, Teigland added. Risk scores are directly correlated with the percentage of contracts representing duals and are 50% higher among members in contracts with fewer than 20% duals compared to those with more than 20% duals.

“Contracts with less than 20% dual/LIS enrollees consistently perform best every year while contracts with more than 20% dual enrollment consistently perform worse every year,” she said. “For a given plan sitting out there with a large population of disadvantaged members, this could be a really significant effect. They could be doing a really great job compared to other plans with similar kinds of disadvantages, but that’s not being reflected in current performance rates.”

Teigland noted that the Inovalon analysis “found the opposite happening as well — we saw plans taking care of wealthier, healthier members who are doing a worse job than we would expect.”

According to Avalere, MA and PDP contracts underperformed for 2017 compared to 2016 on several Part D measures, including:

  • Medication Adherence for Cholesterol (-0.5% overall performance for 2017 compared to 2016).

  • Appeals Upheld (-0.4%).

  • Medication Adherence for Diabetes (-0.4%).

  • High-Risk Medication (-0.4%).

“All three drug adherence measures went from being among the highest [performing measures] in 2016 to the lowest in 2017,” said Donnelly. Inovalon’s research and data-crunching also showed how in 2016, a few “very low-performing contracts” on the medication adherence measures skewed the cut points, leading to significant churn in the star ratings. “Plans that may have had very stable performance year-over-year saw these star jumps,” he said. “Flash forward to 2017, and those previously poor-performing plans brought their performance at least in line with low star clusters,” bringing performance back in line with historical trend lines.

“On the PDP side, the most significant change occurred in the medication adherence for diabetes measure, where the 4- and 5-star cut points — which were extremely high last year, well into the 90th percentile — dropped almost 10% to return to their historical trend,” Donnelly said (see chart, p. 7).

Summary of Characteristics Contributing to the Observed Disparities in Stars

Star Measure

MA Member Characteristic

Rheumatoid Arthritis Mgmt.

Breast

Cancer

Screening

High Risk Meds

Medication Adherence

All Cause Readmission

Hypertension

Diabetes

Cholesterol

Alcohol/Drug/Substance Abuse

+

+

+

+

+

+

 

Lower Home Ownership Area

+

+

+

+

+

 

Disability as Original Reason for Entitlement

 

+

+

+

+

+

 

Living in Primary Care Shortage Area

 

+

-

+

+

+

Living in High Poverty Area

 

+

+

+

+

Male Gender

-

+

+

-

-

Age

-

-

-

+

+

+

Race/Ethnicity

-

-

+

+

+

Percent of Population Never Married

 

+

+

+

 

+ Increases disparity in rates; - Reduces disparity in rates

SOURCE: “An Investigation of Medicare Advantage Dual Eligible Member-Level Performance on CMS Five-Star Quality Measures,” Inovalon White Paper, March 2015. Available at: http://resources.inovalon.com/h/i/54960697-an-investigation-of-medicare-advantage-dual-eligible-member-level-performance. Presented during the Nov. 2 AIS webinar, “The 2017 Star Ratings: How to Translate New CMS Data Into Future Successes.” Visit the AIS MarketPlace to order a recording and accompanying materials.

Medication Adherence Cut Points

Medication Adherence Cut Points rebounded in 2017 Star Ratings, leading to lower Part D and Overall performance.

Medication Adherence Cut Points

Notes: In the pie charts, the star percentages are in descending order clockwise from the 5-star segment, which is the dark segment on the right/upper right. In the bar charts, the star percentages are in descending order from the top bar, which indicates 5 stars.

SOURCE: Inovalon. From the Nov. 2, 2016, AIS webinar, “The 2017 Medicare Star Ratings: How to Translate New CMS Data Into Future Successes.”

© 2016 by Atlantic Information Services, Inc. All Rights Reserved.


For a recording of the Nov. 2 webinar, visit https://aishealth.com/marketplace/c6a24_110216.

It's quick and easy to sign up for FREE access to AISHealth.com!

Why do I need to register?