Intro to Medicare Star Ratings

LinkedIn Note

I'm sharing this article on LinkedIn ahead of it's publication on BHA's website. Since PP2 has already started and 2022 rates are to be published next month, I didn't want to delay getting it out. Please give feedback. This may be an overambitious project, but it'll nicely complement our Stars Tools.

Concept

The Center for Medicare & Medicaid Services (CMS) introduced Star Ratings to help beneficiaries choose higher quality Medicare Advantage plans. (hereafter Plans) Star ratings condense dozens of aspects of performance data into a single number:

5 ***** Excellent
4 **** Above Average
3 *** Average
2 ** Below Average
1 * Poor

The score is updated annually for each health plan.

You are probably already familiar with star ratings from web retailers that display an average of buyer's feedback. If you're selecting a health plan, CMS's website shows Medicare Star Ratings in a similar way, however the ratings are produced in a very different.

For the 2022 fiscal year, there are 40 measures that include such disparate topics as diabetic member's blood sugar control and health plan's accuracy in processing appeals. See the section Star Ratings Calculation for a detailed discussion of the calculation. The simplified version is this:

  • Each Year
    • Each Measure
      • Plans submit data to CMS (the type of data varies)
      • Each Plan gets a numeric score for each measure each year
      • CMS compares scores for all Plans and grades on a curve, awarding 1, 2, 3, 4 or 5 stars for each measure to each plan
    • CMS averages the measure's stars then rounds on the Half-Star

Plans get more than bragging rights from higher Stars. Star Ratings also determine how much money they get from CMS. High performing plans get other rewards and low performing plans get sanctioned. The net is that insurers invest very heavily in improving star ratings.

The definitive guide to Star Ratings is the 175 page Technical Specifications from CMS 2022technotes20210902.pdf (cms.gov), but we think a short tutorial is helpful for some people.

Kinds and Sources of Measures

CMS groups measures into 9 domains that reflect the government's priorities in overseeing MA Plans. At BHA, we usually group the measures differently to reflect the data and mechanism of measurement and how Plans can work to improve them.

Category Type of Data Measurement
HEDIS - Administrative Claims, enrollment, pharmacy data supplemented by feeds from clinical sources Data crunched by the health plan using 3rd party software overseen by a 3rd party auditor. Rates submitted to CMS as a percentage of members who should have a certain service that actually got it.
HEDIS - Hybrid Same as Admin plus manual medical chart review Start with admin rates but augment the score with information found in charts for a sample of members. Improving these results is often the lowest hanging fruit to improve overall star ratings.
Medication Adherence and SUPD Pharmacy Claims, CMS collected enrollment and hospitalization data CMS calculates this and provides monthly reporting of member back to the plans.
Operations Enrollment and other files exchanged between the plan and CMS CMS calculates from data they collect.
CAHPS Surveys Plans contract with one of a handful of accredited survey vendors who conduct the survey and submit data to CMS
HOS Surveys A separate survey: HOS uses cohorts to measure members health over time
Improvement Current and prior year raw scores for star ratings CMS calculates the two heavily weighted improvement measures based on how many measures go up or down (scores, not stars) each year.

Here are CMS's domains. They are quite meaningful, but we will not refer to them again in this tutorial:

  • Part A (Medical plan)
    • Domain: 1 - Staying Healthy: Screenings, Tests and Vaccines
    • Domain: 2 - Managing Chronic (Long Term) Conditions
    • Domain: 3 - Member Experience with Health Plan
    • Domain: 4 - Member Complaints and Changes in the Health Plan's Performance
    • Domain: 5 - Health Plan Customer Service
  • Part D (Drug plan)
    • Domain: 1 - Drug Plan Customer Service
    • Domain: 2 - Member Complaints and Changes in the Drug Plan's Performance
    • Domain: 3 - Member Experience with the Drug Plan
    • Domain: 4 - Drug Safety and Accuracy of Drug Pricing

CMS provides ratings by domain and a rating for each Part C and Part D. In the industry, the vast majority of attention is spent on the Overall Star Rating.

Financial Impact

The Star Rating is calculated as an average then rounded on the half-star. So a score of 3.74 is 3.5 Stars but 3.76 is 4.0 Stars. Benefits, bonuses and penalties accrue at different half-star levels, so the marginal impact of a small difference in one measure can be quite large.

There is no easily calculated Stars Bonus per se. Instead, the star rating plays into 2 of the inputs to the calculation of the Plan's per-member reimbursement from CMS in the following year.

Rating Benchmark Rebate Payment
4.5+ 105% 70%
4 105% 65%
3.5 - 65%
3.0 or lower - 50%

MEDPAC (www.medpac.gov) published an excellent explanation of the actual payment mechanics in Medicare. The core concept is that Plans bid on what they will charge for an average enrollee. The bid is compared with a county specific benchmark calculated from per-capita Original Medicare costs. At 4+ stars, the benchmark is 5% higher. Additionally, CMS give a rebate back to the plan for improved benefits, and the amount of the rebate is tied to Star ratings and the difference between the bid and the benchmarks. All of this is scaled by a risk factor assigned to each individual beneficiary.

Simple enough? If you just want a number for napkin-math, I often use $400/member/year as the value of 4 Stars vs 3.5. That's a very crude estimate; under some conditions it could be nothing. 3.5 Stars is also valuable, about $100 more than 3 Stars, and 4.5 Stars is about another $50. That money is often the entire difference between an annual profit and loss.

If you think about a plan with 100,000 members that is just short of 4 Stars, they could lose out on $40M for falling a handful of records short in one measure like A1C control. We discuss this more in the Stars 201 tutorial.

Cutpoints and Weights

Every measure is a numeric score calculated each year. (usually as a percentage) This number is translated to 1, 2, 3, 4 or 5 stars based on the range of scores for each measure. CMS changes the cutpoints annually after they collect all the data from all the Plans. The September day when you find out the cutpoints is either the most joyous or most frustrating day of the year for a Plan's Stars leader.

The cutpoints are in the Tech Specs, but an easier way to find them is to click on the Measures app on our website. We also graph the cutpoint history.

One last note about the Stars calculation. They are not equally weighted, and the weights change from time to time depending on what CMS decides to emphasize. We discuss this further in the Stars 201 tutorial.

Timeline

CMS publishes Stars information on its website (Part C and D Performance Data | CMS) and in the Federal Register. The same general timeline is followed every year. Here's what has or will happen for 2022 Star Ratings.

2020 All Year Members get healthcare and the Plan processes claims, etc. The 2022 Star ratings are designed to reflect the Plan's performance in 2020.
2021 Spring CMS and Plans collect HEDIS, CAHPS, HOS, Rx and other data to calculate raw scores.
2021 February CMS release the Advance Notice which outlines their ideas for star ratings changes.
2021 March/April CMS Release the Final Call Letter in which they respond to feedback and announce their final plans for the year's Star Ratings and their plans for the future.
2021 August Plan Preview 1 - CMS provides Plans their scores for each measure for review.
2021 September Plan Preview 2 - CMS calculates cutpoints and provides plans a second chance to review their scores with cutpoints and calculated Star Ratings before they're published.
2021 October Star Ratings Published. CMS updates the website plan finder with Star Ratings. CMS provides all data for all plans, along with a trove of analyzed summary data.
2022 January New star ratings are effective with payment impact

Some Jargon

Here are a few critical concepts to understand. See our Glossary for more extensive definitions.

MAPD

Medicare Advantage with Part D. The information in this tutorial is geared towards plans that offer both medical and pharmacy benefits. Star Ratings are calculated for other plans, but they do not have the same financial impact or get the same attention.

Contract

The contract is the basic unit of business between a health insurance company and CMS. A large national insurer may have dozens of contracts, each of which is offered in different counties possibly in different states. Star Ratings are assigned to contracts.

Hcontract a colloquialism because most (but not all) contracts are identified in the form of H and a 4 digit number, e.g. H1234.

Gaps

A gap in care is a particular member who does not meet the criteria to be up to date (compliant) for a particular service. For example, a woman aged 50-74 who has not had a mammogram in the past 2 years is a gap for the breast cancer screening measure. A gap is closed when the member becomes compliant. Gaps can be one and done if a single claim can make them compliant. Or they can be transient -- for example a member can be up to date with his statins for a few months, stop taking it and become a gap and then get back on track and close the year compliant for that measure.

Clinical Campaigns to Improve Ratings

Strategies for improving scores are as diverse as the measures. Most efforts involve operational improvement, benefit changes, improving member communication, expanding data collection, or direct clinical interventions with the members. For the clinical campaign to work, it must happen while there is still time to close the gap before Dec 31.

Clinical campaigns involve 2 pieces:

  • A way to identify which members have gaps in care and predict which ones need a nudge towards closing them.
  • A way to nudge the member towards closing the gap

In another section, we discuss some different forms of nudges. The first thing to understand is just that these campaigns exist. Plans invest heavily in these interventions; it's probably the largest budget item in Stars strategy.

Baltimore Health Analytics

Since this is BHA's website, we get to plug our services in our tutorial. Our most popular -- and free -- tool is the Stars Planner which allows you to view a Plan's most recent stars data and test different strategies to get to 4 Stars. It's also a great way to get a more intuitive grasp of the scoring system.

We, of course, want paid business as well. We can assist in a number of ways:

  • Consult with health plans as an expert resource to help design and guide programs to improve Star Ratings
  • Create database routines to Identify, triage and prioritize gaps in care for outreach
  • Develop reports to help manage Star Ratings programs
  • License our software to manage clinical campaigns
  • Review and help improve HEDISxAE data collection efforts
  • License our software to scan charts and manage the data
  • Analytics and data processing support

Sections of this Tutorial

We are building this tutorial with each section created to be self-standing. I say building as we're going to start publishing articles after the first 5 are ready, then will continue adding as time allows. Here's the planned outline:

  • Basics

    • Star Ratings 101
    • Calculation
    • Data and Documentation
    • BHA Free Tools
    • HEDIS 101
    • Survey Measures 101
    • Med Adherence 101
    • Other Pharmacy Measures
    • Operations/Other Measures
  • Advanced

    • Gaps in Care
    • Stars 201 - Planning
    • HEDIS 201 - Improving Scores
    • Clinical Interventions
    • Providers Perspective
    • Risk & Stars
    • Analyzing Scores
    • Tracking
    • Covid
  • Data & Tools

    • Public Data Sources
    • BHA Open Stars Tools
  • Resources

    • Glossary
    • Other Resources
    • Ask A Question

Copyright

This tutorial is copyright 2021 Baltimore Health Analytics LLC. It may not be reproduced in whole or in part without prior authorization.