To account for differences in disease burden within a patient population with Medicare Advantage (MA) plans, risk adjustment based on the patient’s disease burden is used. In particular, MedPAC notes that:
Medicare uses beneficiary characteristics such as age and preexisting conditions, as well as a risk adjustment model—the CMS Hierarchical Condition Categories (CMS–HCCs)—to develop a measure of their expected relative risk for covered Medicare expenses.
In February 2023, CMS CMS issued a notice of proposed rulemaking to update its HCC risk adjustment algorithm (v28). These changes included (i) the use of ICD-10 codes instead of ICD-9 codes as primary building blocks, (ii) the use of 115 HCC indicators instead of 79, and (iii) the restriction of some coefficients to equality of all levels of severity (e.g. diabetes, heart failure). The new algorithm will be phased in over 2024-2026.
An important question is whether providers with traditional Medicare(TM) plans differ from those with Medicare Advantage (MA) plans. Because CMS’s MA plan payment depends on patient severity, there is an incentive to upcode diagnoses. An article by Carlin et al. (2024) aims to evaluate whether this is happening or not. They first explain the mechanism by which MA plans could better capture patients’ secondary diagnoses:
MA plans have the ability to review medical records to ensure providers have not inadvertently omitted a diagnosis from the medical record. These reviews are even more important when provider reimbursement does not require detailed coding of patients’ secondary diagnoses. MA plans to make corrections to add or (rarely) delete diagnosis via CR records. Additionally, both MA and TM providers can record additional diagnoses through an HRA [health risk assessment] during a wellness visit or a home visit for this purpose.
The authors use 2019 CMS claims data and divide the data into three cohorts: MA plans, TM beneficiaries assigned to ACOs (“TM ACO”), and TM beneficiaries not assigned to an ACO (“TM Non-ACO”). ACO includes patients who can be assigned to Accountable Care Organizations (ACO), such as those participating in the Medicare Shared Savings Program (MSSP). The authors note that the TM non-ACO cohort serves as an important comparison because it is not subject to the same coding intensity incentives as MA plans and TM ACOs (as shared ACO savings are also risk-adjusted).
The authors identify patients who had an HRA based on whether they had an annual health visit, an initial preventative physical examination, or selected home health visits (following the algorithm of Reid et al. 2020). The authors also use information from encounter claims about whether a review of patient records occurred. Using these data, the author’s propensity score matched the MA, TM-ACO, and TM-non-ACO cohorts. The authors then compare the concordant and discordant HCC scores and evaluate how the HRA and CR visits impacted the HCC risk scores. You find:
Incremental health risk due to diagnoses in HRA datasets increased across coverage cohorts in line with incentives to maximize risk scores: +0.9% for TM non-ACO, +1.2% for TM ACO, and +3.6% for M.A. Including HRA and CR records, MA risk scores increased by 9.8% in the corresponding cohort.
Diagnosis codes related to vascular disease, heart failure, and diabetes contributed the most to the average HCC score in all three cohorts. Vascular, psychiatric and congestive heart failure were most likely to increase due to HRA/CR coding intensity activities.
While other publications have claimed that Medicare Advantage upcoded diagnoses for more favorable reimbursement, this document clearly lays out not only the magnitude of the impact, but also the mechanism through which it is most likely to occur. You can read the full article here.