Richard
The Categories That Collapse Under Scrutiny
Not all HCC categories fail audits at the same rate. OIG’s March 2026 audits revealed that certain diagnosis types, specifically single-occurrence acute conditions, fail at dramatically higher rates than chronic disease categories. In the BCBS Alabama audit (A-07-22-01207), acute stroke and myocardial infarction had 100% error rates. In the Priority Health audit (A-07-22-01208), acute stroke and breast cancer hit 100% error rates across all 30 sampled records in those categories.
These aren’t random failures. Single-occurrence conditions follow a pattern that makes them uniquely vulnerable to audit. A patient has a stroke in 2022. The diagnosis is documented in the acute care record. A retrospective chart review identifies the HCC and submits it. The following year, another chart review finds the same diagnosis in the historical record and submits it again. And again the year after that. The original event was real. The problem is that nobody verified whether the condition is still being actively managed in subsequent years.
MEAT criteria require evidence of current clinical management. A stroke that happened three years ago without any subsequent neurological follow-up, monitoring, or treatment documentation doesn’t meet that standard. The diagnosis is historical. The coding treats it as current. The audit rejects it.
Why Retrospective Programs Over-Index on These Categories
Single-occurrence acute conditions carry high RAF coefficients. Under V24, a stroke or MI diagnosis could add significant value to a member’s risk score. Retrospective programs naturally gravitated toward these categories because the financial return per code was substantial. The codes were easy to find in historical records because they’re well-documented during the acute event. The combination of high value and easy identification made them priority targets.
The problem is that high value also means high audit priority. CMS and OIG specifically target high-impact diagnosis categories for sampling because unsupported codes in these categories represent the largest overpayments. Plans that concentrated their coding activity in these high-value categories created audit profiles that concentrated their exposure in the same categories.
Under V28, some of these categories saw coefficient reductions, which means the financial return decreased while the audit risk remained unchanged. Plans still chasing these codes are taking on disproportionate audit risk for diminishing financial reward.
Building Category-Specific Validation
The fix isn’t to stop coding single-occurrence conditions. It’s to apply heightened validation to them. When a retrospective review identifies a historical acute event like stroke, MI, or cancer diagnosis, the system should require evidence of current clinical management before recommending the code. Is there a recent neurology follow-up? Is there current cardiac monitoring? Is there an active oncology treatment plan? Without current management documentation, the code should be flagged as a high-risk submission.
AI-assisted tools can automate this category-specific scrutiny. When the system identifies a single-occurrence HCC in a historical record, it applies a higher evidence threshold. It looks for recent encounters related to the condition. It searches for current labs, specialist referrals, or medication management. If current evidence exists, it recommends the code with supporting documentation. If it doesn’t, it flags the code as indefensible.
This tiered validation approach, where high-risk categories receive more rigorous evidence assessment, reflects how CMS actually audits. Auditors apply heightened scrutiny to the categories that generate the largest payments. Programs that mirror that scrutiny in their own validation process catch the same problems auditors would find.
The Program Design Takeaway
OIG’s 100% error rates in acute stroke, MI, and breast cancer categories aren’t just audit findings. They’re a specification for where Retrospective Risk Adjustment HCC Coding programs need category-specific safeguards. Programs that apply uniform validation across all HCC categories, treating a carried-forward stroke the same as a well-documented chronic condition, will keep producing the failure rates these audits documented. Programs that tier their validation to match audit scrutiny levels will catch the indefensible codes before CMS does.
