The Categories Where Documentation Fails Completely
OIG’s March 2026 audit cycle produced a data point that should reshape how every retrospective program prioritizes its chart reviews. Across three separate audits (BCBS Alabama, Priority Health, and Gateway Health Plan), certain HCC categories hit 100% error rates. Every sampled record in those categories failed. Not 80%. Not 90%. Every single one.
The categories with total failure were acute conditions: stroke, myocardial infarction, and breast cancer. These share a characteristic that makes them uniquely vulnerable in retrospective coding. They’re episodic events with high RAF coefficients that appear prominently in medical records during the acute phase and then fade into historical documentation. A stroke is well-documented when it happens. Three years later, the stroke appears in a problem list but the chart may show no neurological follow-up, no ongoing monitoring, and no active treatment plan.
Retrospective programs that treat these categories the same as chronic conditions, where ongoing management documentation is typically stronger and more continuous, are systematically producing the codes that auditors reject at the highest rates.
Why Uniform Prioritization Creates Concentrated Risk
Most retrospective programs use chase lists that prioritize conditions based on RAF coefficient value. High-value conditions get reviewed first. The logic is financial: spend coding effort where the per-code return is highest. Under V24, acute conditions like stroke and MI sat near the top of these lists because their coefficients were substantial.
The problem is that high coefficient also means high audit priority. OIG specifically samples high-value HCC categories because unsupported codes in those categories represent the largest potential overpayments. Plans that concentrated their coding effort in these categories because of their financial value simultaneously concentrated their audit exposure in the categories with the weakest documentation support. They optimized for revenue and created maximum vulnerability.
V28 reduced some of these coefficients, which means the financial return decreased while the documentation problems remained. Plans still prioritizing these categories based on pre-V28 value rankings are chasing diminishing returns at elevated audit risk.
Tiered Validation by Category Risk
The solution is a validation framework that applies different evidence thresholds based on category-specific audit risk. Chronic conditions with continuous management documentation (diabetes, CKD, COPD) require standard MEAT validation. Single-occurrence acute conditions (stroke, MI, cancer) require enhanced validation: the system must find evidence of current clinical management, not just a historical diagnosis mention.
For acute conditions, enhanced validation means searching for recent encounters related to the condition, current specialist involvement, active medication management, and follow-up care plans. If the system finds a stroke in the problem list but no neurological follow-up in the current year, the code gets flagged as high-risk rather than recommended for submission.
AI-assisted tools can automate this tiering. The system recognizes the HCC category, applies the appropriate evidence threshold, and presents the coder with a risk-adjusted recommendation rather than a uniform confidence score.
Aligning Program Design With Audit Reality
OIG’s 100% error rates aren’t a sampling anomaly. They reflect a structural mismatch between how programs prioritize codes and how auditors evaluate them. Retrospective Risk Adjustment HCC Coding programs that tier their validation to match audit scrutiny levels, applying enhanced evidence thresholds to the categories auditors target most aggressively, will catch the indefensible codes before submission rather than discovering them in audit findings. Programs that apply uniform validation regardless of category risk will keep producing the concentrated failures that generate the industry’s worst audit results.
