Beyond Accuracy: What “Defensible Coding” Really Means Under Today’s RADV Scrutiny

Why It Matters

Defensible coding determines whether health plans face multi‑million dollar settlements or remain audit‑ready, directly impacting financial risk and regulatory standing. The shift reshapes compliance strategies across the industry.

Key Takeaways

• One‑way retrospective adds codes, never removes, raises audit flags
• AI explainability differs from defensibility; intent matters more
• Balanced two‑way coding validates and rejects unsupported diagnoses
• CMS audits six‑year look‑back; penalties rising per finding
• Every diagnosis must be encounter‑linked, clinically evidenced, auditable

Pulse Analysis

Risk‑adjustment programs have long equated coding accuracy with compliance, but recent audits reveal a deeper issue: patterns that prioritize code volume over clinical truth. Investigators examining thousands of charts notice plans that submit endless additions while ignoring deletions, a red flag that the underlying process is engineered for revenue rather than patient‑centered documentation. This shift forces health insurers to reconsider their coding strategies, moving from a checklist mentality to a holistic view that aligns every claim with documented clinical encounters.

Artificial intelligence tools now provide explainability, showing how a recommendation was generated, yet regulators differentiate this from defensibility. Defensibility requires that the system not only surfaces potential codes but also incorporates clinician judgment and removes unsupported entries. A balanced, two‑way retrospective review—one that both confirms appropriate diagnoses and flags excesses—serves as a powerful compliance safeguard. It creates an auditable trail, reduces the risk of costly settlements like the $556 million Kaiser case, and demonstrates intent to uphold clinical integrity rather than chase reimbursements.

CMS’s expanded six‑year look‑back and intensified audit cycles amplify the financial stakes, with penalties per finding climbing steadily. Plans that lack a defensible coding discipline face exposure that can erode profit margins and damage reputation. Building a robust process involves linking every diagnosis to a specific encounter, documenting severity accurately, and ensuring clinicians review AI‑suggested codes before acceptance. By embedding these practices now, insurers position themselves to weather future enforcement actions and maintain regulatory confidence.