← All case studies
Coding Quality Escape Reduction
Calibrated QA, evidence-first workflows, and AQL sampling reduced escapes and rework while improving audit readiness.
Context
- Multi-specialty provider group with decentralized coding (pro-fee + facility).
- Audits flagged inconsistent modifier usage, NCCI edits, and LCD/NCD alignment.
- QA was subjective; no shared defect taxonomy or drift checks across reviewers.
Approach
- Defect taxonomy mapped to payer policy: modifiers, NCCI, ICD specificity, LCD/NCD evidence, and bundling rules.
- AQL sampling by specialty & coder tier with peer review on high-risk classes; double-blind on disputes.
- Calibration cadence (weekly → bi-weekly) with exemplars and tie-outs to payer feedback.
- Playbooks and checklists embedded in coding workbench; evidence tiles required for edge cases.
- Dashboards for FPY, defect mix, rework %, and time-to-detect with drill-downs to encounters.
Measured Results
- Quality escapes −38% (payer/audit findings tied to coding).
- Rework −29% with clearer acceptance criteria and evidence tiles.
- Time-to-detect −21% through targeted sampling and jeopardy views.
- Calibration agreement ≥92% across reviewers by week 6.
Lessons Learned
- Make quality objective: a shared defect taxonomy beats narrative QA.
- Sampling must follow risk: specialty, payer exposure, and coder tier.
- Evidence first: require artifacts at the moment of coding, not later.
- Calibrate continuously; drift returns in 30–45 days without touchpoints.
Stop escapes at the source
Get a coding QA playbook with evidence tiles, sampling rules, and reviewer calibration.
