Claims Scrubbing + ADR Risk

Revenue CycleAsync jobs required

Batch CMS-1500 scrubbing with first-pass scoring and ADR (records-request) risk.

ClaimsCMS-1500ADR riskNCCI/MUECARC

About

Pre-submission scrubbing for professional claims at batch scale. It OCR-extracts scanned CMS-1500 forms (or ingests structured claim JSON), applies a deterministic edit library — format/demographic, eligibility, NCCI/MUE, modifier, authorization, timely-filing — and computes a first-pass score plus a calibrated ADR (additional-documentation-request) probability.

Each flag carries a rule id, layer, severity, the governing rule, a CARC mapping, correction guidance, and an audit verdict; a portfolio summary reports clean-claim rate, dollar-weighted first-pass yield, and ADR-band distribution.

How it works

  1. 1/scrub: OCR (scanned form) or structured JSON → deterministic edits → first-pass score + ADR model → LLM audit of each flag
  2. 2/nonpayable: same upload contract → payable vs non-payable / patient-billable triage

Intended use

  • •Batch pre-bill scrubbing and clean-claim-rate programs
  • •ADR-risk triage — work the claims most likely to draw a records request
  • •Non-payable routing to patient billing

Key outputs

  • ▸claims[] — determination, first_pass_score, score_band, adr {probability, band, factors}, flags[]
  • ▸Per flag — rule_id, severity, governing_rule, carc, correction_guidance, audit_note
  • ▸summary — clean_claim_rate, dollar_weighted_first_pass_yield, adr_band_distribution, reference_versions

Endpoints

Try each endpoint with your signed-in session — usage counts toward your monthly budget.

Use synthetic data only. Do not submit real patient records or PHI when testing endpoints.

Limitations & caveats

  • –Scope is CMS-1500 professional claims (one claim per scanned PDF); UB-04 is not supported
  • –Reference edits (CARC/LCD/MUE/NCCI) are a curated illustrative subset, not the licensed full tables
  • –OCR + rules (30–90s per scanned form); the async /jobs flow is mandatory for scanned uploads