Claims Engine

Given a batch of 100 837P files arrives from the clearinghouse, When the ingestion pipeline runs, Then all 100 are parsed into structured claim records with: patient ID, provider NPI, diagnosis codes (ICD-10), procedure codes (CPT), billed amount, and service dates.

Given a file has invalid segments, When parsing fails, Then the file is quarantined with an error report listing the specific invalid segments and line numbers.

Given an 835 file arrives, When it is parsed, Then each payment/adjustment is matched to the original 837 claim by claim ID, and the claim record is updated with: paid amount, adjustment reason codes, and patient responsibility.

Given an 835 references a claim not yet in the system, When the match fails, Then it is queued in an "Unmatched Remittance" holding area and re-matched on each subsequent 837 ingestion.

Given a claim arrives with a service date in the future, When validation runs, Then the claim is flagged with "Invalid service date — future date detected" and routed to the exception queue.

Given a duplicate claim (same patient, provider, date, codes) already exists, When validation runs, Then it is flagged as "Potential duplicate" with a link to the original claim for manual review.

Given a claim includes CPT 27447 (knee replacement), When the claim is submitted, Then the system checks the payer's prior auth requirements and if required, auto-generates a prior auth request with the clinical documentation attached.

Given a procedure does not require prior auth, When the claim is submitted, Then it bypasses the auth workflow and proceeds directly to adjudication.

Given I have 12 pending auths, When I open the Prior Auth dashboard, Then I see each with: patient name, procedure, payer, submitted date, days pending, and current status.

Given an auth has been pending for more than 14 days, When I view the dashboard, Then it is highlighted in red with an "Escalate" button that generates a follow-up to the payer.

Given a prior auth for MRI is denied, When I click "Appeal," Then the system generates an appeal letter template with the denial reason, relevant clinical notes, and lab results pre-attached.

Given an appeal is submitted, When the payer responds, Then the auth status updates to "Appeal Approved" or "Appeal Denied — Final" and I am notified.

Given claims data for a population of 10,000, When the HEDIS engine runs, Then it calculates numerator/denominator for each applicable measure and produces a HEDIS scorecard with rates and benchmarks.

Given a measure is below the 50th percentile benchmark, When the scorecard is generated, Then it is flagged "Below Benchmark" with a gap analysis showing which patients are in the denominator but not the numerator.

Given a diabetic patient hasn't had an HbA1c test in 12 months, When I open their record, Then a care gap alert shows "Overdue: HbA1c screening (Comprehensive Diabetes Care measure)."

Given a provider bills 50 knee replacements in one day, When the anomaly engine runs, Then a "Volume Anomaly" flag is raised with the provider's billing pattern compared to peer averages.

Given a claim bills E&M level 5 (99215) for a 2-minute visit, When the upcoding detector runs, Then it flags "Potential Upcoding — high complexity code with minimal time."

Given I search by claim ID or patient name, When results appear, Then I see: submitted date, current status (received/in review/adjudicated/paid/denied), and if adjudicated, the payment amount vs. billed amount.