Desktop Medicare Claim Scrutinizer
Budget: $250 – $750 USD
Building a standalone desktop application that audits Medicare and Medicaid claim files is the core objective. The software must open 837, CSV, or PDF batches, parse each record, and then run three consecutive checks:
• Claim validation – confirm every field complies with CMS specifications and reject or flag those that break structural or data-type rules.
• Billing & coding checks – cross-reference ICD-10, CPT, HCPCS, modifier combinations, and NCCI edits to surface under-coding, over-coding, or incompatible code pairs.
• Fraud detection – scan for duplicate billing, impossible service dates, obvious upcoding trends, and statistically abnormal provider patterns.
A concise dashboard should display real-time progress, while a detailed report (CSV and PDF export) lists all flagged lines with category, rule breached, and suggested correction. Users must be able to adjust rule thresholds and toggle individual checks without needing internet access; everything runs locally.
Deliverables:
1. Compiled desktop installer (Windows preferred, cross-platform welcome).
2. Source code with clear build instructions.
3. Rule-set database in human-readable format for future updates.
4. User guide and short video walkthrough.
Acceptance is complete once a 1,000-line sample file processes in under two minutes and all seeded test errors are correctly identified.
• Claim validation – confirm every field complies with CMS specifications and reject or flag those that break structural or data-type rules.
• Billing & coding checks – cross-reference ICD-10, CPT, HCPCS, modifier combinations, and NCCI edits to surface under-coding, over-coding, or incompatible code pairs.
• Fraud detection – scan for duplicate billing, impossible service dates, obvious upcoding trends, and statistically abnormal provider patterns.
A concise dashboard should display real-time progress, while a detailed report (CSV and PDF export) lists all flagged lines with category, rule breached, and suggested correction. Users must be able to adjust rule thresholds and toggle individual checks without needing internet access; everything runs locally.
Deliverables:
1. Compiled desktop installer (Windows preferred, cross-platform welcome).
2. Source code with clear build instructions.
3. Rule-set database in human-readable format for future updates.
4. User guide and short video walkthrough.
Acceptance is complete once a 1,000-line sample file processes in under two minutes and all seeded test errors are correctly identified.