HIPAA Compliance Auditor

A clinical-grade PII/PHI detection and de-identification tool for healthcare text data.

Overview

This skill analyzes text for HIPAA-protected identifiers and automatically redacts or anonymizes them. It uses a combination of regex patterns, NLP entity recognition, and contextual analysis to identify 18 HIPAA identifier categories.

Features

• - 18 HIPAA Identifiers Detection: Names, dates, SSN, MRN, phone/fax, email, geographic data, etc.
• Automatic De-identification: Replace PII with semantic tokens (e.g., [PATIENT_NAME], [DATE_1])
• Context-Aware Detection: Distinguishes between similar patterns (dates vs. lab values)
• Audit Logging: Track all redaction actions for compliance documentation
• Confidence Scoring: Flag uncertain detections for manual review

Usage

Command Line

Python API

Parameters

HIPAA Identifier Categories Detected

1. 1. Names (patient, relatives, employers)
2. Geographic subdivisions smaller than state
3. Dates (except year) related to individual
4. Phone numbers
5. Fax numbers
6. Email addresses
7. SSN
8. Medical record numbers
9. Health plan beneficiary numbers
10. Account numbers
11. Certificate/license numbers
12. Vehicle identifiers
13. Device identifiers
14. URLs
15. IP addresses
16. Biometric identifiers
17. Full-face photos
18. Any other unique identifying numbers

Output Format

De-identified Text

• - [PATIENT_NAME_1], [PATIENT_NAME_2] ...
• INLINECODE13, [DATE_2] ...
• INLINECODE15
• INLINECODE16, [PHONE_2] ...
• INLINECODE18
• INLINECODE19 (Medical Record Number)
• INLINECODE20

Audit Log JSON

Technical Architecture

1. 1. Preprocessing: Normalize text encoding, handle line breaks
2. Regex Engine: Pattern matching for structured identifiers (SSN, phone, email, MRN)
3. NLP Pipeline: spaCy NER for names, organizations, locations
4. Context Filter: Remove false positives (e.g., "Dr. Smith" vs. "smith fracture")
5. Replacement Engine: Sequential replacement with semantic tokens
6. Validation: Ensure no original PII remains in output

Dependencies

• - Python 3.9+
• spaCy (encorewebtrf or encoreweblg)
• regex (for advanced pattern matching)
• Presidio (optional, for enhanced PII detection)

See references/requirements.txt for full dependency list.

Limitations & Warnings

⚠️ CRITICAL: This tool is designed as a helper, not a replacement for human review.

• - Context-dependent PII (e.g., rare disease names + location) may not be fully detected
• Unstructured narrative text may contain identifying information not caught by patterns
• Always perform manual QA on output before HIPAA-compliant release
• AI Autonomous Acceptance Status: 需人工检查 (Requires Manual Review)

References

• - references/hipaa_safe_harbor_guide.pdf - HIPAA Safe Harbor de-identification standards
• INLINECODE23 - Complete regex pattern definitions
• INLINECODE24 - Sample clinical texts with expected outputs
• INLINECODE25 - Python dependencies

Technical Difficulty: High

Complex NLP pipelines, contextual disambiguation, regulatory compliance requirements.

Risk Assessment

Security Checklist

• - [ ] No hardcoded credentials or API keys
• [ ] No unauthorized file system access (../)
• [ ] Output does not expose sensitive information
• [ ] Prompt injection protections in place
• [ ] Input file paths validated (no ../ traversal)
• [ ] Output directory restricted to workspace
• [ ] Script execution in sandboxed environment
• [ ] Error messages sanitized (no stack traces exposed)
• [ ] Dependencies audited

Prerequisites

CODEBLOCK3

Evaluation Criteria

Success Metrics

• - [ ] Successfully executes main functionality
• [ ] Output meets quality standards
• [ ] Handles edge cases gracefully
• [ ] Performance is acceptable

Test Cases

1. 1. Basic Functionality: Standard input → Expected output
2. Edge Case: Invalid input → Graceful error handling
3. Performance: Large dataset → Acceptable processing time

Lifecycle Status

• - Current Stage: Draft
• Next Review Date: 2026-03-06
• Known Issues: None
• Planned Improvements: