NPI
Detects US National Provider Identifier (NPI) values structurally: a ten-digit number beginning with 1 or 2. Detection is structural only — the NPI Luhn checksum (80840 prefix) is NOT enforced. Confidence comes from healthcare context tiers (NPI-specific terms at 85, broader provider/claims context at 75, bare shape discovery-only at 65), and an all-same-digit guard rejects repeated-digit values such as 1111111111.
- Type
- regex
- Engine
- universal
- Confidence
- medium
- Confidence justification
- Medium confidence: the bare shape \b[12]\d{9}\b is a common 10-digit form and is only discovery-only without healthcare context. Even with NPI/provider keywords, the Luhn checksum is not enforced, so medium confidence is honest until a checksum-backed revision lands. Risk rating remains 3 (OFFICIAL) per v25 authoritative assessment of the structural detector.
- Detection quality
- Verified
- Jurisdictions
- us
- Regulations
- CCPA/CPRA, FTC Act s5, HIPAA, State Breach Laws (US)
- Frameworks
- ISO 27001, ISO 27701, SOC 2
- Data categories
- phi, healthcare
- Scope
- narrow
- Risk rating
- 3
- Platform compatibility
- Purview: Compatible, GCP DLP: Compatible, Macie: Compatible, Zscaler: Compatible, Palo Alto: Compatible, Netskope: Compatible
Pattern
\b[12]\d{9}\bCorroborative evidence keywords
MRN, medical record number, patient ID, NPI, DEA, medicare, medicaid, insurance ID, member ID, beneficiary, ICD-10, ICD-9, CPT, NDC, SNOMED, HCPCS, diagnosis code, procedure code, drug code, [object Object] (+2 more)
Proximity: 300 characters
Should match
1234567890— NPI starting with 12345678901— NPI starting with 21234567893— NPI-checksum-valid value with non-repeating digits
Should not match
3234567890— Starts with 3 (must start with 1 or 2)234567890— Starts with 0 (must start with 1 or 2)123456789— Only 9 digits instead of 101111111111— All-same-digit value rejected by the classifier's own AllDigitsSameFilter guard
Known false positives
- Medical terminology in health education materials, research publications, clinical guidelines, or public health documents without patient-specific data. Mitigation: Require corroborative evidence keywords confirming patient context. Look for co-occurrence with patient identifiers such as medical record numbers or dates of birth.
- General wellness and fitness content using medical vocabulary without constituting protected health information. Mitigation: Layer with patient identifier patterns or healthcare-specific document structure detection to distinguish clinical records from general health content.