npi_lookup
ClaudeGet comprehensive provider details by NPI number from the CMS NPPES Registry. WHEN TO USE: - You have a specific NPI and need full provider information - Verifying provider credentials, specialty, or practice location - Following up on results from npi_search to get complete details - Cross-referencing clinical trial investigators with US credentials - Getting raw API response for advanced data extraction USE npi_search INSTEAD FOR: - Finding providers by name, location, or specialty - When you don't have a specific NPI number - Browsing providers in a geographic area WHAT THIS RETURNS: Basic Information: - Provider type: Individual (NPI-1) or Organization (NPI-2) - Name: Full name (individuals) or Legal Business Name (organizations) - Credentials: MD, DO, RN, PA-C, NP, etc. (for individuals) - Status: 'A' (Active) or 'D' (Deactivated) - Enumeration date: When the NPI was first assigned - Last updated: When the record was last modified Specialty/Taxonomy: - Primary taxonomy code: NUCC healthcare provider taxonomy code - Primary taxonomy description: Human-readable specialty name - License state and number: From primary taxonomy Location: - Practice address: Primary practice location with full address - Phone number: Practice telephone (if available) Advanced: - Raw API response: Complete NPPES data including: - All taxonomies (up to 15) - Other identifiers (state license, DEA, etc.) - Other practice locations - Authorized officials (for organizations) - Endpoints (health information exchange) NOTE: Automatically validates NPI format before API call. Returns found=false if NPI is valid but not assigned to any provider. IMPORTANT CAVEATS: - Having an NPI does NOT guarantee the provider is currently licensed - Provider self-reports data; NPPES does not verify accuracy - Deactivated providers remain in database but should not be used for billing - Verify credentials through state licensing boards for official confirmation
npi_search
ClaudeSearch the CMS NPPES Registry for healthcare providers by name, location, specialty, or organization. WHEN TO USE: - Finding providers by name (e.g., 'Find Dr. John Smith') - Finding providers by specialty (e.g., 'Cardiologists in Boston') - Finding organizations (e.g., 'Mayo Clinic locations') - Verifying clinical trial investigators have US credentials - Building provider networks by location or specialty - Finding all providers at a specific facility USE npi_lookup INSTEAD FOR: - When you already have a specific NPI number - Getting full details including raw API response SEARCH REQUIREMENTS: - At least one search parameter required (besides enumeration_type/country_code) - State cannot be the only criterion - must combine with name, specialty, or city - Country code alone only works for non-US countries WILDCARD SUPPORT (min 2 characters before ): - first_name: 'Jo' matches John, Joseph, Jonathan - last_name: 'Smi' matches Smith, Smithson, Smitty - organization_name: 'Mayo' matches Mayo Clinic, Mayo Foundation - taxonomy_description: 'Cardio' matches Cardiology, Cardiovascular Disease - postal_code: '212' matches all ZIP codes starting with 212 NAME ALIAS FEATURE (DEFAULT: ON): Automatically expands first name searches to include common nicknames: - Robert -> also finds Bob, Rob, Robbie, Bobby - William -> also finds Bill, Will, Billy, Willy - Elizabeth -> also finds Beth, Liz, Betty, Eliza - Michael -> also finds Mike, Mick, Mickey Set use_first_name_alias=false for exact match only. PAGINATION: - Maximum 200 results per request (limit parameter) - Skip up to 1000 records (skip parameter) - Maximum 1,200 total results accessible (6 requests × 200) - Use next_skip from response to get next page - Set limit=200 for comprehensive searches COMMON SPECIALTY SEARCHES (taxonomy_description): - Physicians: 'Internal Medicine', 'Family Practice', 'Cardiology', 'Oncology' - Surgeons: 'Orthopedic Surgery', 'General Surgery', 'Neurosurgery' - Mental Health: 'Psychiatry', 'Psychology', 'Clinical Social Worker' - Allied Health: 'Physical Therapist', 'Nurse Practitioner', 'Physician Assistant' - Organizations: 'General Acute Care Hospital', 'Pharmacy', 'Clinical Laboratory' EXAMPLES: - Find cardiologists in Boston: taxonomy_description='Cardiology', city='Boston', state='MA' - Find Dr. Smith (individual): first_name='John', last_name='Smith', enumeration_type='NPI-1' - Find Mayo Clinic (organization): organization_name='Mayo Clinic', enumeration_type='NPI-2' - Find oncologists in California: taxonomy_description='Oncology', state='CA', limit=100 - Find all pharmacies in NYC: taxonomy_description='Pharmacy', city='New York', state='NY' - Find nurse practitioners in Texas: taxonomy_description='Nurse Practitioner', state='TX' - Exact name match: first_name='Robert', use_first_name_alias=false
npi_validate
ClaudeValidate NPI format and Luhn check digit. Instant local validation - no API call. WHEN TO USE: - ALWAYS use before npi_lookup to catch typos and save API calls - Validating NPI numbers from external sources (documents, forms, databases) - Quick format check without network dependency - Pre-filtering lists of NPIs before bulk lookups WHAT THIS CHECKS: 1. Format: Exactly 10 digits, numeric characters only (0-9) 2. Check digit: Valid Luhn algorithm calculation with healthcare prefix (80840) WHAT THIS DOES NOT CHECK: - Whether the NPI is actually assigned to a provider in NPPES - Whether the provider is active or deactivated - Whether the provider is licensed or credentialed - Whether the NPI has ever been issued VALIDATION RESULTS: ✓ VALID: NPI has correct format AND valid check digit -> Proceed to npi_lookup to verify assignment ✗ INVALID FORMAT: Not 10 digits or contains non-numeric characters -> Cannot be a real NPI, check for typos ✗ INVALID CHECK DIGIT: 10 digits but check digit doesn't match -> Likely a typo in one of the digits IMPORTANT: A valid check digit does NOT guarantee the NPI exists! Many mathematically valid NPIs have never been assigned. Use npi_lookup after validation to confirm the NPI is in NPPES. EXAMPLES: - '1234567893' -> VALID (correct format + valid check digit) - '123456789' -> INVALID (only 9 digits) - '1234567890' -> INVALID (wrong check digit, expected 3) - '123456789X' -> INVALID (contains non-digit character) - ' 1234567893 ' -> VALID (whitespace is stripped automatically)
