NPI vs. CMS data: what your physician referral list really tells you
Use NPI records to identify potential providers and Medicare evidence to add context. Learn why missing claims do not mean no opportunity—and why volume is not a referral relationship.
November 14, 20263 min readRepVector
A physician referral list becomes more useful when each field has a clear meaning. A provider name, a specialty and a drug claim count answer different questions. Combining them can improve preparation, but combining them carelessly can create false confidence.
NPI records identify providers, not just Medicare participants
The public NPI Registry contains active National Provider Identifier records, including publicly relevant names, taxonomies and practice addresses. It is not a Medicare-only or Medicaid-only directory. CMS also explicitly warns that issuance of an NPI does not validate licensing or credentials. See the official NPI Registry.
For outreach planning, use the NPI as a stable identity anchor. A name alone is not enough: different clinicians can share a name, and one clinician can have changing addresses. Confirm the office location and appropriate contact before scheduling a visit. Do not treat a taxonomy entry as proof that the office offers your target treatment today.
CMS drug records add historical Medicare context
The Medicare Part D prescriber datasets describe prescription drug events for Medicare Part D beneficiaries and organize the information by NPI and drug. CMS documents the reporting population, limitations and suppression rules in its Part D methodological overview.
That means a row can support a statement about published Medicare activity in the disclosed year. It cannot support a statement about all patients, every payer or current prescribing behavior. An annual dataset is not a live view of an office.
In RepVector, stored import coverage and live lookup limits add another boundary. An absent result may reflect source suppression, no matching published evidence or incomplete application coverage. It is not automatically a zero.
Referral-fit and drug scores answer different questions
A referral-fit score organizes candidates using available matching signals such as provider type and location, with supporting evidence where available. It does not establish that the provider already refers to your practice.
A drug score is narrower: it compares real published drug or procedure evidence within the same search-area cohort. RepVector withholds that score when the evidence or minimum cohort is insufficient. Do not compare a drug score across different cities as though the reference group were identical.
For a request such as “Austin gynecologists relevant to antidepressant care,” a nearby specialty match can still be worth reviewing when Medicare drug evidence is missing. The honest result is “specialty and location match; relevant published prescribing evidence unavailable,” not “this physician does not prescribe.”
Keep these fields with your export
- NPI and provider name for identity.
- Specialty and available practice contact information.
- Search geography and any disclosed radius.
- Score type and available match reasons.
- Source and reporting year for measured activity.
- Coverage notes and missing evidence.
Keeping provenance makes the list easier to review later. It also helps avoid turning an old claims period into a statement about the present.
What still requires verification
Public provider data does not prove existing referral relationships, employment, ownership, treatment availability or patient outcomes. Contact information may be out of date. License checks and compliance review belong in the appropriate official and professional processes.
RepVector supplies provider targeting and evidence, not patient-level referral tracking. Use your own authorized systems to determine which outreach actually produced referrals.
Read the physician referral marketing guide, find drug prescribers, or build a specialty-and-location list.
Filed under Physician referral marketing · Provider data
