Data on the people delivering care, not just the places.
Most healthcare datasets start with a facility — a hospital, a nursing home, a dialysis center. The CMS Doctors and Clinicians dataset starts with an individual. Every clinician with an active Medicare enrollment is included, each carrying a stable unique identifier — the National Provider Identifier (NPI) — that follows them regardless of where or with whom they practice. This is the most granular level at which CMS publicly reports clinical performance.
A "clinician" in this dataset is not limited to physicians. The file spans hundreds of specialty and credential types — anyone billing Medicare for clinical services. That includes medical doctors and osteopathic physicians, but also nurse practitioners, physician assistants, certified registered nurse anesthetists, physical and occupational therapists, clinical social workers, podiatrists, and many others.
The identifiers that follow a clinician across every setting.
Every clinician in the dataset is identified by a consistent set of IDs that appear across all Doctors and Clinicians files. Each one is scoped differently — to the individual, the enrollment record, or the group — which is what makes it possible to link a clinician's demographics, affiliations, procedure volumes, and quality scores into a single picture.
Four distinct files, one complete picture of a clinician.
The Doctors and Clinicians dataset covers four distinct dimensions: who the clinician is, where they work, what procedures they perform, and how they perform against Medicare quality standards. Each dimension is reported separately, but all four link together through a common identifier — the NPI.
Because the demographic file is organized at the clinician / enrollment / group / address level, a single clinician working across multiple locations or organizations will appear on multiple rows. The identifier fields — particularly the NPI — are what tie those rows back to one person.
- NPI, individual PAC ID, and enrollment ID
- Name, gender, credential (MD, DO, NP, PA, etc.)
- Medical school and graduation year
- Primary specialty + up to four secondary specialties
- Group name, Group PAC ID, and number of group members
- Practice address (line 1, line 2, city, state, ZIP, phone)
- Telehealth indicator; Medicare assignment acceptance
- Facility type (hospital, long-term care hospital, nursing home, IRF, home health agency, hospice, dialysis)
- Facility Affiliations Certification Number — the CCN of the unit or facility
- Facility Type Certification Number — the primary hospital CCN, when applicable
- NPI and individual PAC ID linking back to the clinician's demographic record
- Procedure category name and count
- Percentile rank within the national distribution
- Profile display indicator (Y/N)
- Counts below 11 reported as "1–10"
- Quality, Promoting Interoperability, Improvement Activities, and Cost category scores
- MIPS Final Score and MIPS Final Score without Complex Patient Bonus
- Score source: individual, group, virtual group, or APM
- Facility-based certification number, when applicable
Twelve categories that show a clinician's procedural experience.
Procedure volume data is one of the most practically useful elements of the clinician dataset — it answers a question patients actually ask before making decisions: how often does this clinician perform this specific procedure? CMS tracks procedure volume across 12 standardized categories, making counts comparable across clinicians nationwide.1
The methodology applies meaningful constraints: modifier codes that indicate a limited or supporting role in a procedure are excluded, so a count reflects procedures in which the clinician was a primary performer, not an assistant or observer.
Where quality scores become payment adjustments.
The Quality Payment Program is Medicare's framework for tying clinician payment to quality and value. Most clinicians participate through the Merit-Based Incentive Payment System, or MIPS — a composite scoring model that evaluates performance across four categories and applies the result as a positive or negative adjustment to a clinician's Medicare reimbursement rate in a future payment year.
MIPS scores can be reported at three levels: individual clinician, group, or virtual group. A clinician's final score reflects whichever reporting pathway applied to their participation that year.
- Clinician selects from hundreds of quality measures
- Reported via claims, EHR, registry, or web attestation
- Performance rate compared to a national benchmark
- Star rating assigned at the measure level
- Both individual and group-level scores published
- Measures use of certified EHR technology
- Tracks e-prescribing, health information exchange, and patient access to records
- Reported via web attestation
- Certain specialties may apply for hardship exception
- Clinician attests to completing practice improvement activities
- Activities span care coordination, patient safety, population management, and expanded care access
- Reported via web attestation (Y/N)
- Bonus credit available for high-priority activities
- Calculated directly from Medicare claims — no clinician submission required
- Measures total per capita costs and episode-based cost measures
- Risk-adjusted for patient complexity
- Scores assigned where clinician has sufficient case volume
MIPS performance information is publicly reported at three levels, each with its own dataset in the Provider Data Catalog. Individual clinician data is the most granular and is directly linked to NPI-level identity records. Group and virtual group data allow analysis of practice-wide performance patterns. Clinicians in Advanced Alternative Payment Models (APMs) may be exempt from traditional MIPS scoring entirely and instead receive separate incentive payments.
Where the data comes from and what shapes it.
The Doctors and Clinicians dataset is built entirely from Medicare data — licensed sources, board certification records, and residency training information are explicitly excluded.2 Procedure volume draws from Medicare fee-for-service claims and Medicare Advantage encounters, using a one-year rolling window with a three-month lag for claims processing.1
MIPS performance data reflects the performance year for which it was collected. Not every clinician with a MIPS Final Score will have a profile or appear in the demographic dataset — some were assigned scores but did not meet the minimum requirements for public reporting. The small cell size policy also applies throughout: when procedure counts fall below 11, they are suppressed to a range rather than reported exactly — a constraint that affects smaller-volume specialties and rural practitioners disproportionately.2
- CMS Provider Data Catalog: Doctors and Clinicians Data Dictionary, Performance Year 2023 — data.cms.gov
- CMS Provider Data Catalog: Additional Information, Doctors and Clinicians — data.cms.gov
- Quality Payment Program: 2023 MIPS Overview — qpp.cms.gov
- CMS Medicare Care Compare: Doctors and Clinicians — medicare.gov/care-compare