Medicare Fee-For-Service

Provider Utilization & Payment Data

Physician and Other Practitioners Dataset:
A Methodological Overview

August, 2021

Prepared by:
The Centers for Medicare and Medicaid Services,
Office of Enterprise Data and Analytics

1
Table of Contents
1. Background ........................................................................................................................................... 3
2. Key Data Sources................................................................................................................................... 3
3. Population ............................................................................................................................................. 4
4. Data Contents ....................................................................................................................................... 4
4.1. Detailed Data File ............................................................................................................................. 4
Medicare Physician and Other Practitioners by Provider and Service Dataset .................................... 4
4.2. Summary Tables ............................................................................................................................... 4
Medicare Physician and Other Practitioners by Provider Dataset ....................................................... 4
Medicare Physician and Other Practitioners by Geography and Service Dataset ................................ 4
5. Data Limitations: ................................................................................................................................... 5
6. Additional Information.......................................................................................................................... 7
APPENDIX A – Place of Service Code and Description .................................................................................. 8
APPENDIX B– Distribution of HCC Risk Scores .............................................................................................. 9

2
1. Background
The Centers for Medicare & Medicaid Services (CMS) has prepared a public data set, the Provider
Utilization and Payment Data Physician and Other Practitioners Dataset, with information on services
and procedures provided to Medicare beneficiaries by physicians and other healthcare professionals.
The Physician and Other Practitioners Dataset contains information on utilization, payment (allowed
amount and Medicare payment), and submitted charges organized by National Provider Identifier (NPI),
Healthcare Common Procedure Coding System (HCPCS) code, and place of service. The data in the
Physician and Other Practitioners dataset contains 100% final-action (i.e., all claim adjustments have
been resolved) physician/supplier Part B non-institutional line items for the Medicare fee-for-service
(FFS) population. Claims processed by Durable Medical Equipment, Prosthetic, Orthotics and Supplies
(DMEPOS) Medicare Administrative Contractor (MAC) are not included in the Physician and Other
Practitioners Dataset.

2. Key Data Sources

The data for the Physician and Other Practitioners Dataset are based upon CMS administrative claims
data for Medicare beneficiaries enrolled in the fee-for-service program. The data are available from the
CMS Chronic Condition Data Warehouse (CCW), a database with 100% of Medicare enrollment and fee-
for-service claims data. Service counts, beneficiary counts, provider charges, Medicare allowed amounts
and payments and the place of service indicator are summarized from Part B non-institutional claims
processed through Medicare Administrative Contractor (MAC) Jurisdictions (NCH Claim Type Codes '71',
'72'). Please see the CCW website for additional information. The prior years of the Physician and Other
Practitioners Dataset (CY2012/CY2013) are based upon data from the National Claims History (NCH)
Standard Analytic Files (SAFs), which are similar administrative data of 100% of Medicare final action
claims for beneficiaries who are enrolled in the FFS program. We compared the two data sources for
CY2013 and found that across all summary datasets the overall difference was .01% or less.

For all Physician and Other Practitioners Dataset data years, provider demographics (name, credentials,
gender, complete address and entity type) are included from the National Plan & Provider Enumeration
System (NPPES). CMS developed the NPPES to assign unique identifiers, known as National Provider
Identifiers (NPIs), to health care providers. The health care provider’s demographic information is
collected at the time of enrollment and updated periodically. The demographics information provided in
the Physician and Other Practitioners Dataset was extracted from NPPES at the end of calendar year
following the reporting year (e.g. for CY2017 reporting year, the NPPES data was extracted at the end of
calendar year 2018). Prior years of the Physician and Other Practitioners Dataset (CY2012/CY2013) are
based upon information extracted from NPPES at the end of calendar year 2014. Please visit the NPPES
website for additional information.

3
3. Population
The Physician and Other Practitioners Dataset includes data for providers that had a valid NPI and
submitted Medicare Part B non-institutional claims (excluding DMEPOS) during the reporting period. To
protect the privacy of Medicare beneficiaries, any aggregated records which are derived from 10 or fewer
beneficiaries are excluded from the Physician and Other Practitioners Dataset.

4. Data Contents

4.1. Detailed Data File

Medicare Physician and Other Practitioners by Provider and Service Dataset
The spending and utilization data in the Physician and Other Practitioners by Provider and Service
Dataset are aggregated to the following:

a) the NPI for the performing provider,

b) the Healthcare Common Procedure Coding System (HCPCS) code, and
c) the place of service (either facility or non-facility).

There can be multiple records for a given NPI based on the number of distinct HCPCS codes that were
billed and where the services were provided. Data have been aggregated based on the place of service
because separate fee schedules apply depending on whether the place of service submitted on the claim
is facility or non-facility.

4.2. Summary Tables

Two summary type tables have been created to supplement the information reported in the Physician
and Other Practitioners by Provider and Service Dataset: 1) aggregated information by physician or
other practitioner (NPI) and 2) aggregated information by Geography and Service and HCPCS code. The
aggregated reports are not restricted to the redacted data reported in the Physician and Other
Practitioners Dataset but are aggregated based on all Medicare Part B non-institutional claims (excluding
DMEPOS).

Medicare Physician and Other Practitioners by Provider Dataset

The “Medicare Physician and Other Practitioners by Provider Dataset” contains information on
utilization, payments (Medicare allowed amount, Medicare payment, and standardized Medicare
payment), and submitted charges organized by NPI. Sub-totals for medical type services and drug type
services are included as well as overall utilization, payment and charges. In addition, beneficiary
demographic and health characteristics are provided which include age, sex, race, Medicare and
Medicaid entitlement, chronic conditions and risk scores.

Medicare Physician and Other Practitioners by Geography and Service Dataset

The “Medicare Physician and Other Practitioners by Geography and Service Dataset” contains
information on utilization, payment (allowed amount and Medicare payment), and submitted charges
4
organized by HCPCS and place of service in the national table and organized by provider state, HCPCS
and place of service in the state table. The national and state tables also include a HCPCS drug indicator
to identify whether the HCPCS product/service is a drug as defined from the Medicare Part B Drug ASP
list.

5. Data Limitations:

Although the Physician and Other Practitioners Dataset has a wealth of payment and utilization
information about many Medicare Part B services, the dataset also has a number of limitations that are
worth noting.

First, the data in the Physician and Other Practitioners Dataset may not be representative of a
physician’s entire practice. The data in the file only have information for Medicare beneficiaries with
Part B FFS coverage, but physicians typically treat many other patients who do not have that form of
coverage. The Physician and Other Practitioners Dataset does not have any information on patients who
are not covered by Medicare, such as those with coverage from other federal programs (like the Federal
Employees Health Benefits Program or Tricare), those with private health insurance (such as an
individual policy or employer-sponsored coverage), or those who are uninsured. Even within Medicare,
the Physician and Other Practitioners Dataset does not include information for patients who are
enrolled in any form of Medicare Advantage plan.

The information presented in this file also does not indicate the quality of care provided by individual
physicians. The file only contains cost and utilization information, and for the reasons described in the
preceding paragraph, the volume of procedures presented may not be fully inclusive of all procedures
performed by the provider.

Medicare allowed amounts and Medicare payments for a given HCPCS code/place of service can vary
based on a number of factors, including modifiers, geography, and other services performed during the
same day/visit. For example, modifiers (two-character designators that signal a change in how the
HCPCS code for the procedure or service should be applied) may be included on the claim line when the
service intensity was increased or decreased, when an additional physician administered services, or
when the service provided differs from the procedure definition. In some cases, modifiers impact
allowed amounts and payments. In addition, allowed amounts and payments vary geographically
because Medicare makes adjustments for most services based on an area's cost of living. Allowed
amounts and payments can also be adjusted when a physician renders multiple services to a beneficiary
on the same day, which is referred to as a multiple procedure payment reduction. For standard
payment and allowed amount rates by CPT/HCPCS code, please visit the Physician-fee-schedule.

In general, when a provider administers drugs to a patient, the provider purchases the drug and
Medicare pays the provider 106% of the average sales price (ASP) for the drug. Although the ASP list
was used in these datasets to define drug services, the drugs listed on the ASP fee schedule are not a
complete listing of drugs paid under part B, but the ASP fee schedule represents the majority of drugs

5
that are used in the office. For more information on payments for drugs covered under Part B, please
visit ASP Drug Pricing.

Additionally, the data are not risk adjusted and thus do not account for difference in the underlying
severity of disease of patient populations treated by providers. However, we have provided average
beneficiary risk scores in the “Medicare Physician and Other Practitioners by Provider Dataset” (i.e., one
record per NPI) to provide information on the health status of the beneficiaries the providers serve.
Also, since the data presented are summarized from actual claims received from providers and no
attempts were made to modify any data (i.e., no statistical outliers were removed or truncated), in rare
instances the average submitted charge amount may reflect errors included on claims submitted by
providers.

As noted earlier, the file does not include data for services that were performed on 10 or fewer
beneficiaries, so users should be aware that summing the data in the file may underestimate the true
Part B FFS totals. In addition, some providers bill under both an individual NPI and an organizational
NPI. In this case, users cannot determine a provider’s actual total because there is no way to identify
the individual’s portion when billed under their organization.

Medicare pays differently when services are provided in a facility setting versus a freestanding
physicians’ office (or other non-facility setting). When services are delivered in a facility setting,
Medicare makes two payments, one for the physician’s professional fee and one for the facility. For
services delivered in a facility (Place_Of_Srvc=”F”), the data in the Physician and Other Practitioners
Dataset generally represents the physician’s professional fee and does not include the facility payment.
The exception is services delivered in Ambulatory Surgical Centers (ASCs). In these cases, both the
physician’s professional fee and the ASC’s fee are represented in the Physician and Other Practitioners
Dataset. ASCs can be identified using the provider type. For services delivered in a non-facility setting,
such as a physician’s office (Place_Of_Srvc=”O”), the Physician and Other Practitioners Dataset
represents the complete payment for the service.

If users try to link data from this file to other public datasets, please be aware of the particular Medicare
populations included and timeframes used in each file that will be merged. For example, efforts to link
the Physician and Other Practitioners Dataset data to Part D prescription drug data would need to
account for the fact that some beneficiaries who have FFS Part B coverage (and are thus included in the
Physician and Other Practitioners Dataset) do not have Part D drug coverage (and thus not represented
in Part D data files). At the same time, some beneficiaries that have Part D coverage (and are thus
included in the Part D data) do not have FFS Part B coverage (and thus not included in the Physician and
Other Practitioners Dataset). Another example would be linking to data constructed from different or
non-aligning time periods, such as publically available data on physician referral patterns, which is based
on an 18-month period.

Finally, users should be aware that payments from some CMS demonstration programs are included in
the Physician and Other Practitioners Dataset. Since some CMS demonstration programs utilize the
Medicare claims submission process, payments for services under these demonstrations are included in
the data file and may be grouped under specific demonstration HCPCS codes or aggregated under non-
6
demonstration specific HCPCS codes. Demonstration programs that are paid outside of the Medicare
claims submission process are not included in the Physician and Other Practitioners Dataset.

6. Additional Information
Other Data Sources: CMS also releases the “Medicare Fee-For-Service Public Provider Enrollment Data”
that include provider name and address information from the Provider Enrollment and Chain Ownership
System (PECOS). These data are updated on a quarterly basis and are available at data.cms.gov.

Medicare Standardized Spending: Users can find more information on Medicare payment
standardization by referring to the “Geographic Variation Public Use File: Technical Supplement on
Standardization” available within the “Related Links” section of the following web page: Medicare
Geographic Variation.

HCCs (hierarchical condition categories): CMS developed a risk-adjustment model that uses HCCs
(hierarchical condition categories) to assign risk scores. Those scores estimate how beneficiaries’ FFS
spending will compare to the overall average for the entire Medicare population. The average risk scores
of beneficiaries represented in each calendar year of the Physician and Other Practitioners Dataset data
are provided in Appendix A. Beneficiaries with scores greater than the average risk score are expected to
have above-average spending, and vice versa. Risk scores are based on a beneficiary’s age and sex;
whether the beneficiary is eligible for Medicaid, first qualified for Medicare on the basis of disability, or
lives in an institution (usually a nursing home); and the beneficiary’s diagnoses from the previous year.

The HCC model was designed for risk adjustment on larger populations, such as the enrollees in an MA
plan, and generates more accurate results when used to compare groups of beneficiaries rather than
individuals. Please visit HCC risk score for more information.

7
APPENDIX A – Place of Service Code and Description

Place of Service Code Facility Place of Service Description Facility Type

01 Pharmacy O
03 School O
04 Homeless Shelter O
05 Indian Health Service Free-standing Facility O
06 Indian Health Service Provider-based Facility O
07 Tribal 638 Free-standing Facility O
08 Tribal 638 Provider-based Facility O
09 Prison/ Correctional Facility O
11 Office O
12 Home O
13 Assisted Living Facility O
14 Group Home O
15 Mobile Unit O
16 Temporary Lodging O
17 Walk-in Retail Health Clinic O
20 Urgent Care Facility O
25 Birthing Center O
32 Nursing Facility O
33 Custodial Care Facility O
49 Independent Clinic O
50 Federally Qualified Health Center O
54 Intermediate Care Facility/Mentally Retarded O
55 Residential Substance Abuse Treatment Facility O
60 Mass Immunization Center O
57 Non-residential Substance Abuse Treatment Facility O
62 Comprehensive Outpatient Rehabilitation Facility O
65 End-Stage Renal Disease Treatment Facility O
71 Public Health Clinic O
72 Rural Health Clinic O
81 Independent Laboratory O
99 Other Place of Service O
21 Inpatient Hospital F
22 Outpatient Hospital F
23 Emergency Room – Hospital F
24 Ambulatory Surgical Center F
26 Military Treatment Facility F
31 Skilled Nursing Facility F
34 Hospice F
41 Ambulance - Land F
42 Ambulance – Air or Water F
51 Inpatient Psychiatric Facility F
52 Psychiatric Facility-Partial Hospitalization F
53 Community Mental Health Center F
56 Psychiatric Residential Treatment Center F
61 Comprehensive Inpatient Rehabilitation Facility F

8
 APPENDIX B– Distribution of HCC Risk Scores

Number of Average
Calendar Minimum Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile Maximum
Medicare Risk
Year Risk Score 01 05 10 25 50 75 90 95 99 Risk Score
Beneficiaries Score

2012 32,680,448 0.107 0.223 0.268 0.329 0.474 0.751 1.284 2.272 3.190 7.611 58.178 1.149

2013 32,947,265 0.110 0.256 0.266 0.317 0.473 0.748 1.290 2.296 3.251 6.923 47.700 1.136

2014 33,120,069 0.111 0.245 0.264 0.319 0.470 0.740 1.256 2.218 3.197 6.930 46.735 1.116

2015 33,170,347 0.114 0.260 0.272 0.328 0.486 0.769 1.333 2.370 3.380 7.200 48.607 1.174

2016 33,580,752 0.116 0.251 0.274 0.332 0.486 0.774 1.336 2.383 3.454 7.323 47.850 1.182

2017 33,461,183 0.146 0.266 0.295 0.354 0.486 0.761 1.326 2.375 3.459 7.431 48.126 1.185

2018 33,444,902 0.143 0.260 0.289 0.346 0.475 0.761 1.328 2.398 3.484 7.404 47.131 1.185

2019 33,042,229 0.138 0.271 0.284 0.340 0.467 0.759 1.329 2.424 3.508 7.592 51.923 1.191