SEER-Medicare Provider Files (2024)

Researchers may want to include information in their analyses abouthealthcare providers (physicians or hospitals) who are treating cancer patients.Medicare claims include unique provider numbers that identify the hospitals orphysicians associated with the claim. Provider numbers on the SEER-Medicare dataare encrypted to protect the confidentiality of the provider but the encryptionis done in a manner that allows tracking of each provider across years.

Hospital Data

NCI Hospital File

NCI has created a HospitalFile (PDF, 345 KB) that includes information about hospitals that are part of the SEER-Medicare data. Information about these hospitals has been obtained from several sources. These sources include the Healthcare Cost Report(HCRIS) and the Provider of Service (POS) survey from the Center for Medicare and Medicaid Services (CMS). The HCRIS data is created annually from reporting that CMS requires of hospitals that bill to the Medicare and Medicaid programs. CMS also requires institutions to periodically complete the Provider of Service (POS) survey. NCI has extracted selected variables from the HCRIS and POS files from 1996, 1998, and 2000-2016 to be included in the Hospital File. It should be noted that there are differences between the HCRIS and POS files as to exact time periods and how selected variables are defined. NCI has not attempted to resolve any inconsistencies between files, but is making researchers aware of these differences. Close examination of the data may be needed. In addition to information from CMS, the NCI Hospital File includes information about whether the institution was an NCI-designated cancer center as of 2002, 2005, 2010 and 2017 or if the Hospital is a Commission on Cancer Accredited Hospital. There are also historic variables denoting if the hospital participated in any NCI-sponsored cooperative group trial in 2002 or 2005. The NCI Hospital file includes encrypted hospital numbers making it possible to link the NCI Hospital file to the hospital numbers found on the SEER-Medicare claims.

The hospital file will be made available without charge to investigators. Investigators must include a description of the use of the file in their proposal requesting SEER-Medicare data.

Documentation for theHospital file (PDF, 345 KB) is available in PDF format.

Hospital Mergers & Acquisitions

When a hospital undergoes a merger, acquisition or closure, the associatedMedicare hospital provider number often changes, creating challenges fortracking providers over time. The NCI Hospital file itself contains nodistinguishing identifier for researchers to link information on current andpast hospital provider numbers for a facility. For longitudinally tracking ofhospitals in the SEER-Medicare data, researchers are encouraged to use the Mergers and Acquisitionsfile (PDF, 96 KB) maintained by NCI.

The Mergers and Acquisitions file was created from the Provider of Service(POS) files released by the Centers for Medicare & Medicaid Services (CMS),which collect information on previous or associated provider numbers for ahospital. The Mergers and Acquisition file contains one observation per providernumber (current or previous). Along with each provider number, the file includesa linking variable (Group_id) that assigns a unique identifier to all associatedprovider numbers identified across all years of the hospital file, if any. If ahospital has no associated provider numbers, the Group_id variable will have amissing value for that provider. All provider numbers are encrypted in the sameway as other files in the SEER-Medicare data to allow for linkage.

The Mergers and Acquisitions file will be made available without charge toinvestigators. Investigators must include a description of the use of the filein their proposal requesting SEER-Medicare data.

Documentation forthe Mergers and Acquisitions file (PDF, 96 KB) is available in PDF format.

American Hospital Association (AHA) Data

In addition to the NCI Hospital File, information about hospitals can be obtained from the American Hospital Association (AHA) hospital survey. Linking to the AHA data requires unencrypted hospital numbers from the Medicare data. Details on how to obtain unencrypted hospital numbers can be found at SEER-Medicare Encrypted or Restricted Variables.

Individual Providers

Upon request with the SEER-Medicare data, supplemental information on individual providers is available from the Medicare Data on Provider Practice and Specialty (MD-PPAS) file and the American Medical Association (AMA), see below. Requestors will need to determine which data source is most useful for their research proposals. Please refer to the below articles for insights into the strengths and limitations of these two data sources.

• Warren JL, Barrett MJ, White DP, Banks R, Cafardi S, Enewold L. Sensitivity of Medicare data to identify oncologists. J Natl Cancer Inst Monogr. 2020;2020(55): 60–65.
• White DP, Enewold L, Geiger AM, Banks R, Warren JL. Comparison of physician data in two data files available for cancer health services research. J Natl Cancer Inst Monogr. 2020;2020(55):66–71.

MD-PPAS (physicians and other providers)

The Medicare Data on Provider Practice and Specialty (MD-PPAS) uses data from the Provider Enrollment, Chain, and Ownership System (PECOS), the National Plan and Provider Enumeration System (NPPES), and 100% Medicare claims for Part B services to enhance or develop measures concerning provider demographics (sex, birthdate), specialty, place of service, geography, and utilization. The MD-PPAS assigns Medicare providers to medical practices using the national provider identifier (NPI) and the tax identification numbers (TIN). Utilization measures include the number of part B non-institutional line items, total Medicare allowed charges, the number of unique patients, and the percentage of a provider’s line items that are performed in eight place-of-service categories. The MD-PPAS contains a record for any individual provider that had a valid NPI and submitted a Part B non-institutional claim for evaluation and management services, procedures, imaging, or non-laboratory testing with a positive allowed charges amount.

Documentation MD-PPAS V2.3 file (2008-2017) (PDF, 144 KB) is available in PDF format.
Documentation MD-PPAS V2.4 file (2018-2020) (PDF, 195 KB) is available in PDF format.

Physician-specific data

American Medical Association (AMA) via Physician identifiers

Linking to the AMA data requires unencrypted physician numbers from the Medicare data, which are not released. Details on how to link to AMA data can be found at SEER-Medicare Encrypted Variables.

NPI-UPIN Physician Crosswalk File for the SEER-Medicare Data

In 2007, CMS adopted a new physician identification system, changing from the Unique Physician Identification Number (UPIN) to the National Provider Identifier (NPI). In order to track physicians in the SEER-Medicare data across time, a group of researchers at and associated with NCI created a file to crosswalk a physician's NPI to his/her UPIN.
Parsons HM, Enewold LR, Banks R, Barrett MJ, Warren JL. Creating a National Provider Identifier (NPI) to Unique Physician Identification Number (UPIN) Crosswalk for Medicare Data. Med Care. 2017 Dec;55(12):e113-e119.

To create the crosswalk, all NPIs on the SEER-Medicare physician (carrier)claims in 2007 – 2008 were identified. During this period, physicians wererequired to submit claims to Medicare using their NPI; however many claims alsoincluded a UPIN for the physician in addition to the NPI. Using information fromthe claims, the NPI and UPIN Directories and physician name matching, NPIs werematched to UPINs with an overall 90.1% match rate. Of the unique 470,313 NPIsidentified in the claims, 423,912 were assigned a unique UPIN.

All physician identifiers on the SEER-Medicare data are encrypted in order toprotect the privacy of the physicians. The same encryption is used for theNPI-UPIN crosswalk file. This will allow researchers to link the NPI-UPINcrosswalk file to the SEER-Medicare claims. This crosswalk file is madeavailable without charge to SEER-Medicare researchers who need to trackphysicians over the time period when NPIs were implemented. Please contactSEER-Medicare for more details of how thecrosswalk was created or to obtain a copy of the crosswalk file.