SEER-Medicare Specialized Databases

Please be aware that in compliance with CMS, NCI will no longer release SEER-Medicare data outside the USA.

The SEER-Medicare data reflect the linkage of two large population-based sources of data that provide detailed information about Medicare beneficiaries with cancer. NCI created the SEER-Medicare Specialized Databases to make the information included in this data resource more accessible and less resource dependent. In addition to the SEER variables, these specialized databases include variables pertaining to Medicare enrollment, presence of comorbidities prior to cancer diagnosis, and cancer treatments (e.g., systemic, radiation therapy) identified from Medicare administrative enrollment and claims data. Refer to SEER-Medicare Summarized Data for more information.

Database Details

Three databases will be available to all approved requestors. The databases include SEER tumors from the November 2023 data submission diagnosed from 2000-2019 among persons who were also found in the Medicare data at some point from 1999-2020. They include the same fields as the SEER Research Plus databases other than the specialized fields.

Variable Definitions

The full list of variables is available in the SEER-Medicare Specialized Databases Data Dictionary (XLSX).

SEER-Medicare Enrollment

All persons included in the SEER-Medicare linkage are included in the databases to allow for assessments of representativeness. However, not all persons included will have sufficient Medicare data to calculate various attributes (e.g., comorbidity and treatment). Therefore, two non-mutually exclusive cohorts are identified in the databases:

The databases include variables to indicate cohort inclusion status, reason for cohort exclusion, reason for cohort exit (e.g., Medicare enrollment status changed, the person died, the person was diagnosed with a subsequent malignant cancer, or end of available data), number of months included in the cohort, and number of months included in Part D Low Income Subsidy (LIS) in the 4 months prior to and 12 months after diagnosis.

SEER-Medicare Comorbidity

The databases include yes/no indicators for 16 comorbidities, the Charlson comorbidity score, and the NCI comorbidity index, all of which result from running the 2021 version of the Comorbidity SAS Macro against the Medicare claims data. These “prior to cancer diagnosis” comorbidity measures are calculated among persons included in the ABFFS Cohort who had claims in the 12 months prior to the month of cancer diagnosis.

SEER-Medicare Systemic Therapy

The databases include yes/no indicators for the receipt of chemotherapy, hormone therapy, and immunotherapy during year 1, year 2, and year 3+ after diagnosis. These indicators are summarized separately for the Parts A/B claims (i.e., in-patient and out-patient data) and the Part D records (i.e., prescription drug data). For persons who received systemic therapy in year 3+, variables indicating months from diagnosis until first receipt within this year 3+ time window are provided. The availability of these variables depends on inclusion in the ABFFS or Part D cohorts (e.g., the Part A/B indictors will only be available for persons in the ABFFS cohort, the Part D indictors will only be available for persons in the Part D cohort, and only persons included in both cohorts will have both Part A/B and Part D indicators).

Systemic therapy is identified based on International Classification of Disease (ICD) diagnosis or procedure code, Current Procedural Terminology (CPT)/Healthcare Common Procedure Coding System (HCPCS) code, or National Drug Code (NDC) per Cancer Medications Enquiry Database (CanMED), administrative codes (HCPCS: 96400-96549, Q0083-Q0085) and demonstrative codes (HCPCS: G0355-G0363). Specific drugs are categorized based on the CanMED classification; ancillary therapies are not included.

SEER-Medicare Radiation Treatment

The databases include yes/no indicators for the receipt of radiation therapy during the 4 months prior to month of cancer diagnosis and year 1 and year 2+ after diagnosis. During each time period the type of radiation service is classified as proton/neutron, other beam, brachytherapy/radiation implant, radioisotopes, radiopharmaceuticals, and radiation, NOS. For persons who received radiation in year 2+, variables indicating months from diagnosis until first receipt within this year 2+ time window are provided. These variables are available for persons included in the ABFFS cohort.

Radiation therapy is identified based on International Classification of Disease (ICD) diagnosis or procedure codes, or Current Procedural Terminology (CPT)/ Healthcare Common Procedure Coding System (HCPCS) codes. Refer to the CoRe Cancer Treatment- Radiation File for more details.

Data Limitations and Analytical Considerations

The availability of comorbidity and treatment measures included in these databases depends on type of Medicare enrollment; therefore, there may be issues with generalizability of results both within the Medicare population and beyond to the general population. Currently, the comorbidity and treatment measures are largely available only among persons included in the ABFFS cohort. Inclusion in this cohort requires continuous enrollment in Parts A/B FFS and, therefore, most noticeably excludes Medicare Advantage beneficiaries, a growing (51% in 2023) segment of the Medicare population.

Caution should be taken when comparing receipt of treatment based on SEER data versus Medicare data. SEER data focus on initial treatment, whereas the Medicare data can capture longer treatment windows.

Users should acknowledge and discuss these limitations in any publications and employ sensitivity analyses to evaluate robustness of study findings results.

Publication Policy

Researchers must send a copy of any manuscript or book chapter via the SEER Specialized Data Request System prior to submitting it to a journal for review. Manuscripts should be uploaded to the documents section of your original data request.

The review is to ensure the confidentiality of individuals in the data. The review is not intended to assess the scientific merit of a manuscript. Researchers are not required to submit copies of meeting abstracts or power point presentations for review.