Registry-derived Files

• SEER coding and staging manuals
• Cancer and 5% Cancer File: The SEER-Medicare 5% non-cancer sample is not a true non-cancer sample. The NCI generates the 5% non-cancer sample from data pulled from SEER regions by eliminating patients who have SEER records for these cancers; the NCI does not consider the content of Medicare claims. Statisticians should further cull patients using claims-based CCW chronic condition flags and diagnostic codes to create a true non-cancer sample.
  • Data dictionary
  • 5% Cancer FAQ
• MD-PPAS
• Hospital Characteristics File
• Census Tract Characteristics File
• ZIP Code Characteristics File

Medicare-derived Files

Starting with the 2020 data, SEER-Medicare’s MedPAR, Carrier (NCH), Outpatient, Home Health Agency (HHA), Hospice, Durable Medical Equipment, Master Beneficiary Summary (MBSF), Part D files, Minimum Data Set (MDS) - Nursing Home Assessment, and Home Health Outcome and Assessment Information Set (OASIS) utilize the same record layouts and codebooks as the data available from CMS CCW. The files are incompatible with data obtained prior to 2020.

• Data documentation for Medicare files obtained 2020 and later
• Data documentation and SAS input for files obtained prior to 2020