Table 2.

Description of SEER-Medicare Data Files, 1973–2007

File Name	Description	Years of Data Availablea	Comment
PEDSF	SEER cancer cases	1973–2005	One record per person, with up to 10 cancers.
			File includes a flag indicating which SEER cases were in the 5% random sample.
			File includes selected socioeconomic information from the US Census based on census tract.
SUMDENOM	5% random sample of Medicare recipients living in SEER areas, excluding SEER cancer cases	1986–2005	One record per person.
			File includes selected socioeconomic information from the US Census based on census tract.
MEDPARb	Medicare hospital claims for all Part A short stay, long stay, and skilled nursing facility care	1986–2007	One record per hospital stay.
			Each record has up to 10 diagnoses and 6 procedures coded by using the ICD-9.
Carrier claims (NCH)b	Medicare claims from physicians and other noninstitutional medical care providers	1991–2007	Multiple records for the same service from different providers.
			Each record can contain multiple services coded by using the HCPCS.
			Each service has a corresponding ICD-9 diagnosis.
OUTPTb	Medicare claims from institutional outpatient providers	1991–2007	Multiple records per visit are based on revenue center codes.
			Each record must have an HCPCS code.
			ICD-9 diagnosis and procedure codes may be reported but are not required.

Abbreviations: HCPCS, Healthcare Common Procedure Coding System; ICD-9, International Classification of Diseases, Ninth Revision; MEDPAR, medical provider analysis and review; NCH, national claims history; OUTPT, outpatient; PEDSF, patient entitlement and diagnosis summary file; SEER, Surveillance Epidemiology, and End Results; SUMDENOM, summarized denominator.

^aThe heading, “years of data available,” refers to the data files provided from the 2008 linkage.

^bFor cancer cases in PEDSF from 2003 to 2005, Medicare claims data are not available before 1998.