SEER-Medicare: Selecting the Appropriate Comorbidity SAS Macro

SAS Macros to calculate comorbidity weights are provided to assist SEER-Medicare investigators with their analyses. Researchers may be unsure which Comorbidity SAS Macro is appropriate for their study. The table below provides an overview of what is included with each macro. NCI recommends use of the 2014 macro. The 2000 version is provided for those needing to reproduce prior results. Please note that both the 2000 and 2014 macros offer the option of including Medicare hospital (MedPAR) claims only, or also considering physician (NCH) and outpatient claims. Also note that diagnoses using ICD-10 codes, starting in October 2015, are not yet included in these macros. NCI does not accept responsibility for the completeness or accuracy of the codes and weights used in the macros. Investigators may modify the macros if they wish to include different diagnosis codes or condition weights.

Comorbidity SAS Macro Comorbidity Index Included Diagnosis Codes Rule-Out Option* Weights
2014 version NCI Comorbidity Index 2014

Diagnosis codes identified by professional coders in 2014 (HCPCS codes have been removed).

Recommended Weights in macro are for all cancers developed by NCI statisticians in 2014. Weights for cancer-specific sites (breast, lung, prostate, and colorectal cancers) are available in the technical report for researchers who want to customize the macro.
2014 version Charlson Index 2014

Diagnosis codes identified by professional coders in 2014 (HCPCS codes have been removed)

Optional Charlson Index weights
(excluding cancer)
2000 version Charlson Index 2000

Diagnosis and HCPCS codes identified by Deyo and Romano

Optional Charlson Index weights (excluding cancer)

*Rule-out option includes conditions reported twice or more in physician or outpatient claims at least 30 days apart.