Background: Measures to identify comorbidities in administr ative data were developed using elderly or hospitalized patients. We compared the prevalence of five medical comorbidities using different definitions in women undergoing mastectomy compared to national estimates.
Methods: Using a cohort of 7,612 privately insured women aged 18–64 years coded for mastectomy from 1/04–12/08, we identified smoking, obesity, diabetes, hypertension, and iron deficiency anemia using inpatient and outpatient claims for the year prior to surgery. Our revised comorbidity algorithm included outpatient medications for diabetes and hypertension, expanded the timeframe to include the mastectomy admission, and adjusted the interval and number (≥ 1 claims for smoking and obesity) of required outpatient claims. A χ2 test of proportions compared prevalence rates in the cohort to 18–64 year old women in the 2007 Behavioral Risk Factor Surveillance System and the 2000 National Health and Nutrition Examination Survey.
Results: Compared to the standard claims algorithm, the prevalence increased from 0.6% to 6.6% for smoking, 0.6% to 3.6% for obesity, 3.4% to 6.1% for diabetes, 10.7% to 21.4% for hypertension, and 2.9% to 5.0% for iron deficiency anemia. The revised prevalence estimates were more similar to national estimates, however, with the exception of diabetes, were not statistically equivalent (17.1% smoking, 24.7% obesity, 5.9% diabetes, 19.7% hypertension, and 3.0% iron deficiency anemia).
Conclusions: Our revised comorbidity algorithm resulted in prevalence estimates that were closer to expected rates for non-elderly women than the standard medical claims algorithm. The prevalence of chronic medical conditions (i.e., hypertension, diabetes, iron deficiency anemia) was more similar to national estimates as compared with obesity and smoking, possibly since chronic conditions are more likely to impact reimbursement.
