Problem/Objective: To estimate and compare the health care expenditures attributable to smoking by the diabetes status.

Methods: We used the 2000 to 2003 Medical Expenditure Panel Surveys MEPS) data and 1998 to 2002 National Health Interview Survey (NHIS) data to estimate smoking-attributable medical expenditures by diabetes status. The sample includes 38,299 adults. All costs data are inflated to 2005 dollars. Separate two-part econometric models are used to estimate annual smoking-attributable expenditures by diabetes status and service type

Results: Comparing the sample with diabetes (N=3,424) to the sample without diabetes (N=34,875), medical expenditures on inpatient, outpatient and prescription drugs are significantly higher for those with diabetes. People with diabetes are older; are more likely to be non-white, have lower income level, are covered by public insurance, are obese and are less likely to be married. Comparing the sample of ever smokers to never-smokers, on average, ever smokers have significantly higher health care expenditures among people without diabetes. Expenditures on prescription drugs are significantly higher for smokers among people with diabetes. Smoking-attributable fractions (SAF) for those without diabetes are larger than for those with diabetes. Smoking-attributable expenditures (SAE) are larger for all three categories but statistically non-significant (except for prescription drugs) for those with diabetes. Conclusions: Smoking has a larger impact on the level of expenditures on prescription drugs for those with diabetes. Patients with diabetes have a much higher level of total medical expenditures than those without diabetes and SAE is significantly higher among those without diabetes.