OBJECTIVES: Back pain is an expensive medical condition and direct medical costs associated with back pain are significant. Estimating resource utilization costs for diseases like back pain based on disease coding have the potential for under-estimation. The objective of this study was to determine direct costs due to back pain in the US population using an incremental cost approach. METHODS: Analysis of the 2001 Medical Expenditure Panel Survey (MEPS) was conducted. Patients who had back pain related physician visits or treatments during 2001 were identified using International Classification of Diseases (ICD-9) codes. Patients without a back pain diagnosis and without a claim for back pain were treated as controls. Least squares regression was used to estimate the incremental cost of back pain adjusting for age, gender, race, occupation, and co-morbidities using the Charlson co-morbidity index. Sample data was projected to the US population and 95 percent confidence limits for estimates were calculated using the Taylor expansion method. RESULTS: Prevalence of back pain in the US was 8.3% (26,167,199) of the total population. Total annual direct costs for back pain patients were $32,135,937,092 after adjusting for co-morbidities. Mean annual direct cost for a back pain patient was $4241.3 (95% CL $3890.8 – $4591.8). Office-based medical provider visits (29.4%), in-patient visits (27.1%), and prescribed medicines (21.3%) were major cost centers for back pain patients. CONCLUSIONS: With direct medical costs estimated at more than $32.0 billion in 2001, back pain costs represent a significant amount of health care expenditures. The estimate obtained was more than twice the magnitude of earlier estimates based only on expenditures coded for back pain. Potential for disease cost under estimation may be reduced by using the incremental cost approach.