OBJECTIVES: The existence of somatic and psychiatric comorbidity among patients with depressive disorder is well-known and may influence estimates of health-related quality of life (HrQoL) in depression. This study reports on the effect of comorbidity on self-reported EQ-5D in the general population. METHODS: Data from the US Medical Expenditures Panel Study (MEPS) 2003 was used. A weighted linear regression model was constructed to allow for its complex sampling design. Socio-economic covariates were included in the model, dichotomized (any diagnosis vs no diagnosis) comorbidity variables based on ICD-9 chapters were then added. The effects of major chronic diseases were analyzed separately as comorbidity variables. Separate models were constructed for men and women. RESULTS: The EQ-5Dindex adjusted for age and depression was 0.902 for men and 0.887 for women; depression had a disutility of 0.257 for men and 0.207 for women. Adding socio-economic covariates yielded a value for EQ-5D of 0.978 and 0.987 respectively, the disutility of depression decreased to 0.246 for men and 0.172 for women. Among disease conditions that contributed significantly to the estimate of EQ-5D and disutility for depression in both genders were other psychiatric conditions, diabetes, migraine and other neurologic conditions, cardiovascular disease and musculoskeletal diseases. For women only, ischaemic heart disease influenced the the model significantly while thyroid diseases had a significant influence for men only. The final disutility of depression in the model was 0.154 for men and 0.115 for women. CONCLUSION: A generalized estimate of the disutility of depression in the population for use in pharmacoeconomic studies may be difficult to obtain due to the somatic and psychiatric comorbidity in these patients. The comorbidity has a significant influence on the population disutility estimate for depression.