OBJECTIVES: To estimate the prevalence and annual incidence of diagnosed COPD from medical claims data and to compare results with those derived through alternative methodologies. METHODS: Claims data were acquired from a US managed care organization database with records for approximately ten million members in 20 states. The COPD cohort comprised patients aged 25+ years with one or more medical claims with a COPD diagnosis in 2002. For prevalence, patients were required to be eligible for coverage during 2002, while patients in the incidence estimation were required to be eligible for medical coverage in both 2001 and 2002 but without any COPD diagnoses in 2001. Claims-based estimates of prevalence and incidence (overall and by age group) were compared to corresponding published data from other sources. RESULTS: We identified 64,141 patients with a diagnosis of COPD (an overall prevalence of 3.7%). Prevalence progressively increased from 0.3% in patients aged 25-34 years to 11.8% in patients aged 75+ years. These estimates are generally lower than other sources. For example, National Health Interview Survey estimates, based on respondent self-reported COPD or recall of being told they have the disease, ranged from 3.9% to 10.6% in the youngest and oldest age groups, and were about 6.0% overall. A total of 7347 newly-diagnosed COPD patients were found, generating an annual incidence rate of 0.95%. Rates increased from 0.08% in patients aged 25-34 years to 2.07% in patients aged 75+ years, somewhat higher than estimates from European studies that used general practitioner case registrations as a data source. No comparable U.S. studies could be identified. CONCLUSION: Disparities exist in epidemiological estimates of COPD depending on data, patient populations, and methodologies employed. Further studies are needed to clarify the most valid and reliable data and methods for studying the epidemiology of diagnosed COPD.