Methodological Notes

The survey actually comprised two separate surveys, each based on telephone interviews with national probability samples of:

• Patients with COPD, defined as people aged 45 and older who had been diagnosed with COPD, emphysema, chronic bronchitis, or who met stringent symptom criteria for undiagnosed chronic bronchitis. A total of 573 people with COPD were interviewed. This sample was identified by systematically screening a geographically stratified national sample of 26,880 US households by telephone.

• Doctors, including 100 physicians from adult primary specialties of general practice, family practice, and internal medicine, and 103 specialists in pulmonology. National probability samples were drawn from each of these populations.

Patient and physician interviews were conducted by telephone from August to November, 2000. Patient interviews averaged 34 minutes in length, while the physician interviews averaged 22-26 minutes in length. The completed samples were weighted to correct for any disproportionate sampling or unequal probabilities of selection; the weighted survey results should yield unbiased national estimates.

The maximum expected sampling error was ± 4.1 percentage points at the 95% confidence level for a sample of 573 (e.g., patient survey), while the maximum expected sampling error for a simple random sample of 100 (e.g., primary care doctor survey) was ± 9.8 percentage points at the 95% confidence level. Confronting COPD in America may be considered a landmark study in research on chronic obstructive pulmonary disease because:

• The survey is one of the most comprehensive surveys about COPD conducted with a national probability sample of patients with COPD ever conducted in the United States. The 34-minute patient interview covered a broader range of subjects, including symptoms, morbidity, treatment, and quality of life, in more depth than any previous national probability survey of COPD in America.

• The questionnaires for the surveys of the healthcare providers were designed to permit comparisons with patient responses. The study therefore offers insight into the different perspectives of patients and providers on treatment and other issues.