Participation bias in a population-based echocardiography study

PURPOSE: To compare the characteristics of participants and non-participants in a population-based study of cardiac ventricular function. METHODS: Subjects aged 45 years and older on January 1, 1997 were recruited from a sampling frame of Olmsted County residents from the Rochester Epidemiology Project. Subjects were asked to complete a 17-page questionnaire and participate in a 4-hour clinical examination that included a brief physical examination, echocardiography, spirometry, and an electrocardiogram. With specific IRB approval, the community medical records of participants and non-participants from the first wave of recruitment were examined by trained nurse abstractors to elicit details of past medical history. RESULTS: Of the first 963 persons invited to participate in the study, 488 (51%) completed all phases of the examination. Participation rates were similar among men and women (53% vs. 49%, respectively). By age, participation rates were lowest among persons aged 75 years and older (44.7% and 34.9%) and 45 to 54 years (45.4% and 44.3%) for men and women. Participation rates were not different according to past history of coronary heart disease, congestive heart failure, or other cardiovascular disease. Persons with a history of chronic obstructive pulmonary disease were less likely to participate (19.4% vs. 51%; odds ratio, 0.36; 95% confidence interval, 0.18, 0.76) after adjustment for age, sex, and comorbid conditions. CONCLUSIONS: These results provide some reassurance that participation bias in this study may have little influence on its overall findings, although this cannot be conclusive. (C) 2004 Elsevier Inc. All rights reserved.