Evaluation of enrollment bias in a physical-activity-promotion program forseniors

Health-promotion programs' success depends on their ability to enroll repre sentative samples of the target population, particularly those who are hard to reach and those who can benefit the most from such programs. This artic le evaluates enrollment bias in the recruitment process, examines the usefu lness of a 2-phased recruitment strategy in enrolling representative propor tions of eligible individuals in a physical-activity-promotion program for older adults, and explores predictors of enrollment. Of 1,381 randomly sele cted Medicare HMO members, 519 were eligible. Of these, 54% attended an inf ormational meeting and 33% enrolled in the program. Relative to the target population, a representative proportion of women was enrolled, but those wh o enrolled were slightly younger. Of those who were eligible, a representat ive proportion of sedentary participants was recruited, those who were over weight were overrepresented. and the oldest old. less educated, ethnic mino rities, and precontemplators of physical activity were underrepresented. Mo difiable predictors of enrollment included interest in health, previous hea lth-class attendance, and having had a physician recommend exercise.