Enrollment in cardiac rehabilitation has been reported to improve exercise capacity, psychological well-being, and survival. However, participation rates are low and the reasons for nonparticipation have not been adequately defined. The purpose of this study was to evaluate the major correlates of nonparticipation and to examine the level of participation of patients who stand to benefit most on the basis of preenrollment functional status and health behaviors. Three hundred ninety-three patients undergoing coronary artery bypass surgery (1) had baseline functional status and quality-of-life data collected, and (2) were recruited for participation in the Duke Center for Living comprehensive 3-week post-coronary bypass surgery rehabilitation program. Baseline demographic, clinical, catheterization, functional status, psychological status, and health behavior descriptors were analyzed to identify univariate and multivariable correlates of a patient's decision to participate in the program. At baseline, most clinical factors were similar in participants (n = 52) and nonparticipants (n = 341), but the nonparticipants were more often women (26% vs 12%, p = 0.02). Participants were also more likely to be employed (63% vs 45%, p = 0.02) and had a higher education and income distribution than nonparticipants (both p = 0.001). On 2 separate scales, nonparticipants had significantly more baseline functional impairment than participants (both p = 0.001). In multivariable analysis, the independent correlates of higher participation rates were: higher education (college graduates 71% more likely to participate than high school graduates) and better baseline Duke Activity Status Index (patients with mild functional impairment were at least 42% more likely to participate than patients with moderate impairment).(ABSTRACT TRUNCATED AT 250 WORDS)