Relationship of stages of change to attendance in a diabetes prevention program

Purpose. To determine whether preintervention stage-of-change measures are indicative of subsequent attendance at diabetes prevention intervention sessions. Design. Cross-sectional. Setting. Local community. Subjects. Seventy-five adult American Indian women participated. Measures. Attendance, stage-of-change questions for seven diabetes prevention behaviors, and mean stage-of-change score to reflect the combined stages of change for all behaviors. Analysis. Univariate analyses for stage-of-change distribution and Fisher's exact test and prevalence ratios for the association between attendance and stage of change. Results. Participants' readiness for change at baseline was distributed across all stages of change. The most common stage was preparation. There was a significant relationship between the mean stage-of-change scores and attendance. Participants with lower mean stage-of-change scores (mean ± SE, 3.03 ± 0.13) were less likely to attend all 5 sessions than those with higher mean stage-of-change scores (mean ± SE, 3.38 ± 0.10) (p = .04). Participants in the action category (preparation, action, and maintenance stages) before the intervention were 6.7 (95% confidence interval, 1.0-44.1; p < .01) times more likely to be high attenders than those in the preaction category (precontemplation and contemplation stages) before the intervention. Conclusions. Findings from this study suggest that stage of change may be a good predictor of attendance at diabetes prevention intervention sessions and have implications for intervention design and assessment. The mean stage-of-change score may be a more stable estimate of stage of change.