Validity of a stage algorithm for physical activity in participants recruited from orthopedic and cardiac rehabilitation clinics.

OBJECTIVE: For stage-matched interventions, individuals must be classified with respect to their previous behaviors and in conjunction with their future intentions. A novel procedure for the assessment of stages in physical activity was developed. For this, individuals' activity and their regarding intentions were compared with recommended levels of activity. The aim was to examine the psychometric properties. DESIGN: In a cross-sectional study, stages were assessed in 366 study participants (84 in cardiac and 282 in orthopedic rehabilitation) in terms of their previous physical activity and their intention to perform recommended activity levels in the future. MAIN OUTCOME MEASURES: Stages of change were compared to self-reported behavior, intention, planning, self-efficacy, risk perception, pros, cons, and social support. Misclassification, sensitivity, specificity, receiver operating characteristic (ROC) curves, non-linear trends, and planned contrasts were computed. RESULTS: In comparison to previous studies, sensitivity (44%-99%) was high and specificity was similar or low (3%-88%), depending on the type of validation outcome selected. When using less demanding criteria (i.e., less intensive activity), measurement quality decreased, although not always significantly. Applying contrast analyses, more than half of the predicted stage differences were confirmed. No main differences between orthopedic and cardiac, ambulant and stationary rehabilitation appeared and no interactions were found. CONCLUSION: The stage algorithm proved to have acceptable measurement qualities in study participants recruited in both cardiac and orthopedic rehabilitation. Especially in detecting Intenders and Actors the stage algorithm performed well. Mechanisms of adopting and maintaining recommended activity levels seem to operate equally in both groups.