Chapter 4 64 Adherence to prescribed exercise In all studies, adherence was measured with self-reported questionnaires. Twentysix studies used validated questionnaires, and the remaining studies used exercise logs, statements, or the simple question whether patients adhered to the program. The most commonly used questionnaire was the Godin Leisure Time Exercise Questionnaire (GLTEQ). A pedometer to objectify exercise was used by six studies [31-36]. Prognostic factors of exercise adherence The starting point for the prognostic factors of exercise adherence were the five WHO domains. Patient-related factors were evaluated by 41 studies [29-32, 35, 37-72] (Table 1), social-economic factors by 29 studies [29-31, 33, 34, 43, 45, 50, 51, 57, 58, 66, 69, 71-86], therapy-related factors by 5 studies [36, 45, 48, 72, 87], condition-related factors by 15 studies [34, 37, 42, 48, 51, 53, 62, 75, 78, 81, 82, 86-89] and health-system factors by 4 studies [48, 58, 61, 73]. Quality assessment and GRADE recommendations The overall Risk of Bias in patient-related factors was low RoB in 12 studies, moderate RoB in 5 studies and high RoB in 24 studies. In social-economic factors, 11 studies had low RoB, 3 studies had moderate RoB and 15 studies high RoB. In therapy-related factors, 3 studies had low RoB, 1 study had moderate RoB and 1 study had high RoB. In condition-related factors, 3 studies had low RoB, 2 studies had moderate RoB and 10 studies had high RoB. In health-system factors, 1 study had low RoB and 3 studies had high RoB. A concern regarding the quality of most included studies was the likelihood of selection bias and confounding. Prognostic factors, categorized by the five domains, reported by at least two studies were assessed using the GRADE framework (Table 1). Publication bias could only be assessed for the prognostic factors self-efficacy and exercise history (more than ten studies); therefore, this was not included in the grading. The Egger’s regression test showed statistically significant funnel plot asymmetry for the prognostic factor self-efficacy and exercise history (p < 0.0001). Also, visual inspection of the funnel plots by eye-balling suggested asymmetry and thus a potential publication bias. Overall, higher exercise adherence was predicted by patient-related factors (more self-efficacy, exercise history, motivation and PBC), social-economic factors (better education and physical health) and condition-related factors (less comorbidities, depression and fatigue).
RkJQdWJsaXNoZXIy MjY0ODMw