2.3. Outcome: Adherence and contamination

PA levels of the intervention group were assessed by a web‐based diary filled out on daily basis. Automated reminders were sent out by email, and if participants were not responsive within 3 days, phone calls were initiated. In the diary, intervention participants registered each sport session performed, the duration and type of PA performed. The types of activities were manually grouped into jogging/bicycling/swimming, fitness/gym, team sports, hiking/winter sports, and other activities. PA of the control group was assessed by a 7‐day recall questionnaire filled out by the participant at baseline and at the 12‐month assessment. Answers were verified and discussed by an exercise specialist during an interview. It was specifically emphasized to have a typical week reported within the last 3 months. As the controls were allowed to switch to the exercise program after the trial, the 12‐month interview was taken as baseline assessment of their PA behavior to receive the same personalized PA counseling with motivational tools as the intervention group, but without personal follow‐up coaching.

We defined participants of the intervention group to be adherent if they reached ≥2/3 of their individual intense PA goal (100 min of the agreed 150 min additional PA/week) based on the web‐based diary and referred to them as “exercise adherent.” For this, all cumulative PA hours over the 1 year were considered and a binary variable was built to determine whether a participant was adherent or not for each week. For the regression model, we used the percentage of weeks where the PA goal was reached as outcome. Days with no PA entries were set to 0 min PA suggesting that no sport was performed. Control group participants were defined as “control contaminated” if they reported >60 min increase in intense PA/week based on the questionnaire and confirmed by an interview at T0 and T12 done by the coach. ¹¹ Dropouts were classified as non‐adherents for the yearly goal reached, but were included in the prediction model until withdrawal for the intervention group.