Trajectories of physical activity among participants of six-month mobile app-based disease management program for cardiovascular risk: a group-based trajectory modelling approach

Abstract Background Mobile app-based health promotion is an emerging preventive intervention for cardiovascular disease that allows for the frequent and longitudinal monitoring of physical activity (PA). Previous studies exploring PA trajectories categorized participants in a limited number of follow-up assessments and did not consider dynamic nature with changes that occur over the intervention period. Purpose This study explored the trajectories of PA during mobile app-based disease management programs in adults with cardiovascular risk using group-based trajectory modelling (GBTM). Methods Participants who completed a six-month disease management program using a mobile app were enrolled. The program promoted behavioural changes for the management of cardiovascular risk factors in individuals with one or more of hypertension, diabetes, or dyslipidaemia. Daily step count was measured using a wrist-worn activity tracker (Fitbit devices). Each participant was encouraged to wear a tracker for the monitoring of exercise and physical activity. To analyse relatively physically inactive participants, we included those with baseline step count <8,000 steps/day. GBTM was conducted to identify distinct patterns of longitudinal step count changes of baseline, 3-22 weeks, and 24 weeks. Participants with PA measurements <4 time points were excluded. The baseline characteristics and indicators related to mobile app usage during the first two weeks were compared among the trajectory groups. Results A total of 1,760 participants were included in the analysis (age: median 55 years; 83.8% men; body mass index (BMI): median 27.0 kg/m²; baseline step count: median 6,118 steps/day). GBTM identified four step count trajectory groups: rapid increase (T1, rapid increase in the first three months; 5%), gradual increase (T2, gradual increase in the first three months; 24%), moderate activity (T3, maintained at around 7,000 steps/day; 45%), and physically inactive (T4, around 4,500 steps/day throughout the program; 26%) (Figure 1). The trajectory groups that showed increasing PA patterns were older (T1: median 57 years; T2: 55 years; T3: 55 years; T4: 54 years; p=0.040) with a lower baseline BMI (T1: median 26.6 kg/m²; T2: 26.4 kg/m²; T3: 27.0 kg/m²; T4: 27.7 kg/m²; p<0.001) and were predominantly men (T1: 89.5%; T2: 87.4%; T3: 82.8%; T4: 81.1%; p=0.028). PA increasing patterns were associated with a longer time of mobile app use (T1: median 8.5 min/day; T2: 5.5 min/day; T3: 5.5 min/day; T4: 4.9 min/day; p=0.012) and more login times (T1: median of daily average 3.9/day; T2: 3.0/day; T3: 3.1/day; T4: 2.9/day; p=0.003) during the first two weeks of the program. Conclusions GBTM showed that 30% of participants in the mobile app-based intervention showed significant PA improvement during the first 12 weeks. Results of this study suggest that the response to PA intervention may be associated with lifestyle prior to the program and willingness to use the mobile app.Figure 1.Step count trajectories