Abstract
OBJECTIVE: To explore trajectories of 12-week adherence to a digital education and exercise therapy for knee and hip osteoarthritis (OA), associations with baseline characteristics, and trajectories of patient-reported outcomes measures (PROMs) up to 1-year follow-up.
DESIGN: Retrospective cohort (registry) study.
METHODS: Weekly data on adherence (i.e. the percentage of completed activities (exercises, lessons, and quizzes)) were obtained over 12 weeks (n=14,097). Longitudinal k-means clustering was used to identify adherence trajectory clusters. Associations of baseline characteristics with adherence trajectory clusters were assessed using multinomial logistic regression. Trajectories of each PROM (pain, function and general health) from baseline up to 1-year follow-up (measured at 3-month intervals) across adherence trajectory clusters were explored using generalized estimating equations adjusted for baseline characteristics.
RESULTS: Four adherence trajectory clusters were identified: “high-persistent” (68.0%), “high-declining” (16.6%), “moderate-increasing” (8.5%), and “moderate-declining” (6.9%). Multinomial logistic regression suggested that female sex, older age, lower body mass index, lower education, living outside metropolitan cities, higher level of physical activity, less anxiety/depression, no fear of movement, having walking difficulties, and higher readiness to do exercise were associated with a higher probability of assignment to “high-persistent” than other clusters. Beliefs/perceptions and sociodemographic factors accounted for most of the explained variation in adherence trajectory clusters. While “high-persistent” cluster generally reported better outcomes than other clusters, these differences were small.
CONCLUSION: While there were variations in adherence to the digital treatment, participants reported clinically comparable PROMs regardless of their adherence trajectory cluster.
DESIGN: Retrospective cohort (registry) study.
METHODS: Weekly data on adherence (i.e. the percentage of completed activities (exercises, lessons, and quizzes)) were obtained over 12 weeks (n=14,097). Longitudinal k-means clustering was used to identify adherence trajectory clusters. Associations of baseline characteristics with adherence trajectory clusters were assessed using multinomial logistic regression. Trajectories of each PROM (pain, function and general health) from baseline up to 1-year follow-up (measured at 3-month intervals) across adherence trajectory clusters were explored using generalized estimating equations adjusted for baseline characteristics.
RESULTS: Four adherence trajectory clusters were identified: “high-persistent” (68.0%), “high-declining” (16.6%), “moderate-increasing” (8.5%), and “moderate-declining” (6.9%). Multinomial logistic regression suggested that female sex, older age, lower body mass index, lower education, living outside metropolitan cities, higher level of physical activity, less anxiety/depression, no fear of movement, having walking difficulties, and higher readiness to do exercise were associated with a higher probability of assignment to “high-persistent” than other clusters. Beliefs/perceptions and sociodemographic factors accounted for most of the explained variation in adherence trajectory clusters. While “high-persistent” cluster generally reported better outcomes than other clusters, these differences were small.
CONCLUSION: While there were variations in adherence to the digital treatment, participants reported clinically comparable PROMs regardless of their adherence trajectory cluster.
| Original language | English |
|---|---|
| Pages (from-to) | 56-67 |
| Journal | Journal of Orthopaedic and Sports Physical Therapy |
| Volume | 55 |
| Issue number | 1 |
| Early online date | 2024 Nov 23 |
| DOIs | |
| Publication status | Published - 2025 |
Subject classification (UKÄ)
- Orthopedics
