AI Framework Identifies Accelerometer Adherence Phenotypes Predicting Outcomes
Researchers analyzed accelerometer data from the RESILIENT randomized trial (enrolled January 9, 2020–January 10, 2024) of 400 older adults to derive adherence phenotypes using a time-series k-means AI framework on 3-month Fitbit wear. They identified distinct adherence clusters and found the consistently high-adherence phenotype associated with a 38.5-meter greater 6-minute walk distance versus the low-adherence cluster (95% CI 2.2–74.7; P=.04), indicating modest functional benefit tied to sustained device use.
Key Points
- 1Identify distinct adherence phenotypes via k-means clustering on 3-month accelerometer time-series data
- 2Show consistently high-adherence phenotype links to larger 6-minute walk improvements (38.5 m, P=.04)
- 3Suggest tailoring mHealth-CR interventions by adherence phenotype to potentially improve functional outcomes
Scoring Rationale
Moderate novelty with robust RCT evidence supports relevance, but effect sizes are modest and generalizability is limited.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
