Prediction of Physical Activity Patterns in Older Patients Rehabilitating After Hip Fracture Surgery: Exploratory Study.
Dieuwke van Dartel, Ying Wang, Johannes H Hegeman and 1 others
PMID 38032703WHAT IT FOUND
In this small exploratory analysis, a machine learning model correctly predicted two physical activity patterns for all seven patients in the test set using only the first week of wearable data.
This suggests early activity levels may forecast recovery trajectories, but the tiny sample size makes these results preliminary.
Key findings
01The model using the first 7 days of data achieved perfect classification on the test set.
02Physical activity intensity on the first day and morphological features were the most relevant predictors.
03Clinical scores like the Barthel Index were not selected as relevant features for prediction.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study is highly exploratory with a very small sample size (37 patients), and the test set contained only seven patients. The 100% accuracy on the test set may not generalize to larger, more diverse populations. Only two of the five observed physical activity patterns were analyzed; the remaining patterns and 'Else' cases were excluded. The machine learning models used were basic, and the study did not validate the predictions against clinical outcomes like length of stay or functional recovery in this specific analysis.
Declared interests
The authors declared no conflicts of interest.
The easy way to misread this
Do not interpret the 100% prediction accuracy as evidence that this model is ready for clinical use. The perfect score was achieved on a test set of only seven patients, which is far too small to establish reliability or generalizability.