Accuracy of an Algorithm in Predicting Upper Limb Functional Capacity in a United States Population.
Jessica Barth, Kimberly J Waddell, Marghuretta D Bland and 1 others
PMID 34425091WHAT IT FOUND
In 49 US stroke patients, a simple upper limb prediction algorithm was better than chance but less accurate than expected.
It was most uncertain for people with little initial arm strength.
Key findings
01Among 49 participants with complete data, the overall accuracy was 61%, lower than hypothesized.
02The algorithm was overall better than chance and most useful for determining which category a person would not end up in.
03For 20 people with an initial SAFE score below 5, the model was accurate for 50%.
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
Only 49 of 69 enrolled participants had complete data, and the small sample makes the accuracy estimate imprecise. The study excluded people who could not follow a 2-step command, so it does not show how the algorithm performs for people with substantial cognitive or language deficits. The study did not control or prescribe rehabilitation services, so outcomes reflect usual care rather than a defined therapy programme. The analysis changed the algorithm for people with an NIHSS score of 9, predicting a Limited outcome. The sample was younger than the development sample, with 96% of participants under 80 years.
Declared interests
The paper states the study was supported by NIH grant R01 HD068290. The supplied text does not report author conflicts of interest.
The easy way to misread this
Do not treat a 61% overall accuracy as a reliable prediction for every patient. It was lower than hypothesized, and for people with an initial SAFE score below 5 the model was accurate for only 50% of them.