23 lines
2.5 KiB
Text
23 lines
2.5 KiB
Text
|
## A: Predicting change in physical activity after stroke {#studyA}
|
||
|
|
|
||
|
|
At the clinical setting, when considering rehabilitation need for stroke patients, risk profiling regarding risk of low level of physical activity would be interesting to focus on patients at increased need. We hypothesized that change in physical activity after stroke could be predicted based on baseline factors.
|
||
|
|
|
||
|
|
\acr{PA} reduces the risk of stroke and improves functional outcome. We aimed to investigate predictors for decrease and increase in \acr{PA} after stroke.
|
||
|
|
|
||
|
|
Provisional results have been presented with a poster on \acr{ESOC} 2023.
|
||
|
|
|
||
|
|
### Methods
|
||
|
|
|
||
|
|
We used the \acr{PASE} questionnaire score for pre-stroke \acr{PA} at inclusion and \acr{PA} at six months. Main outcomes were decrease in \acr{PA} to the lowest \acr{PASE} score quartile from any higher, and increase from the lowest quartile to any higher. We created two cross validated, regularised, predictive models with available clinical variables.
|
||
|
|
|
||
|
|
### Provisional results
|
||
|
|
|
||
|
|
We included 522 patients from The \acr{TALOS} trial. Median \acr{IQR} age was 68 \[59; 76\], \acr{NIHSS} score 3 \[2; 5\] and 180 (34%) were females. In total, 77 patients (19.7%) decreased in PA and 63 (48.1%) increased. The decrease model had sensitivity 0.862, specificity 0.290 and \acr{AUROC} of 0.607. Variables used were increasing age, living alone, daily or occasionally smoking, hypertension, previous myocardial infarction and stroke severity. The increase model performed with sensitivity 0.882, specificity 0.246 and AUC at 0.606. Variables used were increasing age, female sex, living alone, excess alcohol intake, previous atrial fibrillation, hypertension, diabetes, peripheral arterial disease and pre-stroke \acr{mRS} \>0.
|
||
|
|
|
||
|
|
### Discussion
|
||
|
|
|
||
|
|
We found that change in \acr{PA} after stroke is highly dynamic, and our prediction models performed only modestly, with baseline factors influencing both \acr{PA} decrease and increase. These findings highlight the need for a systematic rehabilitation programme to avoid decrease and stimulate increased \acr{PA} post stroke for all patients.
|
||
|
|
|
||
|
|
### Time frame
|
||
|
|
|
||
|
|
A drafted manuscript has already been prepared, but we have chosen to enrich the data set with data on socio-economic status as well educational level. We are currently awaiting access to these data at Statistics Denmark's servers. When these data become available, I hope to be able to finalise analyses fairly quickly and have the manuscript ready for journal submission before spring 2024.
|