Trajectories of physical activity after ischaemic stroke: Exploring prediction of change

Background and purpose: Physical activity (PA) is associated with lower risk of stroke and better functional outcome. However, low levels of PA after stroke are prevalent. The aim was to identify predictors of PA change after first-time ischaemic stroke and to develop prediction models to predict change in PA.

Methods: Pre-stroke and 6 months post-stroke PA were quantified with the Physical Activity Scale for the Elderly (PASE). Considered predictors were clinical data and demographics including data on socioeconomic status (SES). PASE change was analysed using mixed models of repeated measures. Elastic net regression models were used to predict decrease from higher PASE quartile to the lowest and increase from lowest to higher.

Results: A total of 523 first-time ischaemic stroke patients were included. Median (interquartile range, IQR) age was 69 years (IQR 59, 77), 181 (35%) were female and median National Institutes of Health Stroke Scale score was 3 (IQR 2, 5). Overall PASE score did not change, but 20% of patients decreased to the lowest PASE quartile whereas 48% from the lowest quartile increased to a higher. Prediction performance measured by area under the receiver operating curve was 0.679 for PA decrease and 0.619 for increase. SES factors were the most consistent predictors.

Conclusions: Half of the least active patients increased PA level after stroke whereas a fifth decreased with SES being the most consistent predictor. Despite comprehensive data, the prediction models only performed modestly. Efforts to optimize PA should include all stroke survivors to increase PA for least active patients and to prevent PA decrease.