The Road Ahead: Exploring the potential of sensor-based measurements during inpatient stroke rehabilitation

This research was part of the Making Sense of Sensor Data consortium, which aimed to examine the extent to which sensor technology can contribute to further personalization of inpatient rehabilitation after stroke.

In a systematic review, 30 prognostic models were evaluated for predicting independent walking in the subacute and chronic phases after stroke. Six of the 30 models showed excellent accuracy. However, these models were not externally validated, were time-consuming, and/or were not easy applicable in clinical practice. Given the increasing number of people with stroke and the growing pressure on healthcare systems, these findings highlighted the need for new methods to measure and predict recovery. Ideally, such methods should assess recovery or functioning after stroke in an objective, fast, and continuous way.

Therefore, the aim of this PhD project was to explore the added value of sensor-based measurements compared to conventional clinical assessments for measuring and predicting recovery during inpatient stroke rehabilitation. At three-week intervals, patients’ functioning was assessed during inpatient stroke rehabilitation using both conventional tests and inertial measurement units (i.e., wearable sensors).

In a longitudinal study, sensor-based measurements of postural sway showed limited sensitivity to change during rehabilitation (3.2–23.9%). Additionally, gait parameters derived from sensors did not improve the estimation of walking in daily life compared to walking speed. In a cross-sectional study, the addition of sensor-based variables did not result in a significant increase in explained variance of independence in walking and activities of daily living.

In conclusion, sensor-based measurements do not provide added value besides conventional instruments for measuring or predicting recovery during clinical rehabilitation after stroke. Nevertheless, sensor technology still holds promise for continuous and objective monitoring of physical functioning with minimal burden for healthcare professionals.