This document summarizes the key findings from “An explorative study on movement detection using wearable sensors in acute care hospital patients | Scientific Reports,” published on June 6, 2025. The study investigates the efficacy and patient acceptance of wearable sensors for monitoring physical activity in hospitalized elderly patients, a population particularly vulnerable to Hospital-Associated Disability (HAD).

Main Themes and Key Findings:

1. The Problem of Hospital-Associated Disability (HAD) and Low Physical Activity (PA):

• HAD is a “severe and common problem among hospitalized elderly in acute care,” often leading to “loss of independence” and impairment of Activities of Daily Living (ADLs).
• Over “30% of adults aged 65 and older experience functional decline due to HAD.”
• Low PA during hospitalization is a “primary driver of HAD,” directly linked to “functional decline, prolonged stays, and higher readmission rates among the elderly.”
• Hospitalized patients are highly sedentary, spending an average of “20.8 h/day (86.5%) sedentary.”
• Increasing PA could significantly reduce the Length of Stay (LOS) and hospital costs; for example, a “50% increase in the overall step count of hospitalized elderly patients can lead to a 6% reduction in the length of stay (LOS).”

2. The Need for Systematic Physical Activity Monitoring:

• Current methods of estimating patient PA by patients and staff are “inaccurate.”
• Patients rarely receive “guidance on the PA they need, nor specific instructions on how to achieve PA goals.”
• Wearable sensors offer an “effective means for monitoring patients’ physical activity,” providing “data to set achievable goals and motivation for patients to stay active.”
• Existing commercial activity monitors often exhibit “poor validity in patients with gait aids after total knee arthroplasty,” as they are typically “trained on data from healthy individuals” and do not account for the “distinct gait patterns due to walking aids or co-morbidities” in elderly patients.

3. Optimal Sensor Placement for Accuracy and Patient Comfort:

• The study aimed to develop a machine learning model to classify six activities: lying, sitting, standing, sit-to-stand, walking, and climbing stairs, using accelerometer and gyroscope data from three locations: wrist, ankle, and thigh.
• Accuracy: The ankle model performed best, achieving overall accuracies of 84.6% (accelerometer and gyroscope, AG-Model) and 82.6% (accelerometer only, A-Model).
• In contrast, wrist and thigh models showed significantly lower performance, with accuracies ranging from 72.4% to 76.8%.
• The “ankle location showed a significant advantage in the model’s classification performance over both the wrist and thigh sensor locations.”
• The lower performance of wrist and thigh models is attributed to “increased variability in wrist movement patterns” (especially with walking aids) and “close similarity in leg movement patterns during walking and stair activities” for the thigh.
• Patient Acceptance and Comfort: Patient questionnaire results indicated a high acceptance of 97.7% towards wearing a monitoring device for 8 hours, regardless of location.
• The ankle was reported as the least disturbing location in 87.2% of cases, followed by the wrist (70.9%) and thigh (67.4%).
• Most patients (89.3%) reported that “none of the sensor placements were particularly bothersome.”
• Minor discomfort was reported with the wrist (5.9%), ankle (3.6%), and thigh (2.4%) being the most irritating locations for a small percentage of responses.
• The band-based system for ankle attachment was highly feasible due to its “comfortable wearing and easy, secure setup.” Patch detachments were reported in 12.9% of cases for thigh attachment.
• Conclusion on Optimal Location: The findings “indicate the ankle as the ideal location to track patient activities in a clinical setting, compared to the wrist and thigh,” considering both accuracy and patient comfort.

4. Role of Gyroscope Features:

• Gyroscope features “were particularly influential in classifying sitting, standing, walking, and stairs at the ankle and thigh locations.”
• They are “particularly valuable in scenarios where similar extremity positions occur across multiple activities” (e.g., distinguishing lying, sitting, and standing).
• However, incorporating a gyroscope “increases power consumption,” which can “significantly reduce battery life and limit the suitability of wearable systems for continuous monitoring in clinical settings.”