Monitoring fatigue with consumer wearables in multiple sclerosis following rehabilitation: An observational study
Baumer, A. M.; Haag, C.; Sieber, C.; Stegmann, M.; Sylvester, R.; Gonzenbach, R.; von Wyl, V.
Show abstract
Fatigue places a great burden on people with multiple sclerosis (PwMS), but fatigue assessments are imperfect and rely on subjective surveys. Assessing digital biomarkers derived from wearable devices has shown promise as an objective measurement method. We investigate their association with fatigue with a special focus on within-person changes and differences between cognitive and motor fatigue. PwMS entering inpatient rehabilitation were recruited and equipped with a consumer-grade fitness watch (Fitbit Inspire HR). Participants were monitored during rehabilitation and four weeks after returning home. We used hierarchical modeling to assess cross-sectional associations between digital biomarkers and fatigue scores and linear models to investigate associations between change scores. In total, 45 PwMS were included in the study (61.7% female, mean age = 46.0 (SD = 8.9). Device wearing compliance was high at 92.0% and participants provided on average 51.6 days of recorded data. The assessed digital biomarkers were generally only weakly and non-significantly associated with fatigue. The strongest cross-sectional association with fatigue was found in intraday variability (b=-0.22, 95%CI = [-0.41, -0.02]), while step counts were most responsive to within-person changes in fatigue (b=-0.32, 95%CI = [-0.68, 0.05]). On average, effect sizes were 23.3% greater for change scores compared to cross-sectional associations and 62.9% greater for motor than for cognitive fatigue. Despite the very consistent data collection by participants, digital biomarkers did not show strong associations with fatigue in MS patients, suggesting general limitations of using consumer-grade wearables for fatigue assessments. We found some evidence that digital biomarkers are better suited to reflect within-person change than interpersonal differences. Additionally, digital biomarkers related more closely to motor than cognitive fatigue. Therefore, the most promising target for future research may be within-person changes in motor fatigue.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Smartphone-Based Symbol-Digit Modalities Test Reliably Measures Cognitive Function in Multiple Sclerosis Patients 94%
- Quantifying Device Type and Handedness Biases in a Remote Parkinson’s Disease AI-Powered Assessment 91%
- Passive Detection of COVID-19 with Wearable Sensors and Explainable Machine Learning Algorithms 91%
Similar papers in this journal
- Connecting real-world digital mobility assessment to clinical outcomes for regulatory and clinical endorsement – the Mobilise-D study protocol 95%
- Brain predictors of fatigue in Rheumatoid Arthritis: a machine learning study 92%
- Supervised and self-directed technology-based dual-task exercise training programme for older adults at risk of falling – Protocol for a feasibility study 92%
Similar papers in this journal
- A Conceptual Approach to Comprehensive Interdisciplinary Cognitive Rehabilitation: A description using the TIDieR-Rehab Framework 93%
- Real-time assessment of daytime sleepiness in drivers with multiple sclerosis 92%
- Fear of relapse and quality of life in multiple sclerosis: the mediating role of psychological resilience 90%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.