Digital Biomarkers and Wearable Bioelectronics Across Neurological, Cardiopulmonary, and Mental Health Conditions: Clinical Evidence and Barriers to Adoption

Abstract

Chronic diseases now account for the majority of global morbidity, yet the clinical tools used to track them remain largely episodic, relying on brief in-clinic snapshots rather than a patient's lived, day-to-day physiology. Wearable bioelectronics and the digital biomarkers (dBMs) they generate have been proposed as a remedy, but the evidence supporting their clinical utility is scattered across disparate disease literatures. We conducted a narrative synthesis of clinical, materials-science, and computational evidence on wearable-derived digital biomarkers across neurological, cardiopulmonary, inflammatory, wound-care, and mental-health conditions, drawing on peer-reviewed studies identified through the source literature and organized according to a reproducible, PRISMA-informed search and appraisal framework. Passive sensing demonstrated strong ecological validity: smartwatch-based tremor and gait metrics tracked closely with clinician-rated Parkinson's disease scales, under-mattress bed sensors detected pre-clinical heart-failure decompensation, and closed-loop smart bandages achieved autonomous, infection-triggered antibiotic release. Deep-learning models incorporating autoencoder-derived embeddings predicted depressive symptom severity with a correlation of r = 0.73, outperforming handcrafted statistical features alone. Materials innovations, particularly zwitterionic anti-fouling coatings, reduced biosensor signal drift to under 2% per day. Digital biomarkers have moved credibly beyond proof-of-concept, but scaled clinical adoption remains constrained by biosensor durability, limited external validation, fragmented health-record interoperability, and inequitable access. A coordinated engineering, regulatory, and equity-oriented roadmap is required before these tools can be trusted as everyday clinical instruments.