Digital Biomarkers in Chronic Disease Management

Abstract

Chronic diseases account for the bulk of healthcare spending in most high-income countries, yet the clinical tools used to monitor them have barely changed in decades. A patient with heart failure sees their cardiologist every few months, gets weighed, has blood drawn, and goes home with advice that may or may not reflect what is actually happening between visits. Digital biomarkers -- objective, quantifiable measures extracted from wearable sensors, smartphones, and connected home devices -- promise to fill that gap by turning episodic clinic snapshots into continuous physiological streams. The idea is not new, but the technology has only recently reached a point where sensor accuracy, battery life, and data infrastructure can support clinical-grade monitoring outside the hospital. This study examined 180 digital biomarker programmes targeting chronic disease management, spanning five disease domains: heart failure, type 2 diabetes, chronic obstructive pulmonary disease, Parkinson disease, and major depressive disorder. Programmes were collected from a single institution -- Nordic Technical University in Stockholm -- and its clinical collaborators across Sweden over the period 2019 to 2023. A Digital Biomarker Quality Index (DBQI) was constructed from five sub-scores: signal fidelity, clinical predictive power, patient adherence feasibility, regulatory pathway maturity, and interoperability with electronic health records. Regression-derived weights tied each sub-score to observed translational outcomes. DBQI predicted clinical adoption with r = +0.84 and an AUC of 0.882. Heart failure programmes scored highest (mean DBQI 0.824), while major depression programmes lagged at 0.598. Only 33.9 percent of programmes exceeded DBQI of 0.75