Predicting and Monitoring Symptoms in Diagnosed Depression Using Mobile Phone Data: An Observational Study

Abstract

Abstract Background Clinical diagnostic assessments and outcome monitoring of patients with depression rely predominantly on interviews by professionals and the use of self-report questionnaires. The ubiquity of smartphones and other personal consumer devices has prompted research into the potential of data collected via these devices to serve as digital behavioral markers for indicating presence and monitoring of outcome of depression. Objective This paper explores the potential of using behavioral data collected with mobile phones to detect and monitor depression symptoms in patients diagnosed with depression. Methods In a prospective cohort study, we collected smartphone behavioral data for up to one year. The study consists of observations from 99 subjects, including healthy controls (n=25) and patients diagnosed with various depressive disorders: major depressive disorder (MDD) (n=46), major depressive disorder with comorbid borderline personality disorder (MDD|BPD) (n=16), and bipolar disorder with major depressive episodes (MDE|BD) (n=12). Data were labeled based on depression severity, using the 9-item Patient Health Questionnaire (PHQ-9) scores. We performed statistical analysis and employed supervised machine learning on the data to classify the severity of depression and observe changes in the depression state over time. Results We identified 32 behavioral markers associated with the changes in depressive state. Our analysis classified depressed subjects with an accuracy of 82% and depression state transitions with an accuracy of 75%. Conclusions The use of mobile phone digital behavioral markers to supplement clinical evaluations may aid in detecting the presence and relapse of clinical depression and monitoring its outcome, particularly if combined with intermittent use of self-report of symptoms.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 2 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.