# Links between self-monitoring data collected through smartphones and smartwatches and the individual disease trajectories of adult patients with depressive disorders: Study protocol of a one-year observational trial

## TL;DR

This study explores how smartphone and smartwatch data can track depression severity and help personalize treatment for patients.

## Contribution

It introduces a method to use self-monitoring data for predicting and tracking depression symptoms in individuals.

## Key findings

- Biosensor data from smartphones and smartwatches may serve as objective markers for depression severity.
- Machine learning models could predict self-reported depressive symptoms over time.
- Personalized digital interventions may be tailored based on individual data patterns.

## Abstract

Depression is highly recurrent and heterogenous in its individual course, requiring a personalized treatment approach. Patients today can collect large volumes of personal data via smartphones and smartwatches and may utilize them for their treatment and self-management. We aim to provide proof-of-concept that these data can (i) serve as an objective marker of and (ii) predict the daily and weekly self-reported depression severity within individuals with depressive disorders.

In this exploratory study, 15 adult patients with depressive disorders will collect self-report and biosensor data over the course of one year. Participants will (a) attend three in-person appointments (at baseline, 6 months, and 12 months), (b) self-report daily and weekly depressive symptoms, (c) continuously collect sensor data via the “iTrackDepression” app on their Android smartphone (app usage, phone calls, phonetic parameters from voice recordings), and (d) wear a Samsung Galaxy Watch 5® to record data from the accelerometer, step sensor, light sensor, and heart rate sensor. We will apply multilevel correlations, vector-autoregressive models, and Machine Learning approaches to identify individual patterns in the data, particularly in the relationships between biosensor data and self-reported depressive symptoms.

Enhancing the understanding of individual disease trajectories through data from smartphones and smartwatches could allow for classical, digital, and self-management interventions for depression to be delivered in a manner and at a time specifically tailored to the individual's needs.

Clinical trial registration number: DRKS00032618 (https://drks.de/search/en/trial/DRKS00032618)

•The phenotype of depressive disorders is highly heterogeneous.•Patients are owners of large personal datasets from smartphones and smartwatches.•Data-driven methods can generate personalized disease markers and predictive models.•Patients with depression may use these for disease monitoring and self-management.

The phenotype of depressive disorders is highly heterogeneous.

Patients are owners of large personal datasets from smartphones and smartwatches.

Data-driven methods can generate personalized disease markers and predictive models.

Patients with depression may use these for disease monitoring and self-management.