Cross-platform Prediction of Depression Treatment Outcome Using Location Sensory Data on Smartphones

TL;DR

This work addresses the need for objective, timely assessment of depression treatment response by leveraging passive location data collected from smartphones. It introduces CORAL-based domain adaptation to harmonize location features from Android and iOS, enabling joint modeling of a larger, cross-platform dataset to predict treatment outcome. Using SVM and XGBoost, the study demonstrates that domain-adapted location features, especially when combined with a baseline self-report score, can achieve an $F_1$ score up to $0.67$, approaching the performance of periodic questionnaires. The findings suggest that location sensing offers a promising, low-burden approach to monitor and predict depression treatment progress, with practical implications for real-world clinical support and digital health tooling.

Abstract

Currently, depression treatment relies on closely monitoring patients response to treatment and adjusting the treatment as needed. Using self-reported or physician-administrated questionnaires to monitor treatment response is, however, burdensome, costly and suffers from recall bias. In this paper, we explore using location sensory data collected passively on smartphones to predict treatment outcome. To address heterogeneous data collection on Android and iOS phones, the two predominant smartphone platforms, we explore using domain adaptation techniques to map their data to a common feature space, and then use the data jointly to train machine learning models. Our results show that this domain adaptation approach can lead to significantly better prediction than that with no domain adaptation. In addition, our results show that using location features and baseline self-reported questionnaire score can lead to F1 score up to 0.67, comparable to that obtained using periodic self-reported questionnaires, indicating that using location data is a promising direction for predicting depression treatment outcome.

Figures (9)

• Figure 1: Time duration for which a location measurement is valid when merging GPS and WiFi location data (considering the data collected 6am - 10pm each day).
• Figure 2: (a)-(b) show the distributions of the number of days with location data and the number of location samples in a QIDS interval for the Android dataset. (c)-(d) show the corresponding distributions for the iOS dataset.
• Figure 3: Baseline QIDS score for Android and iOS users.
• Figure 4: Histogram of QIDS score changes for Android and iOS users.
• Figure 5: Android-transformed: distributions of location features; each plot shows the original iOS features, and the original and transformed Android features.