Using Smartphone-Based Digital Phenotyping to Predict Relapse in Serious Mental Disorders Among Slum Residents in Dhaka, Bangladesh: Protocol for a Machine Learning Study
Bibliographic record
Abstract
Background: Serious mental illnesses (SMIs) are associated with high relapse rates and limited access to continuous care, particularly in low-resource settings such as urban slums. Traditional clinical monitoring is constrained by accessibility and scalability challenges. Digital phenotyping, through passive smartphone data, offers a novel approach to predict relapse by capturing real-world behavioral changes. Objective: This study aims to evaluate the feasibility and predictive value of smartphone-based digital phenotyping for detecting relapse in individuals with SMIs living in the Korail slum of Dhaka, Bangladesh. Methods: This prospective 6-month cohort study will recruit 430 participants diagnosed with SMIs who own Android (Google LLC) smartphones. Passive data (eg, screen time, mobility, and call or text frequency) will be continuously collected using a custom-built app (DataDoc). Monthly active data, including symptom and functioning assessments, will be collected via self-report and clinical engagement. Machine learning models will integrate these data to detect early warning signs and predict relapse trajectories. Results: This study was funded by the NIHR (National Institute for Health and Care Research; award number NIHR200846) in October 2022. Data collection commenced in August 2025 and is ongoing. A total of 14 participants have been recruited, as of January 2026. Preliminary data analysis is ongoing, with expected results to be published in fall 2026. Conclusions: This study is one of the first to apply smartphone-based digital phenotyping and machine learning for relapse prediction in low- and middle-income countries' slum settings. The findings will inform scalable, low-cost digital interventions to address the mental health treatment gap in underresourced communities.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.011 | 0.013 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.004 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.044 | 0.011 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".