Development of a Cohesive Predictive Model for Substance Use Disorder Rehabilitation Using Passive Digital Biomarkers, Psychological Assessments, and Automated Facial Emotion Recognition: Protocol for a Prospective Cohort Study
Bibliographic record
Abstract
BACKGROUND: Substance use disorder (SUD) involves excessive substance consumption and persistent reward-seeking behaviors, leading to serious physical, cognitive, and social consequences. This disorder is a global health crisis tied to increased mortality, unemployment, and reduced quality of life. Altered brain connectivity, circadian rhythms, and dopaminergic pathways contribute to sleep disorders, anxiety, and stress, which worsen SUD severity and relapse. Factors like trauma and socioeconomic disadvantages heighten risk. Digital health technologies, including wearables and machine learning, show promise for diagnosis, monitoring, and intervention, from relapse prediction to early detection of comorbidities. With high relapse rates and younger patient cases, these innovations could enhance the treatment outcomes of SUD. OBJECTIVE: The objective of this study is to develop and validate a predictive model with machine learning for the duration of therapy and the rehabilitation or relapse in patients with SUD, using digital physiological measurements, psychological profiles, automatic facial emotion recognition, and the emotional state during craving. METHODS: The study will be conducted with adult male patients with SUD at a rehabilitation center and control volunteers. Participants will undergo a self-reported demographic and psychological assessment, a clinician-administered craving and emotional reaction test, and will also be monitored using a smartwatch. SUD participants will be monitored for a total of 18 months (6 months during rehabilitation, an additional 12 months post discharge), and control participants for a total of 6 months. All participants will be reassessed at the sixth month of monitoring. The collected data will then be used to train models with a neural network, which will then be validated against other models and compared with other algorithms. Demographic, psychological, digital biomarkers, and craving profiles will be created, correlations will be analyzed, and they will be compared with controls to generate a digital phenotype of SUD. When the model achieves adequate validity (area under the curve of ≥0.80), a graphic user interface will then be designed for clinical use. RESULTS: The study is supported by the Program for the Improvement of Working Conditions for Members of the SNII and SNCA (PROSNII U006EST), and APPAC-VII-CUCS-2025 for Article Publication Fees, from the University of Guadalajara. The research protocol was approved by the University of Guadalajara (reference CI-01225) in January 2025. Recruitment of patients with SUD and control participants will take place from January 2025 through January 2027. CONCLUSIONS: As shown in recent studies, accessible and affordable wearables, like commercial smartwatches, combined with psychological, demographic, and emotional state data, used with a machine learning predictive model, may be able to be used as tools to enhance SUD rehabilitation and prevent relapse. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): PRR1-10.2196/71374.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".