Ethical aspects and user preferences in applying machine learning to adjust eHealth addressing substance use: A...

Area of Science:

• Digital Health
• Data Science
• Behavioral Health

Background:

• Digital health interventions are increasingly used for substance use disorders.
• Data science, including machine learning (ML), can enhance intervention effectiveness but raises ethical questions.
• User perspectives on ML in digital interventions for substance use and gambling disorder are underexplored.

Purpose of the Study:

• To explore ethical aspects and user preferences regarding supervised machine learning (ML) in digital interventions.
• To understand user acceptance and concerns about ML application in tailoring online substance use and gambling disorder treatments.

Main Methods:

• Mixed-methods approach combining qualitative focus groups (n=10) and a quantitative online questionnaire (n=157).
• Participants were recruited from an evidence-based online intervention (Jellinek Digital Self-help).
• Data analysis guided by biomedical ethics principles.

Main Results:

• Users found ML application ethically acceptable, anticipating benefits for intervention and well-being.
• Preserving user autonomy emerged as a critical factor in the acceptance of ML-driven adjustments.
• Trust in the intervention provider's integrity was high; initial qualitative concerns about data security and control were not quantitatively confirmed.

Conclusions:

• Digital intervention users exhibit limited ethical concerns regarding ML, prioritizing autonomy and privacy.
• Supervised ML can be ethically applied in digital interventions, provided user autonomy and privacy are maintained.
• User trust and perceived benefits outweigh potential ethical risks when implementing ML in digital health for substance use disorders.