Conversational Agents Supporting Self-Management in People With a Chronic Disease: Systematic Review

Area of Science:

• Digital Health
• Behavioral Psychology
• Biomedical Engineering

Background:

• Conversational agents (CAs) offer scalable, personalized self-management support for chronic diseases.
• Interdisciplinary approaches integrating biomedical, behavioral, and technological aspects are crucial for CA effectiveness.
• Structured evaluations are essential for CAs in chronic disease self-management.

Purpose of the Study:

• To examine the design and evaluation of CAs for chronic disease self-management.
• To identify characteristics, behavioral change techniques, and evaluation methods of existing CAs.
• To guide future research and inform the design of effective CA interventions.

Main Methods:

• Systematic literature search in PubMed and Embase (Jan 2018 - Apr 2024).
• Inclusion of full-text articles on CA efficacy/effectiveness for adult chronic disease self-management.
• Data extraction using the behavioral intervention technology model and CONSORT-EHEALTH checklist; risk of bias assessment.

Main Results:

• 25 studies included, mostly text-based, rule-based CAs via mobile apps.
• Predominant chronic diseases targeted were diabetes and cancer.
• Common behavior change techniques included 'shaping knowledge,' 'feedback and monitoring,' 'natural consequences,' and 'associations,' but reporting was limited.

Conclusions:

• Transparent intervention descriptions, rigorous methodologies, and standardized reporting are needed.
• Integration of AI-driven personalization and focus on healthcare implementation are key for future CA research.
• Consistent use of validated scales and standardized taxonomy will advance the field.