A Virtual Counselor for Breast Cancer Genetic Counseling: Adaptive Pedagogy Leads to Greater Knowledge Gain

Title of the Paper

A Virtual Counselor for Breast Cancer Genetic Counseling: Adaptive Pedagogy Leads to Greater Knowledge Gain

Paper Information

• Field of Study: Intersection of information technology and health education, particularly virtual assistant-based genetic counseling for breast cancer
• Keywords: Virtual conversational agent, educational agent, intelligent tutoring system, genetic counseling, risk communication, health education

Research Background and Issues

• Background:
  With the significant increase in the demand for genetic counseling, genetics and genomics play a crucial role in breast cancer prevention and treatment. However, due to the shortage of qualified genetic counselors in the U.S. and limitations in health literacy, many individuals lack access to genetic counseling services. Substantial challenges remain in risk communication and behavior promotion within breast cancer genetic counseling. Additionally, existing information processing models have not been effectively applied to the field of genetic counseling.

• Identified Issues:

  1. Complex genetic risk information is often difficult for the general public to understand, especially for individuals with low health literacy.
  2. Merely ensuring that audiences receive and understand genetic risk information does not directly lead to ideal behavior changes, such as adhering to screening guidelines or taking preventive measures.
  3. When faced with "personalized needs," genetic counseling lacks sufficient adaptive adjustments.

• Research Significance:
  Efficient breast cancer genetic counseling not only helps improve risk comprehension but also increases screening participation rates and reduces mortality. Automated and personalized genetic counseling can bridge existing service gaps, particularly in situations where face-to-face counseling is unavailable.

• Research Motivation and Related Work:
  Virtual conversational agents (ECA) with multidimensional adaptive adjustments can simulate face-to-face interactions to provide personalized health counseling with educational functions, overcoming the limitations of traditional genetic counseling.

Solution

• Proposed Method:
  The authors designed a virtual genetic counselor system that uses ECA to simulate face-to-face genetic counseling and dynamically adjust its educational content. The system's adaptive adjustment mechanisms include dynamic assessment of user knowledge states, adaptive conversational strategies based on user information processing preferences, and personalized content selection based on health literacy levels.

• Innovations:

  1. Introduced unique dimensions of adaptive adjustments, such as health literacy and information processing patterns, into the traditional intelligent tutoring system architecture for breast cancer genetic counseling.
  2. Designed a multidimensional dynamic educational mechanism by integrating information processing theories (e.g., Heuristic-Systematic Model, HSM) to enhance risk comprehension and behavior change outcomes.
  3. The interactive experience provided by the ECA is more intuitive and suitable for users with low health literacy compared to traditional graphical interfaces or chatbots.

• Implementation Steps and Technology:

  1. Architecture Design:
     • Includes an intelligent tutoring system with domain knowledge models, user models, educational models, and an ECA interface.
  2. Personalized Counseling Process:
     • Dynamically assesses users' knowledge states and adjusts educational content through Q&A interactions.
     • Provides visual or narrative risk explanations, selecting the appropriate explanatory method based on user preferences.
  3. Technical Support:
     • Developed the system using the Unity 3D engine, incorporating multimodal interaction features (e.g., gestures and voice), dynamic content generation, and predefined dialogue templates.

Research Outcomes

• Main Results:

  1. Compared to the non-adaptive version and control group, the adaptive virtual counselor significantly improved participants' understanding of breast cancer genetic risks and prevention guidelines.
  2. Participants perceived the adaptive virtual counselor as more satisfactory, rating it as providing tailored information based on individual needs.

• Advantages Over Existing Solutions:

  • The adaptive virtual counselor dynamically adjusts content and provides a highly personalized health education experience, which traditional genetic counseling cannot achieve with such precision.
  • It performs particularly well for users with lower health literacy and higher genetic risk.

• Experimental or Evaluation Results:

  • Among 30 female participants in the experiment, users under the adaptive condition showed significant improvements in breast cancer genetic knowledge scores (average increase of 6.9 points).
  • In terms of effectiveness and adaptive cognition evaluations, the adaptive version of the virtual counselor scored significantly higher than other conditions.
  • Participants across all conditions exhibited high behavioral intentions and satisfaction, but the adaptive version was preferred due to higher satisfaction with user preferences.

• Limitations and Future Directions:

  • Limitations:

    1. Small sample size, with a low proportion of high-risk and low health literacy users.
    2. Single-session experimental design cannot verify long-term knowledge retention or behavior change.

  • Future Directions:

    1. Expand to multiple long-term interventions to evaluate the impact of sustained behavior changes on breast cancer prevention.
    2. Design more refined emotional interaction mechanisms to address anxiety issues in high-risk users, focusing on psychological support needs.
    3. Extend this technology to other health education and remote counseling fields, such as cardiovascular disease risk education or vaccine counseling.

This study demonstrates the potential of ECA in health education, particularly in the multidimensional dynamic adjustment model for automated genetic risk counseling. It holds significant implications for improving health literacy and promoting behavior change.