We explored the design of chatbots for depression screening, leveraging advancements in natural language processing and artificial intelligence. Using a research-through-design approach, we developed a chatbot prototype that integrates the Patient Health Questionnaire (PHQ-9), a validated depression screening tool. Our findings emphasize the importance of ethical considerations, tailored personality design, effective conversational flow, and predictability to ensure safety and user engagement in healthcare chatbot applications.
What was the context of our study?
Depression is the leading cause of disability worldwide, affecting over 300 million individuals. Despite its prevalence, it often remains underdiagnosed and untreated. Chatbots, with their ability to mimic human dialogue, offer potential solutions for addressing gaps in mental health care by engaging users in nonjudgmental interactions. However, designing chatbots for healthcare scenarios presents unique challenges due to the sensitive nature of the context and the need for reliability.
What did we do?
We employed a research-through-design methodology to create a chatbot prototype named “IGOR” for depression screening. The chatbot uses the PHQ-9 to assess depression severity and transmits results to healthcare professionals for evaluation. The development process combined a literature review of chatbot applications in healthcare, analysis of existing practices for chatbot design, and an iterative design and testing phase involving usability heuristics and participant feedback.
What did we find?
Our findings identified four critical design considerations:
- Ethical implications: Ensuring chatbot reliability and managing off-topic user inputs to safeguard patient safety.
- Personality: Selecting appropriate personality traits and tone to foster user comfort and trust.
- Conversational flow: Designing clear and efficient pathways for user interaction while minimizing redundant steps.
- Predictability: Balancing user input flexibility with the chatbot’s ability to handle unexpected scenarios without compromising safety or accuracy.
Why is this important?
Chatbots have the potential to expand access to mental health screening and support, especially for underdiagnosed populations. However, their success relies on careful attention to design features that ensure safety, user engagement, and efficacy. Our insights provide a foundation for developers and researchers to refine chatbot interfaces, making them viable tools for healthcare. By addressing these challenges, chatbots can become integral components of digital health solutions, improving access and outcomes for individuals with mental health needs.
Reference
Giunti, G., Isomursu, M., Gabarron, E., & Solad, Y. (2021). Designing Depression Screening Chatbots. *Nurses and Midwives in the Digital Age*. International Medical Informatics Association. https://doi.org/10.3233/SHTI210719