An AI chatbot for talking therapy referrals

Jacqueline Sin

DOI: https://doi.org/10.1038/s41591-023-02773-y

IF: 82.9

2024-02-06

Nature Medicine

Abstract:An AI-enabled chatbot increased self-referrals to psychological therapies for common mental health disorders; but further research is needed to ensure that better access translates into high-quality treatment experiences and outcomes for everyone.

biochemistry & molecular biology,cell biology,medicine, research & experimental

What problem does this paper attempt to address?

The problem this paper attempts to address is: how to improve the self-referral rate of patients with common mental health disorders (such as anxiety and depression) through artificial intelligence (AI) chatbots, especially for minority and non-binary gender groups. Specifically, the paper explores the effectiveness of a personalized AI chatbot named Limbic in a multi-site study, aiming to increase the self-referral rate of these patients, thereby improving their access to psychological therapy.

### Main Issues:
1. **Improving Self-Referral Rate**: Many people with anxiety and depression are reluctant to seek treatment proactively due to various reasons (such as feeling down, social withdrawal, etc.). Traditional referral pathways (such as referrals by general practitioners) face issues of insufficient time and diagnostic skills, making it difficult for patients to receive timely treatment. Therefore, improving the self-referral rate is crucial.
2. **Reducing Treatment Barriers for Minority and Non-Binary Gender Groups**: These groups face more barriers in accessing mental health services, such as underdiagnosis, misdiagnosis, and cultural barriers. The paper specifically focuses on whether Limbic can effectively improve the self-referral rate of these groups.
3. **Ensuring High-Quality Treatment Experience and Outcomes**: While increasing the self-referral rate, further research is needed to ensure that these patients can receive a high-quality treatment experience and good treatment outcomes.

### Research Background:
- **High Prevalence of Common Mental Health Disorders**: Anxiety and depression are common mental health issues, affecting about 1/5 of the global population.
- **Effectiveness of Cognitive Behavioral Therapy (CBT)**: CBT is recommended as the first-line treatment for mild to moderate cases.
- **Limitations of Existing Referral Pathways**: Traditional referral pathways face issues of insufficient time and diagnostic skills, making it difficult for patients to receive timely treatment.

### Research Methods and Results:
- **Study Design**: Habicht et al. conducted a multi-site study involving nearly 130,000 individuals, comparing the effectiveness of the Limbic chatbot and standard web forms in promoting self-referral.
- **Main Findings**:
- Limbic increased the self-referral rate by 15%, while standard web forms only increased it by 6%.
- Limbic was particularly effective in increasing the self-referral rate of minority (29%) and non-binary gender groups (179%).
- Patients using Limbic completed self-referrals more quickly and had a better understanding of their treatment needs.

### Conclusion:
- **Positive Impact**: Limbic successfully increased the self-referral rate, especially among minority and non-binary gender groups, helping to reduce the barriers these groups face in accessing mental health services.
- **Future Research Directions**: Further research is needed to ensure that the increased self-referral rate translates into high-quality treatment experiences and good treatment outcomes. Additionally, consideration is needed on how to train human therapists to collaborate with AI-assisted tools to provide psychological therapy that meets different cultural and individual needs.