Closing the accessibility gap to mental health treatment with a conversational AI-enabled self-referral tool

Accessing mental health care can be challenging, and minority groups often face additional barriers. This study investigates whether digital tools can enhance equality of access to mental health treatment. We evaluated a novel AI-enabled self-referral tool (a chatbot) designed to make entry to mental health treatment more accessible in a real-world setting. In a multi-site observational study, data were collected from 129,400 patients who referred to 28 separate NHS Talking Therapies services across England. Our results indicate that the tool led to a 15% increase in total referrals, which was significantly larger than the 6% baseline increase observed in matched services using traditional self-referral methods during the same time period. Importantly, the tool was particularly effective for minority groups, which included non-binary (235% increase), bisexual (30% increase), and ethnic minority individuals (31% increase). This paints a promising picture for the use of AI chatbots in mental healthcare and suggests they may be especially beneficial for demographic groups that experience barriers to accessing treatment in the traditional care systems. To better understand the reasons for this disproportional benefit for minority groups, we used thematic analysis and Natural Language Processing (NLP) models to evaluate qualitative feedback from 42,332 individuals who referred through the AI-enabled tool. We found that the tool's human-free nature and its ability to improve the perceived need for treatment were the main drivers for improved diversity. These findings suggest that AI-enabled chatbots have the potential to increase accessibility to mental health services for all, and to alleviate barriers faced by disadvantaged populations. The results have important implications for healthcare policy, clinical practice, and technology development.