AI Chatbot for Mental Health
The experiment saw an increased rate of referrals among those less likely to seek help
A major challenge around mental wellness across the world, and particularly so in India, relates to the lack of awareness around how and when to seek help. Mental health experts blame this on the social stigma and ready availability of support, though a recent experiment by the British authorities around artificial intelligence (AI) could help resolve this to some extent.
The British Medical Association said demand for mental health services grew post the covid-19 pandemic to 4.6 million referrals in 2022 and credit an AI-led chatbot for the referrals via the country’s National Health Service. The spike was visible across hitherto underrepresented groups that are less likely to seek health, says an article in the MIT Tech Review.
A non-judgmental approach to mental health
The chatbot has been created by AI company Limbic, which was evaluated by Nature Medicine in a recent published report. It highlights the impact of the chatbot called Limbic Access on the referrals to the mental health referrals including therapies for anxiety and disorder. On its part, the AI company was seeking to probe if AI could lower the barrier to healthcare.
The study examined data from over 1.29 lakh people visiting websites to refer themselves to various talking therapy services across England. Half of these had used the chatbot on their website while the other half used data-collecting methods such as web forms. The referrals from the Limbic chatbot rose 15% during the study period compared to 6% rise elsewhere.
On deeper analysis, the data also indicated that minority groups including ethnic and sexual ones, grew significantly where the chatbot was available – a rise of 179% among nonbinary people, 39% for Asians and 40% for Black patients. Given India’s diverse social structure, such an intervention could drastically enhance referrals around mental wellness.
Proves that AI will enhance existing jobs
The report also indicated that in England, higher referrals did not increase waiting periods or cause reduced clinical assessments. It was found that the data collected by the chatbot reduced the total time human clinicians required to spend assessing patients while automatically improving the quality of assessments, which in turn, freed up resource time.
Yet another instance of AI actually not taking up human jobs but enhancing their productivity. Experts quoted by the MIT article noted that chatbots appear to collect better quality data than those from static web forms, which could further ease the ability of wellness services to reach the target population.
Here’s how the process works
The chatbot-enabled websites had a pop-up explaining the purpose of the robotic assistant and its design to help people access psychological support. Based on a set of evidence-based screening, the chatbot asks supporting questions around existing issues, past diagnoses etc. Thereafter, it follows-up with queries to measure symptoms of the patients’ problems.
This data is then used to create a detailed referral that is shared with the electronic record system that the government uses. The next stage involves a human professional accessing the data and reaching out to the person within a couple of days to make a detailed assessment and then start the treatment process.
Limbic says its chatbot combines multiple AI models including natural language processing to assess the user’s typed responses and provide empathetic answers, followed by probabilistic models that use data to tailor its own responses to the patient. The report noted that these models had classified eight common mental-health issues with 93% accuracy.
