Using Conversational AI to Facilitate Mental Health Assessments and Improve Clinical Efficiency Within Psychotherapy Services: Real-World Observational Study (Preprint)

Abstract

BACKGROUND Most mental health service providers face the challenge of increasing demand in the absence of increases in funding or staffing. To overcome this supply-demand imbalance, providers need to increase efficiencies to cope with the demand. OBJECTIVE Here, we test whether artificial intelligence (AI) enabled solutions can enable mental health practitioners to use their time more efficiently, and thus reduce strain on the service and improve patient outcomes. METHODS In this study, we focus on the usage of an AI solution (Limbic Access) in the referral and assessment process in UK’s national health service (NHS) first-line psychotherapy service. Data was collected from 9 Improving Access to Psychological Therapies (IAPT) services across England from 64,862 patients. RESULTS We show that the use of this AI solution improves clinical efficiency by reducing the time clinicians spend on mental health assessments. Furthermore, we find improved outcomes for patients using the AI solution in a number of key metrics, such as reduced wait times, re- duced dropout rates, improved allocation to accurate treatment pathways and, most importantly, improved recovery rates. When investigating the mechanism by which the AI solution achieved these improvements, we find that the provision of clinically relevant information ahead of a clinical assessment was critical for these observed effects. CONCLUSIONS Our results emphasise the utility of using AI solutions to support the mental health workforce and highlight that AI solutions can increase efficiencies and in parallel improve mental healthcare for patients.

AI-Enabled Clinical Decision Support Tools for Mental Healthcare: A Product Review

“…These inputs trigger more detailed questions to create a thorough symptom profile for each user. The obtained data is assessed in connection with known mental disorders, subsequently suggesting potential diagnoses (Limbic AI, 2023;Rollwage et al, 2023).…”