Using Natural Language Processing–Derived LGBTQ+ Status to Quantify Inequalities in Mental Health Services: A Secondary Analysis of Health Records
Context and challenge: LGBTQ+ people experience higher rates of mental ill health than the general population, yet little is known about their experiences of secondary mental health services because sexual orientation and gender identity are rarely recorded in NHS records. This limits research on inequalities despite evidence of poorer outcomes, including higher suicide risk.
Aims: Building on our validated AI tool that identifies LGBTQ+ status from free-text clinical notes, this project will analyse de-identified NHS mental health records to: (1) compare access, care pathways and outcomes for LGBTQ+ and non-LGBTQ+ people; (2) examine inequalities across demographic and clinical groups; and (3) generate evidence to inform equality monito