<b>Using Speech Analyses for the Detection of Suicide Risk and Symptom Relapse in Eating Disorders: Protocol for a Longitudinal Cohort Study</b>

<p dir="ltr"><b>Background:</b></p><p dir="ltr">Rates of suicidality and symptom relapse are unacceptably high amongst young people with eating disorders (EDs), yet current approaches inadequately predict when individuals are most at risk. Despite national priorities emphasising early intervention and dynamic monitoring, few studies have captured real-time indicators of increased risk during critical periods following discharge from acute medical care. Speech has shown promise for identifying suicidality in other psychiatric cohorts but remains underexplored in ED populations.</p><p><br></p><p dir="ltr"><b>Objectives:</b></p><p dir="ltr">The primary study objectives are to examine whether acoustic features of speech can predict momentary suicide risk and eating disorder (ED) symptom severity among 12–25-year-olds with EDs, to identify which acoustic voice markers are most strongly associated with these outcomes over time, and to assess the accuracy of these markers in classifying suicide risk and ED severity in real time. We also aim to explore whether these relationships are influenced by demographic and clinical factors such as age, sex, ED diagnosis, age of illness onset, and co-occurring mental health conditions. Additional exploratory objectives include investigating potential mechanisms that may underlie associations between acoustic markers, ED symptoms, and suicidality.</p><p><br></p><p dir="ltr"><b>Methods:</b></p><p dir="ltr">A four-month prospective cohort study will be conducted over three in-person visits (baseline, two months, four months), with additional smartphone-based ecological momentary assessments (EMAs) embedded over the first month. Speech recordings, suicide risk assessments, and ED symptom severity ratings will be performed at all in-person visits and across all EMA prompts. We aim to recruit 192 adolescents and young adults (aged 12–25) with a diagnosis of anorexia nervosa, bulimia nervosa, binge-eating disorder, or another specified eating disorder.</p><p><br></p><p dir="ltr"><b>Results and Conclusions:</b></p><p dir="ltr">This study is a two-year project, with work commencing in March 2025. Study recruitment is planned to commence in October 2025, with data collection scheduled to begin in November 2025 and continue through to November 2026. Given the limited research on speech acoustics in EDs, findings from this study may lay the groundwork for future studies and interventions that use speech as a passive, scalable, real-time indicator of psychological distress in young people with diverse ED presentations.</p>