Beyond psychiatric diagnosis: detection-gap perspective on suicide decedents in a population-based cohort

A considerable proportion of individuals who die by suicide have no recorded psychiatric diagnosis, yet their psychosocial and clinical profiles remain poorly understood.

To examine whether conventional demographic, physical health, psychosocial and mental health characteristics distinguish suicide decedents with, versus without, psychiatric diagnoses in a population-based cohort.

This retrospective case-control study used UK Biobank baseline data linked with national mortality records. Suicide decedents were classified as having a recorded psychiatric diagnosis prior to death (PtD) or no recorded diagnosis (NPD). Group comparisons and logistic regression analyses examined associations between psychosocial variables and diagnostic status, including domain-specific multivariable models adjusted for age and gender.

Among suicide decedents, 44% had no recorded psychiatric diagnosis. There were no significant differences between groups in smoking-related variables, diabetes, hyperlipidaemia, body mass index, social support or social connectedness. Hypertension, as well as the use of immunosuppressants and anti-inflammatory agents, were more prevalent in the PtD group. PtD individuals exhibited higher levels of anxiety-related traits, neuroticism and healthcare utilisation. In multivariable analyses, several unadjusted comparisons were attenuated. Neuroticism, 'suffering from nerves', healthcare utilisation and lower risk-taking remained significantly associated with PtD status whereas mood-related variables, including depressed mood and anhedonia, were no longer statistically significant following adjustment.

Nearly half of suicide decedents had no recorded psychiatric diagnosis, and conventional psychosocial markers demonstrated limited discriminative value under current measurement conditions. These findings support a detection-gap framework and highlight the need for suicide prevention strategies extending beyond diagnosis-based approaches towards broader, population-level detection.