Predicting suicide attempts in people with autism and co-morbid mood or psychotic disorders
- 1University of Utah
Abstract
Individuals with autism spectrum disorder (ASD) and co-occurring mood or psychotic disorders face substantially elevated suicide-attempt risk, but prior research is largely limited to population-level incidence and cross-sectional risk-factor studies, leaving unclear whether individual-level, prospective prediction is possible in this population. We conducted a retrospective prediction model study using Epic Cosmos electronic health record (EHR) data from 322,879 individuals with ASD and co-occurring mood or psychotic disorders to predict suicide attempts at the next clinical encounter using only information available at or before the preceding encounter. Five machine-learning algorithms were compared, with performance evaluated across discrimination, calibration, and clinical utility. All algorithms achieved strong discrimination using the full predictor set (AUROC 0.87-0.91). A parsimonious feature set (22 predictors) matched or exceeded full-model performance, with extreme gradient boosting (XGBoost) achieving the highest discrimination and standardized net benefit post-ablation (AUROC = 0.90; AUPRC = 0.34), though net benefit was comparable for elastic net and logistic regression. A prior suicide attempt was the strongest individual predictor. Decision-curve analysis indicated that model-guided decisions provided greater net benefit than treating all or no individuals across a broad range of clinically plausible thresholds. These findings demonstrate that near-term suicide-attempt risk can be predicted with high decision utility in this high-risk population using data already captured in routine care, suggesting a feasible basis for future EHR-based clinical decision-support tools.
Lay SummaryPeople with autism spectrum disorder who also have a co-occurring mood disorder (such as depression or bipolar disorder) or a psychotic disorder face a higher risk of suicide attempts, but clinicians have had few tools to identify which patients are at greatest risk before an attempt occurs. Using medical records from over 320,000 people, we built a predictive model that estimates future risk of suicide attempt, using information already documented in routine care. The model performed strongly and points toward a practical way to help clinicians identify at-risk patients and intervene earlier.
Related
- From psychotic symptoms to suicide probability and depressive symptoms: the mediating role of hopelessness in schizophrenia spectrum disorders — evidence from covariate-adjusted structural equation modelling
- Prefrontal responses to negative emotional cognitive stimuli are associated with future suicide attempts in patients with mood disorders: a near-infrared spectroscopy study
- Association between Multimorbidity and Suicidal Behaviors among People with Psychotic Disorders.
- Sex differences in the association between body mass index and recent suicide attempts in treatment-naïve young adults with major depressive disorder
- Cannabis use patterns, depression trajectories and suicide attempts: a 10-year longitudinal study
BibTeX
@misc{baker2026predicting,
title = {{Predicting suicide attempts in people with autism and co-morbid mood or psychotic disorders}},
author = {M. Baker and W.-Y. Lam and M. Virtosu and N. de Lacy},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.25.26364048},
url = {https://www.medrxiv.org/content/10.64898/2026.09.25.26364048v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Predicting suicide attempts in people with autism and co-morbid mood or psychotic disorders
AU - M. Baker
AU - W.-Y. Lam
AU - M. Virtosu
AU - N. de Lacy
PY - 2026
DA - 2026/09/25
PB - medRxiv
DO - 10.64898/2026.09.25.26364048
UR - https://www.medrxiv.org/content/10.64898/2026.09.25.26364048v1
AB - Individuals with autism spectrum disorder (ASD) and co-occurring mood or psychotic disorders face substantially elevated suicide-attempt risk, but prior research is largely limited to population-level incidence and cross-sectional risk-factor studies, leaving unclear whether individual-level, prospective prediction is possible in this population. We conducted a retrospective prediction model study using Epic Cosmos electronic health record (EHR) data from 322,879 individuals with ASD and co-occurring mood or psychotic disorders to predict suicide attempts at the next clinical encounter using only information available at or before the preceding encounter. Five machine-learning algorithms were compared, with performance evaluated across discrimination, calibration, and clinical utility. All algorithms achieved strong discrimination using the full predictor set (AUROC 0.87-0.91). A parsimonious feature set (22 predictors) matched or exceeded full-model performance, with extreme gradient boosting (XGBoost) achieving the highest discrimination and standardized net benefit post-ablation (AUROC = 0.90; AUPRC = 0.34), though net benefit was comparable for elastic net and logistic regression. A prior suicide attempt was the strongest individual predictor. Decision-curve analysis indicated that model-guided decisions provided greater net benefit than treating all or no individuals across a broad range of clinically plausible thresholds. These findings demonstrate that near-term suicide-attempt risk can be predicted with high decision utility in this high-risk population using data already captured in routine care, suggesting a feasible basis for future EHR-based clinical decision-support tools. Lay SummaryPeople with autism spectrum disorder who also have a co-occurring mood disorder (such as depression or bipolar disorder) or a psychotic disorder face a higher risk of suicide attempts, but clinicians have had few tools to identify which patients are at greatest risk before an attempt occurs. Using medical records from over 320,000 people, we built a predictive model that estimates future risk of suicide attempt, using information already documented in routine care. The model performed strongly and points toward a practical way to help clinicians identify at-risk patients and intervene earlier.
ER -