Increasing self-harm and suicide rates among Black youth are an urgent public health concern. Historically, Black youth had lower rates of self-harm and suicide compared with their peers from other racial and ethnic groups; however, this trend has shifted. Between 1991 and 2007, self-reported suicide attempts among Black youth rose by 73%, while rates among white youth declined. Suicide attempts by Black youth aged 14-18 years also increased from 2009 to 2019. Suicide is currently the third leading cause of death among Black youth aged 10 to 24 years, with rates increasing from 8.2 to 10.7 per 100 000 between 2018 and 2023. These troubling trends highlight the critical need for effective prevention strategies among Black youth.
One strategy for suicide prevention involves the use of machine-learning algorithms applied to electronic health record (EHR) data to identify individuals at higher risk for suicide and self-harm. Integrating these algorithms into EHR systems facilitates early detection and intervention for preventive care. Self-harm prediction models have been developed, validated, and implemented across multiple health care settings, demonstrating their reliability and effectiveness in identifying high-risk individuals, including a model presented by Simon et al. External validation studies have shown the model performs adequately among adolescents, in a large multiracial adult population, and in American Indian and Alaskan Native populations, whereas another study raised concerns about racial disparities in predicting suicide death. However, the efficacy of existing models for predicting self-harm risk among Black youth has not been examined, to our knowledge. To address this gap, we validated the self-harm prediction model presented by Simon et al. in more than 380 000 Black youth from 8 health systems.