University of New Mexico Study Reveals Machine Learning Detects Hidden Self-Harm Histories in Veterans
A study led by the University of New Mexico School of Medicine has utilized machine learning to uncover hidden self-harm histories in veterans, significantly surpassing traditional diagnosis codes. The research analyzed electronic health records of over 1.3 million patients from the Veterans Health Administration. The machine learning method identified self-harm in 7.9% of patients, compared to just 1.85% detected through diagnosis codes. This approach, combining machine learning with expert chart review and statistical calibration, highlights a substantial gap in visibility between documented self-harm and what is captured in diagnosis codes. The findings, published in the Journal of Medical Internet Research, suggest that relying solely on diagnosis codes may lead to underestimating the need for mental health services. 
