This article is more than 5 years old. Information may no longer be current.

EHR data accurately predicts suicide risk

Data from electronic health records may be utilized to predict future risk for suicidal behavior, according to recent findings.

“We previously showed that data commonly available in EHRs can accurately predict future domestic abuse diagnoses an average of 2 years in advance. Using this information to support early detection of individuals at high risk for suicide and self-inflicted injury could help prevent significant morbidity and mortality and ensure that at-risk patients receive the professional care they need,” Yuval Barak-Corren, MS, of Boston Children’s Hospital Informatics Program, and colleagues wrote.

To determine if longitudinal historical data from EHR systems can predict future risk for suicidal behavior, researchers analyzed EHR data from a large health care database for inpatient and outpatient visits from 1998 to 2012. Individuals with three or more visits were included (n = 1,728,549). Suicidal behavior was defined by expert clinician consensus review of 2,700 narrative EHR notes from 520 individuals, supplemented by state death certificates.

Overall, 1.2% of the cohort met the case definition for suicidal behavior.

The model achieved 33% to 45% sensitivity, 90% to 95% specificity, and early prediction (3 to 4 years in advance) of future suicidal behavior.

Opioid abuse, personality and bipolar disorders, infections (OR = 6.1; 95% CI, 4.5-8.1), alveolitis of the jaw (OR = 5.8; 95% CI, 3.7-9.4), osteomyelitis (OR = 4.85; 95% CI, 2.9-8), cellulitis (OR = 4.6; 95% CI, 3.9-5.4) and numerous codes associated with wounds and injuries were the strongest predictors of suicidal risk.

“These findings suggest that the vast quantities of longitudinal data accumulating in electronic health information systems present a largely untapped opportunity for improving medical screening and diagnosis,” the researchers wrote. “Beyond the direct implications for prediction of suicide risk, this general approach has far-reaching implications for the automated screening of a wide range of clinical conditions for which longitudinal historical information may be beneficial for estimating clinical risk.” – by Amanda Oldt

Disclosure: Barak-Corren reports no relevant financial disclosures. Please see the full study for a list of all authors’ relevant financial disclosures.