Models that predict patients most at risk of suicide may be less accurate at predicting this risk in people who identify as Black or American Indian/Alaskan Native, suggests a study published today in JAMA Psychiatry.
Because clinical prediction models frequently rely on health records, they may be less accurate for patients who experience disparities in access to health care, wrote R. Yates Coley, Ph.D., of the Kaiser Permanente Washington Health Research Institute and colleagues. The authors noted that current suicide prediction models are accurate at the population level, but research is lacking on their accuracy for racial and ethnic subgroups.
Coley and colleagues gathered data on outpatient visits to a mental health specialist, including health record and insurance billing information, from seven health systems between January 1, 2009, and September 30, 2017. Suicide predictors used in the analysis included demographic characteristics (age, sex, race, ethnicity, and insurance type), comorbidities, mental and substance use diagnoses, dispensed psychiatric medications, prior suicide attempts, prior mental health encounters (including at emergency departments and hospitalizations), and Patient Health Questionnaire-9 (PHQ-9) responses. The researchers ran two prediction models for suicide deaths that occurred within 90 days after an outpatient visit. Because patients self-reported their race/ethnicity information during clinic visits, the information may not have been recorded if the patient had too few encounters during which it was collected, they did not identify with any of the categories offered, or clinic staff did not make inquiries about the information.
The analysis included nearly 14 million visits by 1.4 million patients, and 768 suicide deaths were observed within 90 days of 3,143 visits. Suicide rates were highest for patients who did not have a race or ethnicity recorded, followed by patients who were Asian or White. Overall the models predicted who would die by suicide with about 82% accuracy when including the entire data set at the population level. Within racial/ethnic subgroups, the model was most accurate at predicting suicide among for White, Hispanic, and Asian patients, and least accurate at predicting suicide in Black and American Indian/Alaskan Native patients, as well as those with unrecorded race or ethnicity.
“Health records data may poorly predict suicide death in some racial/ethnic groups for several reasons,” the authors wrote. These include barriers to affordable, culturally competent mental health care; practitioner bias and institutionalized discrimination that lower the likelihood that underrepresented populations will receive a mental health diagnosis or treatment; and the potential misclassification of suicide deaths as unintentional or accidental, or vice versa.
“[P]otential benefits and harms of using a prediction model within particular populations must be considered in the context of existing health inequities,” the authors concluded. “BIPOC [Black, Indigenous, and people of color] populations already face significant barriers to accessing mental health care and, as a result, have poorer outcomes. In this context, deploying a prediction model that provides less benefit to already underserved populations will widen this care gap.”
For related information, see the Psychiatric Services article “Reconciling Statistical and Clinicians’ Predictions of Suicide Risk.”
(Image: iStock/AzmanL)
Don't miss out! To learn about newly posted articles in Psychiatric News, please sign up here.
