In the Journals

Algorithm more accurate than clinical assessment in identifying veterans at risk for suicide

A suicide-risk algorithm that used data from the Veterans Health Administration health system electronic medical records accurately identified patients at very high risk for suicide, many of whom had not been identified by clinicians.

To develop the algorithm, John F. McCarthy, MPH, PhD, director of the Serious Mental Illness Treatment Resource and Evaluation Center, Office of Mental Health Operations, Department of Veterans Affairs in Washington D.C., and colleagues examined the manner of death and predictors of suicide and other deaths across the Veterans Health Administration patient population from 2009 to 2011. They divided the patient population in half, using one half to develop the predictive model and the other half to test it.

John F. McCarthy, MPH, PhD

John F. McCarthy

The study cohorts included 3,180 patients who committed suicide and 1,056,004 control patients.

One-third of the patients considered to have the highest predicted suicide risk based on the algorithm had been clinically identified as at-risk for suicide, suggesting that the model was more sensitive than the typical method of identifying patients based on information obtained during clinical visits.

“This is valuable, because it gives the VA more extensive information about suicide risk,” study researcher Michael Schoenbaum, PhD, of the NIMH, said in a press release. “If the VA can identify small groups of people with a particularly high risk of suicide, then they can target enhanced prevention and treatment services to these highest-risk individuals.”

Thomas Insel

Thomas Insel

The study findings are particularly promising because they draw on data available to all large health care systems, according to Thomas Insel, MD, director of the NIMH.

“It’s particularly encouraging that these analyses use the types of data available to any large health care system,” he said in the release. “These methods could help us prevent civilian as well as veteran suicides.”

Disclosure: Please see the full study for a list of authors’ relevant financial disclosures.

A suicide-risk algorithm that used data from the Veterans Health Administration health system electronic medical records accurately identified patients at very high risk for suicide, many of whom had not been identified by clinicians.

To develop the algorithm, John F. McCarthy, MPH, PhD, director of the Serious Mental Illness Treatment Resource and Evaluation Center, Office of Mental Health Operations, Department of Veterans Affairs in Washington D.C., and colleagues examined the manner of death and predictors of suicide and other deaths across the Veterans Health Administration patient population from 2009 to 2011. They divided the patient population in half, using one half to develop the predictive model and the other half to test it.

John F. McCarthy, MPH, PhD

John F. McCarthy

The study cohorts included 3,180 patients who committed suicide and 1,056,004 control patients.

One-third of the patients considered to have the highest predicted suicide risk based on the algorithm had been clinically identified as at-risk for suicide, suggesting that the model was more sensitive than the typical method of identifying patients based on information obtained during clinical visits.

“This is valuable, because it gives the VA more extensive information about suicide risk,” study researcher Michael Schoenbaum, PhD, of the NIMH, said in a press release. “If the VA can identify small groups of people with a particularly high risk of suicide, then they can target enhanced prevention and treatment services to these highest-risk individuals.”

Thomas Insel

Thomas Insel

The study findings are particularly promising because they draw on data available to all large health care systems, according to Thomas Insel, MD, director of the NIMH.

“It’s particularly encouraging that these analyses use the types of data available to any large health care system,” he said in the release. “These methods could help us prevent civilian as well as veteran suicides.”

Disclosure: Please see the full study for a list of authors’ relevant financial disclosures.