Automated screening system might detect suicidality after trauma

LOS ANGELES – About half of trauma patients report suicidal ideation within a year of their injury, and about 20% do so at any given time during that year, according to an investigation of 206 patients at Harborview Medical Center, the University of Washington’s Level I trauma center in Seattle.

A patient’s ties to assault and some of the other risk factors identified in the study could be spotted automatically by electronic medical records (EMR) systems. With a few programming tweaks, records systems can alert clinicians to those most likely to be considering suicide so preventive steps can be taken, said lead investigator Stephen S. O’Connor, Ph.D., formerly of the University of Washington, but now with the department of psychological sciences at Western Kentucky University, in Bowling Green.

"We feel that this is a really pressing need. Suicidality is endemic within the trauma population, and a lot of it is under the radar." An automated screening system might prevent suicidal people from falling through the cracks amid the tumult of trauma and follow-up care, he said at the annual conference of the American Association of Suicidology.

At some point during the 12-month investigation, 101 (49%) patients acknowledged that they had thought they’d be better off dead or had thought about hurting themselves, as assessed by the Patient Health Questionnaire-9.

About 26% admitted to those feelings while in the hospital, and the number dipped only moderately during regular, semimonthly reassessments, with 18% of the sample acknowledging suicidal thoughts at month 12.

While in the hospital, suicidal ideation was predicted by a previous mental health visit (relative risk,1.89; 95% confidence interval, 1.02-3.5; P = .04); total score on the Personal Health Questionnaire Depression Scale-8 (RR ,1.12; 95% CI 1.07-1.19; P less than .001); and general mental health function on the Short Form Health Survey-12 (RR, 0.98; 95% CI 0.97-1.00; P = .04). There was also a nonsignificant trend for injury severity.

Having children was protective (RR, 0.55; 95% CI 0.35-0.85; P = .01).

After discharge, patients who had been assaulted were most likely to be considering suicide (RR 2.03; 95% CI 1.28-3.20; P less than 0.01), as were patients who had injury-related legal proceedings (RR, 4.32; 95% CI 2.24-8.33; P less than .001), thoughts of suicide while in the hospital (RR, 3.34; 95% CI 2.28-4.89; P less than .001), and histories of PTSD (posttraumatic stress disorder) (RR, 1.72; 95% CI 1.09-2.71; P = .02).

Being older (RR, 1.03; 95% CI 1.01-1.04; P less than .001) and in pain (RR, 1.11; 95% CI 1.00-1.22; P = .05) also increased the risk. Once again, a nonsignificant trend was found for injury severity.

"Many of these are the kinds of risk factors that you could pick up from sweeping the EMR. Being assaulted by someone quickly gets an ICD code," for instance, Dr. O’Connor said.

Harborview "already has an automated screening system for PTSD; you could do the same thing with suicidal ideation. We’re moving in that direction," he said, noting that the efforts, besides saving lives, would help meet the Joint Commission’s call for better identification of suicidal patients.

Data for the analysis came from a previous PTSD intervention trial at Harborview. All of the subjects had screened positive for PTSD at 1 month, as part of the earlier work. They were about 38 years old on average, about half were women, and about half had completed high school. About a quarter of the subjects reported annual incomes of $30,000 or more.

Dr. O’Connor has no relevant disclosures. The work was funded by the National Institutes of Health.

[email protected]

LOS ANGELES – About half of trauma patients report suicidal ideation within a year of their injury, and about 20% do so at any given time during that year, according to an investigation of 206 patients at Harborview Medical Center, the University of Washington’s Level I trauma center in Seattle.

A patient’s ties to assault and some of the other risk factors identified in the study could be spotted automatically by electronic medical records (EMR) systems. With a few programming tweaks, records systems can alert clinicians to those most likely to be considering suicide so preventive steps can be taken, said lead investigator Stephen S. O’Connor, Ph.D., formerly of the University of Washington, but now with the department of psychological sciences at Western Kentucky University, in Bowling Green.

"We feel that this is a really pressing need. Suicidality is endemic within the trauma population, and a lot of it is under the radar." An automated screening system might prevent suicidal people from falling through the cracks amid the tumult of trauma and follow-up care, he said at the annual conference of the American Association of Suicidology.

At some point during the 12-month investigation, 101 (49%) patients acknowledged that they had thought they’d be better off dead or had thought about hurting themselves, as assessed by the Patient Health Questionnaire-9.

About 26% admitted to those feelings while in the hospital, and the number dipped only moderately during regular, semimonthly reassessments, with 18% of the sample acknowledging suicidal thoughts at month 12.

While in the hospital, suicidal ideation was predicted by a previous mental health visit (relative risk,1.89; 95% confidence interval, 1.02-3.5; P = .04); total score on the Personal Health Questionnaire Depression Scale-8 (RR ,1.12; 95% CI 1.07-1.19; P less than .001); and general mental health function on the Short Form Health Survey-12 (RR, 0.98; 95% CI 0.97-1.00; P = .04). There was also a nonsignificant trend for injury severity.

Having children was protective (RR, 0.55; 95% CI 0.35-0.85; P = .01).

After discharge, patients who had been assaulted were most likely to be considering suicide (RR 2.03; 95% CI 1.28-3.20; P less than 0.01), as were patients who had injury-related legal proceedings (RR, 4.32; 95% CI 2.24-8.33; P less than .001), thoughts of suicide while in the hospital (RR, 3.34; 95% CI 2.28-4.89; P less than .001), and histories of PTSD (posttraumatic stress disorder) (RR, 1.72; 95% CI 1.09-2.71; P = .02).

Being older (RR, 1.03; 95% CI 1.01-1.04; P less than .001) and in pain (RR, 1.11; 95% CI 1.00-1.22; P = .05) also increased the risk. Once again, a nonsignificant trend was found for injury severity.

"Many of these are the kinds of risk factors that you could pick up from sweeping the EMR. Being assaulted by someone quickly gets an ICD code," for instance, Dr. O’Connor said.

Harborview "already has an automated screening system for PTSD; you could do the same thing with suicidal ideation. We’re moving in that direction," he said, noting that the efforts, besides saving lives, would help meet the Joint Commission’s call for better identification of suicidal patients.

Data for the analysis came from a previous PTSD intervention trial at Harborview. All of the subjects had screened positive for PTSD at 1 month, as part of the earlier work. They were about 38 years old on average, about half were women, and about half had completed high school. About a quarter of the subjects reported annual incomes of $30,000 or more.

Dr. O’Connor has no relevant disclosures. The work was funded by the National Institutes of Health.

[email protected]

LOS ANGELES – About half of trauma patients report suicidal ideation within a year of their injury, and about 20% do so at any given time during that year, according to an investigation of 206 patients at Harborview Medical Center, the University of Washington’s Level I trauma center in Seattle.

A patient’s ties to assault and some of the other risk factors identified in the study could be spotted automatically by electronic medical records (EMR) systems. With a few programming tweaks, records systems can alert clinicians to those most likely to be considering suicide so preventive steps can be taken, said lead investigator Stephen S. O’Connor, Ph.D., formerly of the University of Washington, but now with the department of psychological sciences at Western Kentucky University, in Bowling Green.

"We feel that this is a really pressing need. Suicidality is endemic within the trauma population, and a lot of it is under the radar." An automated screening system might prevent suicidal people from falling through the cracks amid the tumult of trauma and follow-up care, he said at the annual conference of the American Association of Suicidology.

At some point during the 12-month investigation, 101 (49%) patients acknowledged that they had thought they’d be better off dead or had thought about hurting themselves, as assessed by the Patient Health Questionnaire-9.

About 26% admitted to those feelings while in the hospital, and the number dipped only moderately during regular, semimonthly reassessments, with 18% of the sample acknowledging suicidal thoughts at month 12.

While in the hospital, suicidal ideation was predicted by a previous mental health visit (relative risk,1.89; 95% confidence interval, 1.02-3.5; P = .04); total score on the Personal Health Questionnaire Depression Scale-8 (RR ,1.12; 95% CI 1.07-1.19; P less than .001); and general mental health function on the Short Form Health Survey-12 (RR, 0.98; 95% CI 0.97-1.00; P = .04). There was also a nonsignificant trend for injury severity.

Having children was protective (RR, 0.55; 95% CI 0.35-0.85; P = .01).

After discharge, patients who had been assaulted were most likely to be considering suicide (RR 2.03; 95% CI 1.28-3.20; P less than 0.01), as were patients who had injury-related legal proceedings (RR, 4.32; 95% CI 2.24-8.33; P less than .001), thoughts of suicide while in the hospital (RR, 3.34; 95% CI 2.28-4.89; P less than .001), and histories of PTSD (posttraumatic stress disorder) (RR, 1.72; 95% CI 1.09-2.71; P = .02).

Being older (RR, 1.03; 95% CI 1.01-1.04; P less than .001) and in pain (RR, 1.11; 95% CI 1.00-1.22; P = .05) also increased the risk. Once again, a nonsignificant trend was found for injury severity.

"Many of these are the kinds of risk factors that you could pick up from sweeping the EMR. Being assaulted by someone quickly gets an ICD code," for instance, Dr. O’Connor said.

Harborview "already has an automated screening system for PTSD; you could do the same thing with suicidal ideation. We’re moving in that direction," he said, noting that the efforts, besides saving lives, would help meet the Joint Commission’s call for better identification of suicidal patients.

Data for the analysis came from a previous PTSD intervention trial at Harborview. All of the subjects had screened positive for PTSD at 1 month, as part of the earlier work. They were about 38 years old on average, about half were women, and about half had completed high school. About a quarter of the subjects reported annual incomes of $30,000 or more.

Dr. O’Connor has no relevant disclosures. The work was funded by the National Institutes of Health.

[email protected]