A brief scalable suicide risk assessment tool accurately predicts suicide risk in patients with serious mental illness (SMI), a new population-based study shows.
The 17-question Oxford Mental Illness and Suicide Tool (OxMIS) assessment is designed to predict 12-month suicide risk in people with schizophrenia spectrum disorders and bipolar disorder based on risk factors such as familial traits, antisocial traits, and information about self-harm.
Amir Sariaslan, PhD
“We have demonstrated the clinical utility of OxMIS in two separate studies and countries. As with any clinical risk prediction tool, it will not improve outcomes unless coupled with effective interventions,” lead investigator Amir Sariaslan, PhD, a senior research fellow in psychiatric epidemiology at the University of Oxford, Oxford, United Kingdom, told Medscape Medical News.
The findings were published online April 18 in Translational Psychiatry.
Twice Validated
Sariaslan and his team originally developed and validated the OxMIS in a cohort of 75,000 people with SMI in Sweden. Recognizing the lack of externally validated prognostic models in the mental health field, the team wanted to validate the instrument in a new, population-based sample in Finland.
The investigators accessed information about patient diagnosis and treatment from the Finnish Care Register for Health Care, which contains de-identified information for all individuals between ages 15 and 65 years diagnosed with an SMI between January 1, 1996, and December 31, 2017.
They included 137,000 patients with somatic symptom disorder or bipolar disorder for a total of more than 5 million episodes of inpatient or outpatient treatment. Investigators linked the cohort to the Causes of Death Register to identify those who had died by suicide within 12 months of an index treatment episode, which investigators randomly selected for each person.
The investigators found that 1475 individuals in the sample died by suicide within 1 year of their index episode (1.1%).
Each patient was assigned a clinical suicide risk score based on their clinical information, familial traits, prescription information, and comorbid conditions. Using OxMIS, the investigators found that the instrument accurately predicted suicide with an area under the curve of 0.70.
In other words, in 70% of the instances where the investigators randomly selected two people from the sample, one of whom died by suicide and the other of whom did not, the individual who died by suicide had a higher OxMIS risk score.
The investigators note the model overestimated the risk for patients who were at extremely high risk for suicide (those with a predicted suicide risk of
> 5%). “In our complementary sensitivity analysis, we observed improved calibration in these patients when we assigned them a suicide risk prediction of no more than 5%,” they write.
Sariaslan said that the findings highlight the importance of safety planning interventions. “It is also essential to remember that OxMIS is not intended to replace clinical decision-making, but rather to support it,” he said.
As to whether the tool could be used in other populations, such as the United States, Sariaslan said, “there is no good evidence that the contribution of risk factors to suicide in this population is different in the US than in northern Europe, so there is no a priori reason to have to do multiple external validations before it can be used for research or clinical purposes.”
One Size Does Not Fit All
Commenting on the study for Medscape Medical News, Ronald Kessler, PhD, McNeil Family Professor, Department of Health Care Policy at Harvard Medical School in Boston, Massachusetts, said that he’d be “surprised” if OxMIS was adopted in the United States because there is already an existing tool that is “slightly more accurate,” which he helped develop.
Ronald Kessler, PhD
“In addition, when we start thinking about uses for such scales, it becomes clear that different scales should be used for different segments of the population, depending on intervention options,” Kessler said.
“So, for example, a different scale would probably be optimal in deciding how to manage psychiatric inpatients in the transition back to the community after hospital discharge than [it would be], say, in deciding how to respond to suicidality among patients presenting at an emergency department. No one scale will fit for all the scenarios in which prediction is desired,” he added.
The study was funded by the Academy of Finland. Kessler receives funding from the National Institute of Mental Health, Department of Defense, and Veterans Administration to develop suicide prediction models. Sariaslan has no disclosures to report.
Translational Psychiatry. Published online April 18, 2023. Full text
For more Medscape Psychiatry news, join us on Twitter and Facebook
Follow Medscape on Facebook, Twitter, Instagram, and YouTube
Source: Read Full Article