Suicide is a pervasive health problem — and it may be because clinicians’ still lack effective diagnostic tools, according to new study out of Australia. The findings have prompted medical professionals to review how they allocate suicide prevention resources.

The research team, led by clinical psychiatrist Matthew Large, analyzed data from 37 studies published all over the world. These studies looked at risk factors that caused low- and high-risk patients to commit suicide. They found that there was not a single reliable method of predicting who would take their own life — half of all suicides occurred in supposedly low-risk patients, while 95 percent of high-risk patients did not commit suicide. What's more, researchers found a lack of progress in accurately predicting suicide in the last 50 years.

“It is widely assumed that the care of psychiatric patients can be guided by a mental health professional’s estimate of suicide risk and by using patient characteristics to define high-risk patients,” Large explained in a press release. “However, the reliability of categorizing suicide risk remains unknown.”

Suicide risk assessments predicted the act only a little better than pure chance, according to the study. And a more complex method of prediction, one that considers multiple factors into account, didn’t have an advantage over methods that focus on a single factor.

“Much of what happens when a mentally distressed person presents to a hospital depends on suicide risk assessment, based on a whole range of risk factors,” Large said. “Lower-risk patients can be denied treatment, while some high-risk patients get hospitalized, sometimes against their wishes, based on an inaccurate risk assessment.”

Suicide is the tenth leading cause of death in the United States, but the second among 15-24 year olds. It takes 38,000 lives ever year. Depression, which affects 20 to 25 percent of Americans, and some other mental illnesses are major risk factors for suicide, so doctors pay special attention to this population when assessing suicide risk. The current study focused only on mental health patients, rather than the general community. But even within this focused group, accuracy in predicting suicide was a problem.

“A statistically strong and reliable method to usefully distinguish patients with a high risk of suicide remains elusive,” the paper concluded.

Source: Large M, Kaneson M, Myles N, Myles H, Gunaratne P, Ryan C. Meta-analysis of Longitudinal Cohort Studies of Suicide Risk Assessment Among Psychiatric Patients: Heterogeneity in Results and Lack of Improvement Over Time. PLOS One. 2016.