Researchers Develop App That Predicts Suicidal Tendencies With 92 Percent Accuracy


Researchers from the Indiana University School of Medicine and the Indianapolis VA Medical Center have published promising findings from a study evaluating the predictive power of online mood surveys when it comes to identifying patients struggling with suicidal thoughts. Researchers set out to establish “improved, objective, and quantitative” ways of predicting suicidal behavior, ultimately incorporating both an online mood survey and blood test results in an algorithm-based approach that demonstrates an impressive 90-percent accuracy of predicting suicidal ideation.

The study, which was published in this month’s issue of Molecular Psychiatry, followed 217 male patients recruited from the Indianapolis VA Medical Center over the course of several years. Researchers drew blood and had participants complete mood surveys every three to six months for the duration of the study. Within the 217 patients, the team was hoping to identify a subset of patients that would ultimately go through a period of suicidal thoughts during the study, as measured on the mood questionnaires. Researchers wanted to use data collected from these participants to compare the blood work results from when they were non-suicidal with the same results from when they were having strong suicidal thoughts. Of the 217 initial participants, 37 reported experiencing both no suicidal ideation and high suicidal ideation at varying points in the study. Over the course of the study, 106 blood samples were taken from the group. The results from these samples were compared with blood samples taken from 26 deceased males of the same demographic makeup that had recently committed suicide. Samples taken from the initial 37 participants, along with the 26 post-mortem samples collected, became the focus of the data analysis for the remainder of this study. Researchers combing through the blood work results in this dataset discovered 11 biomarkers that independently predict suicidal ideation with a 70-percent accuracy.

To improve upon those results, researchers developed a mood questionnaire and turned it into a mobile app. The questionnaire solicits information about issues that have a known correlation to suicide, including struggles with addiction, physical health, stress levels, and having a history of physical abuse. On its own, the survey was able to predict suicidal ideation with an 80 percent accuracy rate, but taken together with the blood test results, researchers were able to accurately predict which participants would experience suicidal ideation over the next year with a 92 percent accuracy rate. Lead researcher Alexander Niculescu explains, “The app assesses the context in which the biomarkers are elevated. If you have the biomarkers in the context of having these other risk factors, then you’re at very high risk—that’s what our study is showing.”

Enjoy HIStalk Connect? Sign up for update alerts, or follow us at @HIStalkConnect.

↑ Back to top

Founding Sponsors

Platinum Sponsors