How Predictive Analysis is Protecting Our Veterans

By Melissa Tilley

It is easy to overlook the importance of proper data management, particularly when a system has been set up that is running smoothly and preventing the types of problems that can occur when we fail to track important information. However, occasionally a story comes along that reminds us of exactly why content and data management is so important, not just for companies and government departments, but also for the people whose lives they can influence. The data these organizations gather can be used not just to increase efficiency or uncover hidden patterns, but also to help and protect individuals.

Data Isn’t Just Data

The data that crosses our computer screens can often seem to simply be made up of strings of numbers and letters, completely detached from the real world and the people in it. However, once we consider what this data might represent, we can see that it might be used not just to learn more about people in general, but to help particular individuals who are in need. One way in which this is being done is by using the data produced by social media to monitor soldiers and veterans who are at risk of Post-traumatic Stress Disorder (PTSD) and suicide.

Post-traumatic Stress Disorder

The shocking truth is that more active duty service personnel lost their lives due to suicide in the last year than were killed in action. A survey conducted by the Iraq and Afghanistan Veterans of America found that 30 percent of the group’s members had themselves considered suicide, while 37 percent knew another veteran who had taken their own life. The National Center for PTSD estimates that Post-traumatic Stress Disorder may have affected between 12 and 25 percent of the military personnel who have served in Iraq and Afghanistan since 2008. For both the veterans who are affected by this condition, and the military and veterans groups that aim to support them, learning how to deal with PTSD at both an individual and organizational levels is a priority. While psychologists and medical professionals are finding new ways to treat and support people who are living with PTSD, the organizations that work with soldiers and veterans need to find ways of reducing the risk of developing the condition, and of identifying and helping those who are affected. Data management systems are one of the tools that can be used to tackle these problems.

The Durkheim Project

The Durkheim Project, supported by DARPA’s Detection and Computational Analysis of Psychological Signals project, has been developing a system for making use of online data to detect veterans who are at high risk of depression, suicide and PTSD. The nonprofit group has been using data gathered, with the permission of the participating individuals, from social networking sites in order to determine which veterans might be at risk. The aim is to find a way of predicting when someone is in need of help, in order to save lives and ensure that people obtain the support they need.

The new tool makes use of predictive methods similar to those employed in analyzing the stock market, applying them to data gathered from the social media profiles of veterans. Although people often find it difficult to reach out to those around them directly when they are feeling suicidal, it is often possible to recognize signs that they are struggling in the posts they make online. Analyzing these messages allowed the Durkheim Project to identify textual signals of suicidal thinking that could be analyzed in real time, as they were being posted. The predictive model, which continuously monitored the collected data using a system based on machine learning, was not perfect, but it achieved an accuracy of 65 percent during the initial stages of development.

The Future of PTSD Treatment

As work on the project continues, it is hoped that it will allow veterans who opt in to the service on their social media accounts to obtain the help they need even if they are unable to ask for it themselves. One version of the tool, designed for use in a clinical setting, will enable doctors to monitor their patients remotely, offering predictions of suicide risk based on continuously analyzed streams of social media data and information taken from medical records. Another version will enable other active duty military personnel and veterans to opt in to the service on their social media accounts in order to enable their activity to be monitored. The system will ensure that the danger signs that might suggest someone is feeling suicidal are no longer missed.