The Army says it’s found a way to use data it already has to learn more about its soldiers and the stressors that might affect their readiness and resilience.
In June, the service finished a roughly five-year project as part of the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS), which looked at the characteristics of suicide and other mental and behavioral health issues.
Army STARRS collected and organized 40 different datasets, mostly administrative information the Army already had about its soldiers. Data scientists studied demographic information, medical and law enforcement history, in addition to military characteristics like a soldier’s rank, deployment location, number of tours and when a soldier moved up a rank.
To protect soldiers’ identities, the Army STARRS team detached individual names from the actual data.
It’s the first time the service has gathered this much data on its soldiers, said Dr. Kenneth Cox, scientific liaison with the STARRS program.
“These tools are not being used at all with real people,” Cox said. “They’re not putting out lists that go to commanders. They’re not taking actions. It’s something that’s under consideration and a possible tool to add to the things that we have available to take care of our force and our people.”
The Army, in partnership with the National Institutes of Health, Harvard Medical School, the University of California, San Diego and other universities, began its study in 2009, when the military suicide rate surpassed that of the civilian population.
Out of about 500,000 people in the Army, the service identified a smaller group of 3,000 people who may be at higher risk for attempting suicide.
Taking a more holistic approach
The Army Resilience Directorate (ARD) is taking a broader look at physical, psychological, social and spiritual factors and how they may impact a soldier’s readiness.
In addition to more typical demographic information, it collected alcohol and substance abuse data, information on who was enrolled in substance abuse training, and attempted and completed suicide numbers.
Based on the data, ARD identified 540 factors that might influence a soldier’s resilience and readiness, said Randy Lane, division chief of the analytics, assessment and systems division at ARD.
“Now we’re trying to take a more holistic [approach], which is going to really get a more complex view of how the different factors start to interact,” Lane said. “The alcohol and substance abuse, the risk that our soldiers and their families undergo, the stresses, everything that goes into that picture, [and] trying to get a more complex interaction.”
Both Army STARRS and the ARD had the challenge of bringing all of this information from separate databases together. The ARD is working with an Army analytics group to meld some of those disparate systems together, Lane said.
The Army is also using the STARRS and Resilience Directorate data to create or update training guides on behavioral and mental health for commanders and leaders, said Wendy Lasko, manager for the Army Suicide Prevention program.
Based on the trends, Lane said the ultimate goal is to develop an “intervention methodology,” or a model or plan for speaking with — and helping — soldiers who might be at risk.
“We know [what] will affect a certain population, but that’s not good enough for us,” he said. “What we want to get to eventually is that risk-based methodology, really that interaction of the factors … in a more complex, behavioral modeling effort that we’re undergoing with the Army analytics group. That’s where we feel we’ll start to neck down and get more specific on who we intervene with.”
Approaching from a positive perspective
Though funding for the Army-specific STARRS program ended this past summer, the Defense Department is supporting a similar five-year study, the STARRS Longitudinal Study of Army soldiers.
The Army is also following 70,000 soldiers who participated in past surveys and interviews who indicated they were willing to be followed again, Cox said.
Army leaders are evaluating the trends the STARRS program and ARD uncovered. They’re now tasked with deciding how the service should use this information, what policy changes it should drive and how commanders can — or should —intervene.
“It is very easy to take these kinds of predictive analytics when they’re being used on people and not on predicting whether sales of male hygiene products, [for example], will be up in the third quarter of this year and we should not stock up more,” Cox said. “It’s a lot different when you’re deciding, well maybe this person needs to be given information, maybe we won’t keep them in the Army, you have to decide what you’re going to do with that and you can cause inadvertent harm.”
Though the question of intervention is a tricky line to walk, Lane said Army leaders are looking at where the data points them now and where it may point them in the future.
“[We’re] really walking ourselves forward and creating a running assessment of where we are as a force in readiness in resilience,” he said. “That, we think, will then inform the policy, the training and how we approach the problem more positively.”