Eliminating Veterans' Homelessness: Connecting a Deserving Population with the Right Resources

September 2012
Topics: Veterans Affairs, Modeling and Simulation, Probability and Statistics, Veterans Health and Benefits Modernization
When the Obama Administration launched its ambitious Eliminating Veteran Homelessness initiative, the Department of Veterans Affairs enlisted MITRE's statistical modeling expertise to help VA best allocate funds to the program.
MITRE suppoprts eliminating veterans homelessness program

Imagine a running faucet, filling a tub faster than you can drain it. What happens when the water overflows the rim?

That's akin to the situation faced by society when it comes to homelessness——particularly among those who have served our country. But thanks to some help from MITRE's statistical modeling expertise, the Department of Veterans Affairs has maintained funding to combat this challenge —and is making measurable headway on draining away the problem.

The longstanding problem of homelessness among U.S. veterans arises from complex factors, including the economy and the physical and mental health of the veterans themselves. Today, veterans are disproportionately represented in America's homeless population, making up nearly 10 percent of the adult population, but as much as 16 percent of homeless adults.

In late 2009, the Obama Administration launched Eliminating Veteran Homelessness (EVH), a program to end homelessness among veterans by FY 2015. The EVH program gives veterans access to services such as safe housing, healthcare treatment and support services, opportunities to return to employment, and benefits assistance. At the time the program launched, however, there was a major problem: the troubled economy, coupled with EVH's need for increased funding to meet its goals.

With multiple government programs competing for limited resources, some scenarios showed EVH's budget being cut by as much as 35 percent. The VA needed to figure out how to best allocate its funds—and that's where MITRE came in.

"Our involvement started with the VA's Deputy Secretary, who looked at the wide array of VA homeless programs and asked, 'How much do we need to invest? Are we investing enough, and in the right programs, to meet the President's goal of zero homeless veterans by 2015?'" says Bill Burns, program manager in the Center for Enterprise Modernization, the MITRE federally funded research and development center sponsored by the IRS and co-sponsored by VA. "That's a hard question to answer, —but statistical modeling can help."

Creating a Model

In 2011, a MITRE team began developing a statistical model that would shed some light on the problem. "We needed an idea of the impact that different investments in VA homelessness programs would have," says Walter Ellis, a lead information systems engineer who supports VA programs. "What executives can really influence is the budget, and the idea of the model was to help them see what those choices would yield.

"We wanted to answer the question, 'If the budget increases or decreases in different areas, what will the homelessness level look like in 2015?' In our initial meetings, it became clear that we should use system dynamics for this particular problem."

System dynamics is a methodology and mathematical modeling technique for framing complex issues over time, and uses the concept of "stocks," or variables measured at a specific time, and "flows," or variables measured over an interval. (See "Stocks, Flows, and Bathtub Models," below.) Once the team agreed upon a method, Ellis and his team went looking for MITRE colleagues who could make it happen. "We only had a week to prepare for a meeting with the Deputy Secretary," he says.

By using the "expertise finder" feature on MITRE's intranet, he came across the work of Chris Glazner, a lead information systems engineer and modeling and simulation expert. "I called him, and our project tied in perfectly with his research interests," Ellis says. Glazner arranged a teleconference with the VA National Center on Homelessness Among Veterans (known as the National Center), traveled to its offices in Philadelphia the next day, and had a prototype of the model ready for the briefing.

Ellis credits working across the corporation as one key to the success of the project. "Chris was able to get us through the first critical hurdles, and he was soon joined by Steph Glasser, a lead information systems engineer with extensive system dynamics modeling expertise," he says. "We have a great team that's picked up elements from all over MITRE."

Working closely with the National Center, the team built a program performance and budget simulation that gave VA insight into the way their programs interacted. "The model helped VA understand the complex interdependencies in their programs," Burns says.

Visibility into Impact Yields Results

After putting the MITRE-developed model through external review and validation, VA used it to compare results from a series of budget scenarios——developed by the Homeless Program Office, other VA organizations, and external partners such as the Department of Housing and Urban Development (HUD)——against the baseline budget. The new model enabled program leadership to see what impact changes to the program's budget would have on the goal of ending veteran homelessness.

"This capability, used in conjunction with input from program stakeholders, allows VA to run performance and budget scenarios far more quickly and accurately than ever before," says Jim Chin, a group leader in MITRE's Systems Design & Integration department. "This brings them closer to their goal of eliminating homelessness among veterans in the next three years."

As the program grows to reach increasing numbers of veterans, MITRE will provide strategic support to the continued progress of this crucial initiative. According to Ellis, his team is currently working on a regional version of the model. Burns points out that the greatest benefit of the model is that it quantifies the intuition of the EVH program's subject matter experts and validates long-held assumptions.

"VA's concern was that without a solid justification for the EVH budget, those funds would be cut," he says. "After running the different budget scenarios through the model, VA sustained the funding for their homeless initiatives. MITRE's ability to help them justify their budget has resulted directly in VA serving more veterans."

——by Tricia C. Bailey


Publication Search