CEM FFRDC Innovation Exchange hosted by MITRE and the Center for Enterprise Modernization FFRDC

Veteran Experience and Engagement
Useful Citizen Feedback for Modeling and Decision-Making Sherri Condon
Alison Dingwall
To improve customer service, government agencies need cost-effective ways to obtain feedback about citizens’ attitudes and concerns. Meanwhile, citizens publicly and openly share their thoughts and experiences online in millions of communications daily. Our research is demonstrating that citizens’ communications online and in social media can be analyzed to deliver useful feedback that agencies can use to improve service to the public. 
Veteran-Centric Mobile-Based Service System Peter Whitehead By establishing the veteran at the center of a network of wireless devices (smartphones and tablets), we establish a platform for improved services and innovation. We are exploring a design-thinking approach to a presentation for the VA describing and demonstrating this capability including: the privacy issues and how to promote opting-in, system risk reduction, and the potential for innovation and technology transfer. We will show with select use cases and mock-ups how this network could improve services. We also are exploring the trade-space for major potential obstacles to realizing the vision. We focus initially on VHA use cases, but the envisioned network would have the potential to improve all VA services. 
Empowering Patients with the Patient Toolkit Kristina Sheridan
Jason Ji 
Roughly four out of five U.S. healthcare dollars are spent on behalf of people with chronic conditions. Managing and treating chronic conditions present many challenges.  Providers and patients must often make critical decisions based on limited, incomplete patient-provided data—data pulled from memory and hand-written notes.  The intertwined processes involved in patient communication, collaboration, and coordination undermines the daily processes necessary to control chronic conditions.  MITRE researchers searched for ways to empower patients and caregivers on a daily basis and allow them to capture a complete set of patient-generated data necessary to support diagnosis and treatment evaluations.  Looking for something easy-to-use and habit forming—something a person would use all day long—the team explored mobile technology and iPad applications. This work led to the "Patient Toolkit," a patient-facing application designed to support chronic patients and their caregivers.  The team is partnering with the University of Virginia to quantify the reliability of patient-generated symptom data, measure the value of these data to providers, and collect patient feedback on the usability of the toolkit.  In collaboration with Carnegie Mellon University, the team surveyed and interviewed physicians to learn what patient-generated health data would be most useful from a clinician’s perspective and how to visualize and share this data.  This research will increased patient empowerment and enable patients and caregivers to fully partner with their care teams, resulting in a cultural shift in how a patient is viewed.  It will shift the view of a patient from being a passive recipient of care, to being an active participant of their care team.
Social Radar Alison Dingwall
Mike Fulk
There is a growing need for real-time access to Veteran information in order to improve Veteran health, awareness of VA services and benefits, and participation in government programs. The Department of Veterans Affairs can gain earlier insight into Veteran needs by looking at timely and relevant on-line information. This can aid in communicating tailored messages to specific groups that can be identified via demographics, topics of interest, or regional location. Online and social media can be used to disseminate health and safety information and tailor strategic communication messages to communities in order to improve public engagement and empower individual decision making and behavior change.
Cyber Security
Biometrics-Based Authentication for Cyber-Physical System Security Qian Hu
Josh Kraunelis
Stelios Melachrinoudis
This team has developed a secure method for voice authentication on mobile devices, such as smartphones, laptops, and other special-purpose embedded computing devices. MITRE is transitioning the prototype technology to sponsors and industry. 
AJACS Lindsley Boiney
Clem Skorupka
This team has developed a rapid, light-weight methodology to assess a Security Operations Center’s maturity along five dimensions that are essential for both effective cyber defense and effective threat sharing. 
Non-ID ID Rick Knowles
Katherine Schoenfelde
Exploring the landscape of options for on-line identity proofing. As government agencies interact more and more with citizens over the Internet, they need rigorous methods to authenticate that an individual is who he says he is and lives where he says he does. This problem is especially acute when a pre-assigned ID number is not appropriate, permissible, or cost-effective. We are exploring and evaluating possible approaches and their applicability to a variety of government services (starting with the Census Bureau’s decennial data collection) and will recommend and possibly prototype one or more solutions.
Leaner, More Effective Operations
A Quantitative Approach to Forecasting Demand Sam Steckley
Angus Chen
This effort has investigated quantitative methods to improve forecasts for the demand for government services.  The methods include developing models incorporating demographic, economic, socio-political, and other environmental variables that may drive demand.  We have also investigated the performance of ensembles of models in which multiple quantitative forecasting models are combined to make projections.  The methods have been tested on a case study involving the demand for Veteran benefits. 
Modeling agency-individual performance alignment within the government service ecosystem Amy Squires
Suzanne Geigle
This project team is exploring the feasibility of creating a model for optimally aligning individual performance objectives to agency outcomes. The goal is to develop a “what if” modeling capability that would not only make transparent any alignment/misalignment of objectives within a service but would also enable agencies to test alternate individual performance objectives to assess the implications for overall service results. VBA is a case study for this work. 
Policy Lab: Exploring Effects of Policies on Complex Sponsor Environment David Allen
Will Lee
All government policy changes have complex consequences. For example, for any given policy change, there will be benefits, costs, and risks. Some will be expected, some will be surprises. This team is investigating and maturing the use of serious games as a tool to improve the development of policies by providing a capability to test policies to verify they meet their intent before they are implemented. In our Policy Lab, we subject proposed policy changes to independent attack and review, using scenarios that elicit information about real stakeholder behaviors from participants.
Computational Modeling of Sponsor Rule-sets to Help Measure Policy Effectiveness David Slater Last year we began exploring how network analysis can provide useful insights into dependencies among different levels of policies (laws, rules, forms). Based on sponsor interest, this year we are exploring ways to gather additional insights from these networks, including the use of human language technology to better understand the meaning and context of the references, how the computational aspects of the network analysis can be used, and ways to make the tool more robust and generic so that it can be applied to multiple sponsor domains. The initial use case was inspired by VBA challenges.
Dynamic, Citizen Centric Government Services Ecosystem Modeling Zoe Henscheid
James Thompson
Alison Dingwall 
As government communicates and provides services to citizens, how can it adapt to rapidly changing citizen preferences and behaviors? We will expand on previous work in this area to develop behavior models that employ citizen and ecosystem stakeholder behaviors. Our goal is to build a computational tool that will allow agencies to evaluate the effectiveness of different outreach campaigns on increasing citizen participation and/or compliance. 
Health Transformation
Transcript: Medication Reconciliation David Tresner-Kirsch This team is working to improve the safety of prescription instructions through the application of automated language processing and artificial intelligence. Typically, medication instructions are conveyed in natural language, which is difficult for automated systems to reason with. We have prototyped and leveraged a novel system that can extract a structured knowledge representation from those natural language instructions and, subsequently, reason with or transform that representation.
Rapid Screening of Mild Traumatic Brain Injury Michael Fine
Elizabeth Brokaw
Rapid Screening of Mild Traumatic Brain Injury, Michael Fine. This team is developing the technology necessary for rapid screening of mild traumatic brain injury (mTBI), which will allow clinicians to diagnose and treat mTBI at its earliest and least problematic stages—leading to better outcomes and reducing costly long-term care. We plan to develop, test, and integrate the screening technology into a small, portable device for field testing.
BrainKit: TaskPlanner Lindsay Kaye BrainKit: TaskPlanner team designed and built a prototype iPad application to help patients with prospective memory issues accomplish the activities of daily living. The team has tested the application with patients in the field and gathered feedback, which it will use to refine the tool. For example, we are exploring the most effective ways of displaying data collected by the app so that the dashboard can provide insight about the patient’s skills both to the patient and the clinician. (The prototype app is available in the Apple App Store.)
Patient and Medical Product Safety Eric Bloedorn
Greg Nelson
Joy Tobin
Elizabeth Canzone
MITRE has spearheaded a collaborative and voluntary public-private national partnership to develop and share predictive analytics to systematically reduce medical errors and other harm to patients. This National Patient Safety Partnership will improve patient safety through 1) cross-partner sharing of patient data, 2) collection and multi-INT analysis of data across normally stove-piped modes (EHR + physiologic + incident reports) 3) identification of cross-partner commonalities and differences, 4) predictive and discovery algorithms applied to fused partner data.
Healthcare Fraud: Untangling Webs of Fraud Networks Mike Fulk
Megan Ward
Leveraging MITRE’s expertise in advanced analytics with healthcare analysis, we are collecting data sources that can be effectively mined for high-value fraud leads. By applying analytics techniques to multiple data sources, we will find associations among multiple, linked organizations to identify unusual patterns beyond typical analysis done on a per-provider basis.
Healthcare Fraud Prevention Partnership  Alanna Lavelle
Ted Sienknecht
A public-private partnership—the Healthcare Fraud Prevention Partnership—can identify trends and schemes previously unknown to payers through healthcare claims data fusion. This unprecedented partnership has the goal of reducing fraud nationally and includes CMS and private payers as partners with law enforcement and national associations.
Better Business and IT Functionality
Exploring Alternate Cloud Development Strategie Rick Cagle We are examining how to deploy complex, distributed software systems.  Specifically, we are modeling a scaled DevOps environment, including elements of Continuous Integration and Continuous Deployment in a Cloud Environment using an OSEHRA VistA instance to illustrate scenarios for applying patching, updates, and customizations across diverse production endpoints such as clinical environments.  Results are being used to shape acquisition guidance for developing and sustaining applications and reduce expense and complexity for sponsors.

Approved for Public Release; Distribution Unlimited. 15-1137
©2015 The MITRE Corporation. ALL RIGHTS RESERVED.