Woman in glasses analyzes data in a dark room

MITRE Adds a Special Element of Trust to Data Sharing and Analysis

MITRE is finding limitless applications for a data sharing and analysis model that has dramatically improved aviation safety. That model is now being used to keep children and patients safer and help the government better detect fraud.

Eleven years ago, MITRE engineers helped launch a new kind of public-private collaboration: the Aviation Safety Information Analysis and Sharing (ASIAS) initiative. At the time, they had no idea the model would not only produce dramatic improvements in U.S. aviation safety but would also spur the creation of similar programs across the globe and serve as a model for a wide variety of other safety initiatives.

But it did. Today, MITRE is applying the experience gained on ASIAS to initiatives in healthcare, child welfare, and tax fraud, among others. These are just a few of the ways MITRE has applied the model—so far.

The work demonstrates MITRE's unique vantage point—enabling us to apply what we learn from one sponsor domain to significant challenges in other arenas. What started with the skies now covers more ground than ever before.

Each new endeavor relies on diverse groups being willing to share their data with us, based on our code of trustworthiness and objectivity.

Improving Aviation Safety Through a Pioneering Partnership

At its core, ASIAS is a public-private partnership of data sharing for the common good. The partnership includes government agencies, aviation stakeholder organizations, aircraft manufacturers, and dozens of airlines and corporate operators. The program obtains and fuses data from these partners and other sources so that safety trends can be identified and addressed before accidents or other serious incidents occur.

The Federal Aviation Administration selected MITRE to facilitate the data sharing and analysis aspects of ASIAS. Despite the FAA's decades-long collaboration with us, that choice wasn't as simple as it sounds.

"We had to demonstrate to our ASIAS partners that we could protect their data, keep it anonymous, and conduct analyses that would reveal existing or potential safety issues that they could then use to introduce mitigations," says MITRE's Ed Walsh, department head for aviation safety analysis.

MITRE did all that, with profound results. Today, the program is credited with dramatically improving commercial aviation safety in the United States and is considered a world standard for aviation safety data sharing and analysis efforts.

ASIAS's applications now extend beyond aviation. In the healthcare domain alone, we've already found several areas where a data sharing and analysis approach shows promise.

Taking a Bite out of Healthcare Fraud

In 2012, the Department of Health and Human Services (HHS)—with MITRE's help—launched the Healthcare Fraud Prevention Partnership (HFPP). Built on the ASIAS model, HFPP is a public-private partnership. Federal government, state agencies, law enforcement, private health insurance plans, and healthcare anti-fraud associations share their data and information with a trusted third party for analysis. MITRE served as that third party, analyzing HFPP partners' data to help them detect and prevent healthcare fraud.

Since HFPP's launch, our analyses have helped the HFPP partners save hundreds of millions of dollars—money that might otherwise have been paid as the result of healthcare fraud.

Reducing Preventable Medical Errors

In 2015, our researchers decided to see if the ASIAS approach could help prevent avoidable—and often fatal—medical errors that occur in health facilities. They created the National Patient Safety Partnership (NPSP) to collect data from three pediatric hospitals. The team then analyzed that data for factors that correlated with patient safety issues.

During their initial study, they focused on patient deterioration and medication safety events.

"We were able to predict with a high degree of specificity which patients' health would deteriorate, as well as factors associated with medication errors," says Sybil Klaus, who leads MITRE's healthcare research program.

MITRE is now using similar data for other research. "Our next phase is to reimagine how to measure quality of care using that data," Klaus says.

Preventing Child Welfare Fatalities

Also in 2015, MITRE began another project for HHS—this time to assess the potential benefits of a data sharing and analysis approach in the field of child welfare.

When portfolio manager Mark Thomas learned about the work of the Congressionally established Commission to Eliminate Child Abuse and Neglect Fatalities, he suggested that an ASIAS-like data model might support the commission's goal. That outreach led to a recommendation from the commission =to Congress that HHS adopt a data sharing and analysis approach to combatting child welfare system fatalities.

Additionally, MITRE developed guidance documents that HHS could share with state and local child welfare agencies to help them apply analytics to improve the safety of vulnerable children.

Testing the Approach at the Local Level

MITRE also conducted research in collaboration with the County of San Diego's Health and Human Services Agency to test a data sharing and analysis approach at the local level.

"Child welfare systems are often only as effective as they information they know about and can act upon," explains model-based analytics engineer Chris Teixeira, who led the study. MITRE researchers collected and analyzed data from a variety of sources to provide a fuller picture of the adults who interacted with children in the system.

"We were able to determine not only the risk factors that exist, but also which ones are impactful and distinguishable from random effects," Teixeira says. Those findings are now informing efforts at the national level.

Putting the Brakes on Automotive Hazards

More recently, we've begun translating our success with ASIAS to the automotive industry, which is critical because accidents claim tens of thousands of lives each year. 
In 2018, the National Highway Transportation Safety Administration and auto manufacturers agreed to participate in a prototype initiative, called the Partnership for Analytics Research in Transportation Safety (PARTS). Like the aviation model, the PARTS vision is for the automotive industry to voluntarily share safety data—with the latest security and privacy protections in place—for the sole purpose of proactively advancing safety.  
Through that prototype, our researchers collect and analyze data about crashes and vehicle characteristics to create a fuller picture of how advanced safety technologies perform in real-world situations.
"In PARTS, we're focusing on collaboratively gaining real-world safety insights to understand how new and emerging automotive technologies are actually working," says PARTS project leader Jessica Lascara. "Initially, we focused on how automatic emergency braking is reducing crash rates. We now have plans to expand that research into other areas in 2019 and 2020.” 

Spurring Millions of Dollars' Worth of Tax Fraud Prevention

Preventing identity fraud in the tax system has proved fertile ground for applying a data sharing and analysis approach as well.

"When electronic tax return filing took off, so did identity fraud," says Andy Taylor, portfolio manager for IRS information technology. "And that's not just an IRS problem."

It also concerns state revenue agencies and the companies that submit electronic tax returns on behalf of their clients. "Because the problem involved all of these organizations, we recommended IRS work with them to form an information sharing and analysis center [ISAC]" Taylor says.

IRS agreed, and MITRE is now operating the center. "We serve as the trusted third party. We collect data from the different partners, analyze it, and provide anonymized results back to the ISAC community."

"The ISAC has created an analysts' community of practice," adds business strategist Andy Dziewulski, who is now leading this work. "The fraud analysts from each of the sectors—the tax industry, the state departments of revenue, and the IRS—are able to share best practices, processes, and findings daily. Through this sharing process they've been able to prevent millions of dollars in false refunds."

Exploring Other Applications

These are just a few ways that we are using the experiences gained from ASIAS and other PPPs to support other MITRE sponsors in a variety of domains. Members of the PPP Community of Practice have also shared insights with MITRE and industry colleagues seeking to establish data sharing partnerships to address veterans' health, cyber incidents, and railroad safety.

And that's just the beginning. More applications for the ASIAS model are bound to follow. The potential for such public-private data sharing and analysis initiatives is vast.

—by Marlis McCollum