Better together

In the growing field of data science, UMKC researchers take an inclusive approach to solving complex problems

What’s the point?

This sentence is often said in exasperation. But also, this statement sums up the levels of complexity related to data.

In projects big and small, data science and data analytics — often used interchangeably,  despite distinct differences — drive decision making in business, government and academia. Using the right data correctly and for the right purposes requires expertise and sensitivity. It’s a high-stakes challenge.

Brent Never, Ph.D., says UMKC is more than ready to meet that challenge.

“We’re a comprehensive university, and that makes us different from other universities, particularly in this region,” Never says.

Part of what makes UMKC different is its recently created Institute for Data Education, Analytics and Science (IDEAS), of which Never is coordinator. The vision of the institute is to position UMKC as the top option for data science education, research and service in the region, building on the university’s strengths in biomedical informatics, big data analytics, image analysis, natural language processing and geospatial analysis.

The institute is just one way UMKC is harnessing the depth and breadth of its resources to use data for good.

Putting patients first 

Mark Patterson, Ph.D., M.P.H., engages both small data and big data in his pharmacy research, which focuses on improving the accuracy of the prescription process and, therefore, health outcomes.

Since 2016, Patterson has worked to build a registry of prescription discrepancies that occur during transitions of care between hospitals and nursing homes in rural areas. With support from the American Association of Colleges of Pharmacy, he works with another pharmacist and a nurse to analyze mismatches in discharge orders, looking for breakdowns in communications that adversely affect high-risk patients.

Patterson’s registry links patients, prescription orders and communication failure locations. The goal is to pinpoint where the discrepancy occurred and follow up with quality improvement standards to fix system flaws.

Another of Patterson’s projects uses big data to investigate prescribing patterns of long-acting antipsychotic injectables administered during hospital stays. Together with fellow School of Pharmacy professors Yifei Liu, Ph.D., and Steve Stoner, Pharm.D., Patterson uses millions of points of data from thousands of patients to measure and compare anti-psychotic utilization rates across multiple drug brands. The data comes from Cerner Health Facts, a massive de-identified, real-world health database.

Much of Patterson’s work has been creating the parameters to clean and design the data sets, making big data more manageable to analyze. Patterson says it’s a majority of the work.

“It takes a lot of patience and attention to detail to quality check and build an analytic dataset,” Patterson says. “Getting there takes months.”

The desired result of improving health across the region, though, is worth it.

Fighting blight, one data point at a time 

Never and Bloch School colleague Jim DeLisle, Ph.D., have teamed up to use data analytics for urban-core neighborhoods. In the case of abandoned housing, for instance, they work to identify which houses could tilt or flip a neighborhood.

“How do you pick the 1,000 houses that if you could clean up, are going to make that neighborhood better for residents?” he says.

DeLisle and Never created “Abandoned to Vibrant,” a database of abandoned homes in the Kansas City area. The site pulls reported data from public officials and citizens: code violations, 311 calls and the like. Using spatial analysis and the data analytic software Alteryx, the researchers created what Never calls a “Zillow for distressed homes” that transforms data points into a visual platform that can be used by officials and citizens.

In particular, the project provided residents key information on homes they could buy from the Land Bank of Kansas City to flip or occupy themselves. This information is especially important because of a city law stating that residents have 120 days to bring homes bought by the Land Bank up to code.

In many cases, Never says, families buy the wrong house and are not equipped to deal with structural deficiencies. They put in their money and time, only to have the city repossess the house in the end. This new platform can help potential home buyers avoid that situation long before they buy. For their work, DeLisle and Never were awarded the 2018 Alteryx Excellence Award.

Never says his research on abandoned homes will continue through a proposed grant with the National Science Foundation, which will allow him and his colleagues to examine blight on residential streets using cameras on trash dumpster trucks. The goal is to head off potential homes before they are abandoned and create blight.

“It’s easier to intervene now and pay $1,000 to mow the lawn, than to pay $10,000 to deal with abandonment,” Never says.

Where data meets privacy 

Tony Luppino, J.D., believes that, when it comes to the intersection of data and the law, it’s crucial for lawyers to be included in the conversation. Luppino is working with the MetroLab Network, a national consortium of cities and universities focusing on civic innovation, on a policy that will help cities pursue the benefits of data while respecting privacy and civil liberties.

UMKC students and faculty developed the policy with local government and other organizations. It addresses such questions as: Who decides what data to collect and how to store it? When and how is consent sought to collect, use and share a person’s data? What oversight systems are appropriate to manage data handling processes?

The objective is to balance the value of data with risks of intrusions on privacy.

“Everyone sees the value of data-driven decision-making, and cities have been wanting that,” Luppino says. “This should not be left to a few decision-makers. The public ought to weigh in on some of these things.”

The School of Law has organized collaborations at intersections of law, technology and public policy since 2014, when a team led by former Dean Ellen Suni, supported by the Ewing Marion Kauffman Foundation, worked to help law faculty and students across the country explore uses of technology. While not a data scientist, Luppino sees the need for attention to law in data science endeavors.

“Even though I focus on unintended consequences, I appreciate the value of data science in improving public services, health, safety and economic opportunities,” he says. “Data collection and analysis can productively inform policy and decision-making, and I’m all for that if done responsibly.”

Keeping connected

The one common element for successful data research is connectivity. That’s why IDEAS is so effective. The institute works to provide guidance and expertise in data to researchers across all fields, not just a chosen few. That spirit of collaboration pulls researchers like Never, Patterson and Luppino together to support data science.

The institute has three main functions: offering for-credit educational classes and certifications through various academic units, using UMKC researchers to collect and analyze data for industry and community good and, finally, fostering the multidisciplinary research capabilities across UMKC.

IDEAS, though, is just one facet of Chancellor Mauli Agrawal’s strategic goal to promote data research on campus. That goal is also being realized through the health informatics work being done through the UMKC Health Equity Institute, led by Jannette Berkley-Patton, Ph.D., associate professor in the UMKC School of Medicine. Another important group is the NexGen Data Science and Analytics Center, headed by associate professor Praveen Rao, from the School of Computing and Engineering, and housed on campus through $20 million in support by the University of Missouri System. And, of course, UMKC researchers continue to work with local municipalities and organizations like Children’s Mercy hospital.

The overall hope is to create synergy in both the sciences and humanities, using data-driven principles to advance a mission for the common good.

“It blows my mind what people are doing at UMKC,” Never says. “It’s been a great opportunity.”

And that is the point.

EMPOWERING INDIVIDUALS THROUGH HEALTH AND NEIGHBORHOOD DATA

KC Health CORE, a project led by the UMKC Center for Economic Information and Children’s Mercy, works with agencies throughout the Metro to research regional health disparities and intervention.

What is KC Health Core?

The KC Health Community-Organized Resource Exchange (CORE) is a multi-year initiative award by the Health Forward Foundation to facilitate and coordinate regional health disparities research and intervention via a neutral data exchange platform.

Who are the data partners in this initiative?

Children’s Mercy leads the project alongside the UMKC Center for Economic Information. Data contributors also include:

  • mySidewalk for its data library, apportionment tool and web interface
  • Mid-America Regional Council for primary geospatial and regional data contributor
  • The Kansas City Missouri Health Department for micro-level pediatric blood lead testing
  • The UMKC Center for Neighborhoods as a community engagement partner

How is the community involved?

The project utilizes a model for neighborhood engagement and advocacy in the research and policy-design process, developed by the UMKC Center for Neighborhoods. Another example of community involvement is in Wyandotte County, Kansas, where researchers are collaborating with WYCO Livable Neighborhoods and the H.E.A.T. project neighborhood leaders and leadership team.

What will neighborhoods and people able to do with the data?

The data will help empower neighborhood leaders through neighborhood planning activities, which are supported by a network of researchers and students continually involved in the process of solving problems at the local level. This could lead to grant funding and other resources to benefit people subject to health disparities.