The MITRE Corporation recently announced the winner of its 2018 Healthcare Anti-Fraud Academic Competition. We are proud to announce that the SBMiners from the McWilliams School of Biomedical Informatics at UTHealth Houston, formerly UTHealth Houston School of Biomedical Informatics (SBMI) is the winning team.
The SBMiners team includes Ph.D. student Kang Lin Hsieh and M.S. students Adeola Akinfaderin, Avi Raju, who served as team lead, and Bin Yao. Associate Dean for Academic Affairs and Professor Susan Fenton, Ph.D. is the team’s academic advisor.
“Each student played an important role during the competition,” said Fenton. “Avi was a good team leader while Kang handled the complex analysis and Bin provided additional analysis support. Adeola was instrumental in creating the presentation and gathering our references.”
To launch the competition, MITRE partnered with the National Health Care Anti-Fraud Association (NHCAA). Students from U.S.-based colleges and universities worked collaboratively to create innovative approaches for finding fraud while using a dataset and tools provided by MITRE. The competition encouraged teams to develop innovative solutions that could help government and private healthcare payers reduce dollars lost due to healthcare fraud.
“MITRE decided to pursue this competition to raise awareness of fraud in health care, expose competitors to the behind-the-scenes efforts to combat these issues, and generate interest in fraud detection and prevention,” stated Lisa Cannady, MITRE’s project lead for the competition. “Advanced analytics skills are key to successfully combating healthcare fraud and new talent and ideas are needed.”
Each team had access to a large, de-identified set of recent medical and pharmacy claims, a dedicated cloud-based Amazon Web Services workstation, and a suite of open source tools to use to analyze the data. “We designed the competition to provide the students with resources, tools, and a de-identified set of real-world healthcare claims for the teams to analyze the data, uncover potential schemes, and understand the potential significance of their work,” Cannady said.
Teams also worked with law enforcement and private industry experts throughout the competition. The experts offered guidance on the importance of data in investigations and gave helpful information on their personal career paths.
During the competition, which began in Feb. and ended in mid-April, teams were required to perform data analysis to detect anomalies, prioritize findings and develop an investigation recommendation plan for submission. Judges from government agencies and private insurance companies evaluated the submissions. The judges assessed the sophistication and innovation of the analytic design, the overall technical competence of the team and the ability to link the data analysis results to an investigative recommendation.
As the competition winners, the SBMiners team demonstrated advanced technical skills to identify potentially fraudulent schemes using cluster analysis. The team’s analysis approach included a geographical-based focus and analysis by ratios. Additionally, the team chose impactful visualizations to convey their recommendations.
“Winning the MITRE competition demonstrates not only how diligently the team members worked, but also speaks to the caliber of the students we educate at McWilliams School of Biomedical Informatics,” stated Fenton. “The students take courses focused in data visualization and biomedical data analysis and successfully applied the knowledge gained to win a national anti-fraud competition.”
The two runner-up teams are Texas State University and University of Chicago, Illinois. The SBMiners will present their solution in August at the NHCAA Skills and Schemes for Healthcare Investigators conference in Denver.
published on 6/6/2018 at 5:00 p.m.