Thought leaders from the University of Maryland, IBM and the federal government joined on Nov. 17, 2017, at the Robert H. Smith School of Business to provide perspectives on the application of advanced analytics to behavioral healthcare challenges at the Sixth Annual Smith School and IBM Business Analytics Workshop, “Leveraging the Power of Analytics for Behavioral Health.”
“Behavioral healthcare challenges in the U.S. have increased dramatically to the point of reaching what some have characterized as a state of crisis,” said conference organizer Michael Ball, senior associate dean and Dean's Chair in Management Science at the Smith School, at the opening of the conference. Ball and longtime associate Frank Stein, director of the IBM Analytics Solution Center, launched the conference in 2011 to highlight the ever-growing importance of business analytics in all sectors. Over the years, the conference has put the focus on analytics in risk, healthcare, internet, supply chain management, and cybersecurity.
Keynote speaker Ira Katz, senior consultant for Mental Health Program Evaluation at the Department of Veterans Affairs and professor emeritus of psychiatry at the University of Pennsylvania, discussed suicide risks, prevalence and mortality in the U.S. veteran population. Katz said that with nine million enrolled, the VA healthcare program is the largest in the country. Suicide was the 10th leading cause of death and in 2015 there were 44,000 suicides as compared to 16,000 murders.
“One death is too many. Every suicide is a tragedy,” said Katz. “We all know people who have died from suicide. The suicide rate for veterans is especially high with 20.2 suicides among veterans each day.”
Katz discussed how data analytics is helping the VA manage the suicide crisis. Launched earlier this year, REACH VET (Recovery Engagement and Coordination for Health - Veterans Enhanced Treatment) is a predictive analytics system for suicide prevention funded by the Department of Veterans Affairs Office of Mental Health and Suicide Prevention. Using a statistical model with 380 variables and interactions, REACH VET uses information from health records to identify those who are at a higher risk for suicide, hospitalization, illness, or other negative outcomes. Once a veteran is identified, his or her VA mental health or primary care provider reaches out to check on the veteran's well-being and review their treatment plan to determine if enhanced care is needed.
“Although REACH VET began as an enhancement to VA’s suicide prevention program, it has evolved into something broader,” said Katz. “REACH VET identifies vulnerable patients and provides clinicians opportunities to enhance care.”
Over one year, the model will identify 30,000 patients who are at risk for suicide based on clinical signs and symptoms. The system indicates they are at risk, but maybe they don’t personally feel like they’re at risk, explained Katz. The VA has the largest crisis line in America with more than 800 responders.
Katz said that much work still needs to be done on the suicide crisis and the VA system should be viewed as a national resource laboratory.
The presentations analyzed relevant behavioral healthcare data, identified problem areas and shortcomings, and also presented solutions where analytics and cognitive computing can be integrated with aggregated multi-sourced data to deliver improved healthcare with a focus on the opioid crisis.
- Amelia M. Arria, director of the Center on Young Adult Health and Development at the University of Maryland’s School of Public Health, associate professor in the Department of Behavioral and Community Health, and director of the Office of Planning and Evaluation, discussed using data analytics in novel ways to study and predict substance use. As principal investigator on the College Life Study, a longitudinal prospective study of health-risk behaviors among college students, Arria’s research focuses on risk and resiliency factors associated with the development of mental health and substance use among adolescents and young adults, as well as the consequences of untreated mental health conditions and substance use. Her most recent area of interest is understanding how these issues can interfere with academic achievement, and what can be done to promote student success. Predictive models for academic success typically don’t include behavioral health, which is an oversight, she perceived.
- Amy Windham, director of Behavioral Health Research and Policy at Truven Health Analytics, an IBM Company, discussed the role of big data and advanced analytics in transforming behavioral health. She said that now more than ever people are empowered to monitor their own health by using technology.
- Mustafa Karakus, a health economist and vice president of Behavioral Health Services and Policy Research at IBM Watson Health, discussed the opioid crisis and potential role of cognitive analytics.
- Margrét Bjarnadóttir, assistant professor of management science and statistics at the Smith School, said it’s time we raise the alarm bell on opioids. She said that in 2015, 12.5 million people misused prescription opioids and that early detection is key.
Bhanu Vinzamuri, a postdoctoral scientist at IBM Research AI, as part of IBM’s Science for Social Good Program, explored the causal relationships between an initial opioid prescription and subsequent addiction. Some of the factors positively correlated to addiction included the length of the initial prescription, the use of synthetic opioids, age, and region of the country. Both Bjarnadóttir’s and Vinzamuri’s research suggested similar changes in opioid prescription regimes.
Ritu Agarwal, Robert H. Smith Dean's Chair of Information Systems and senior associate dean at the Smith School, moderated the panel discussion on key behavioral health challenges and how to use behavioral health analytics. Agarwal, who is director of the Center for Health Information and Decision Systems (CHIDS) at Smith, helped organize this year’s workshop.
Tari Rangel, a senior decision science analyst with the U.S. Army Medical Command Program Analysis & Evaluation department specializing in behavioral health analytics, discussed the Army’s Behavioral Health Data Portal and how they use data in the clinical setting.
Yang Yu, a senior health planner with Behavioral Health and Crisis Services within Montgomery County Dept. of Health and Human Services, discussed the case management system for predictive modeling that they use in Montgomery County.
The National Alliance on Mental Illness (NAMI) estimates that approximately one in every five adults in the U.S. – 48.5 million – experiences some form of mental illness in a given year. In the past decade, addiction to opioids has assumed center stage in policy discussions. Substance abuse is ravaging many communities and suicide rates have seen a dramatic rise. Yet, the resources dedicated to behavioral health have lagged those devoted to other types of health conditions.
This situation suggests a strong need for an evidence-based approach to making decisions that can exploit the power of information technologies and analytics to more effectively address behavioral health in the U.S.
In his closing remarks, Frank Stein stressed the importance of continuing this conversation on behavioral health analytics. He urged the 100+ attendees to make connections with others in the audience and reach out to speakers and conference organizers to keep up the momentum sparked by the workshop. Want to join the conversation? Contact Prithvi Rao at firstname.lastname@example.org and Ritu Agarwal at email@example.com.
For more information about the IBM Analytics Solution Center, visit: https://www.ibm.com/industries/federal/analytics
For more information about the Center for Health Information and Decision Systems (CHIDS) at Smith, visit: http://www.rhsmith.umd.edu/chids
For more details on the Smith School and IBM Analytics Workshop, visit: http://www.rhsmith.umd.edu/AnalyticsWorkshop
- Alissa Arford, Office of Marketing Communications