Today, business and other organizations collect a staggering volume of data. However, the data is not useful when they don’t know what it means or how they can use it. Helping organizations extract practical information and insights from mountains of numbers is what business analytics professionals do.

Business analytics can be used to monetize data through information and insights that can lead to action and results, including revenues and profits for companies, said Lara Khansa, professor of business information technology.

Analytics at the center of decision making

Khansa, who gave a campus talk on the topic earlier this year, does research in health care analytics and teaches graduate courses in Virginia Tech’s highly ranked Master of Information Technology program.

Analytics professionals, she said, are at the center of decision making today, helping to guide organizational policy, strategy, and operations, and enhance business competitive advantage.

Their roles go beyond the arenas of manufacturing, marketing, and sales. In sports, analytics is being used to boost game tactics and player training; in political campaigns, to model outcome scenarios and target voter groups; in health care, to evaluate physician performance, reduce waste, and improve patient care.

Many of today’s cars are “internet of things” machines, Khansa noted. “Every time you make a turn, brake, choose a channel, everything is being recorded and stored somewhere.” Auto manufacturers may sell the information to banks, insurers, or media companies. Radio stations and politicians, she said, might be interested in information about driver demographics, music tastes, and listening times.

“Analytics is everywhere,” Khansa said. “It’s really about identifying a need for change, and analytics professionals are agents for change.”

She points out the four types of analytics: “Descriptive analytics looks at what is going on or what happened. Diagnostic analytics explains why it happened. Predictive analytics suggests what might happen next, and prescriptive analytics recommends how to make a desired action happen.”

Investigating ER overcrowding

Khansa and her collaborators on a Virginia Tech-Carilion Clinic research team used all four analytics techniques to investigate overcrowding in the ER of a local hospital.

ER overcrowding occurs when patients are not processed within a given period, resulting in longer waits, care delays, and patients not receiving the required level of care, the research team wrote in an article in Journal of Operations Management.

A better understanding of how emergency departments perform during disaster-level overcrowding would strengthen their resilience to such events — that is, their capacity to resist these events and recover from them more rapidly. It would help lead to a better patient experience and operational and resource allocation effectiveness.

The researchers developed a modeling framework in collaboration with the emergency department at Carilion Clinic’s Roanoke Memorial Hospital. Carilion Clinic is the only level-one trauma center in the region and serves nearly one million people in southwestern Virginia.

The hospital provided data from actual disaster-level overcrowding events over a one-month period. “Our techniques were so powerful, we were able to make a big impact,” Khansa said.

“With only a single month’s worth of data, we were able to identify the most impactful root causes of the problem, which were, in this scenario, high patient throughput and insufficient staff to triage and care for this many patients.”

The other members of the research team were then Virginia Tech doctoral student Zachary Davis (Ph.D., BIT ’18), now an assistant professor of decision sciences and information management at Jacksonville University; Christopher Zobel, professor of business information technology at Pamplin; and Roger Glick, an instructor at the Virginia Tech Carilion School of Medicine and former emergency management consultant for Carilion Clinic.

Their research project identified other contributing factors to overcrowding. It resulted in a predictive model to determine when overcrowding is likely to occur and to quantify the impact of each contributing component.

“Our key practical contribution was to provide Carilion Clinic with a decision-support process to predict overcrowding and to improve resilience to such events,” Khansa said.

But the impact of the research can be more far-reaching. Emergency departments worldwide experience the effects of disaster-level overcrowding “on a recurring and surprisingly regular basis,” the researchers noted in their study.

Their model can help hospitals everywhere assess the effectiveness of various intervention strategies and determine what changes they can make that would have the greatest impact on improving overall resilience. Their ability to manage or cope with overcrowding would be strengthened, and patient outcomes improved.

At its core, business analytics is about people, Khansa said. “If you don’t take people into consideration, you won’t be a good analytics professional.”

– Sookhan Ho

Loading player for