UK HealthCare Exploring Ways Big Data Analytics Can Improve Patient Care

You finish some online window-shopping for a new pair of Nikes and move on to check your Facebook feed. Something to the right of the feed catches your eye: it's that same pair of Nikes you haven't yet bought, with a link to a website that has it in your size.

Big Brother? No, Big Data.

In The Digital Doctor, big data's daunting definition is: "High volume, high velocity and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery, and process optimization."

But Joe Labianca describes big data in simpler terms.

"Our use of computers, smartphones and other devices generates massive amounts of data, much of which can be used to customize your shopping experience and make life easier," he said. "For example, with Google, perhaps the world's leading corporate user of big data analytics, users can download their smartphone app that tells you how long it will take you to get home (taking traffic into account) and whether the books you ordered from Amazon have arrived on your doorstep."

According to Labianca, Gatton Chair in Management at the University of Kentucky's Gatton College of Business and Economics, companies around the world are rushing to adopt big data analytics as a means to streamline product delivery, improve the shopping experience and boost sales. And that, in turn, is catching the eye of the health care industry, where the health records of a single patient can easily top thousands of pages and a larger health system, processes about 10 million computerized transactions a day — twice the number of transactions that takes place every day on the NASDAQ.

Dr. Mark Williams, interim chief of the Division of Hospital Medicine and director of the Center for Health Services Research, is one member of the team committed to explore how UK HealthCare can use big data analytics to make the patient experience better.

"Health care is a contradictory enterprise, generating terabytes of data in the course of a month but still requiring a high level of human touch," Williams said. "The challenge for us is to find ways to use that data to help patients get better faster while maximizing efficiency and lowering costs — all without compromising the human element of the patient experience."

A visit to one of UK Chandler Hospital's Intensive Care Units (ICU) illustrates Williams' assertion. A single patient can be connected to as many as 12 machines monitoring blood pressure, IV drips, dialysis, ventilators and so forth. The nurse assigned to that patient must watch each and every one of those monitors for signs of trouble and coordinate care with each element in mind. Though technology has made health care better, it has also fostered new challenges for the people who are responsible for delivering it.

The greater UK health care enterprise already has big data analytics available. Jeff Talbert, professor in the UK College of Pharmacy, runs the UK Center for Clinical and Translational Science Enterprise Data Trust, a repository of health care data from a number of sources, including links to the state's Medicaid program, the Kentucky Cancer Registry and health data from the insurance industry. He sees data as a strategic asset with enormous opportunity to inform policy and change the way patients are treated, and points to patient research participation as a prime example.

"We can cull through the data to find patients who have a certain disease and might be eligible for a clinical trial, which can increase research participation and collect feedback about new treatments on a faster timetable," said Talbert.

But there are large pieces of it scattered around campus, and none of it thus far directly addresses the enabling technology for big data and patient experience. To tackle that, UK has recently brought acclaimed computer scientist and informatician GQ Zhang to campus to lead the newly formed Institute of Biomedical Informatics. Zhang is charged with integrating and leveraging large data systems across the academic and medical enterprises to improve patient care, research and education, creating what is increasingly known as a "Learning Health System."

"Many people refer to volume, velocity and variety when discussing big data, but I like to add two extra 'Vs' — vision and value," said Zhang. "If someone doesn't have the big idea that will help leverage the data in the right way to answer a relevant question or provide a new way to solve a problem, the untapped potential of big data will not be realized."

As an example, Zhang tells the story of a program developed by the mayor of the city of Boston to improve the driving experience in an unusual way: by fixing potholes.

Bostonians can download a free app to their smartphones and let the app run in the background while driving through town. The jolt of a pothole is sensed by the phone, which then transmits georeference data to the city's databases, requiring no active input from the driver. As increasing numbers of drivers bump through the same pothole, the accuracy of the pothole's location increases significantly and road crews are dispatched to the area to repair the pothole with pinpoint precision. Saves time, saves money and makes countless drivers happy.

The same principle can be applied to critical care, said Zhang. Currently, nurses collect basic data about an ICU patient from the monitors at periodic intervals and enter it into the patient's chart.

"A trend can be more important than an absolute value and this sporadic recording of data may not be responsive enough in spotting subtle trends that predict whether that patient is in need of immediate attention," said Zhang. "If that equipment recorded data second-to-second, we could potentially develop a real-time index or score for each patient in the ICU that would allow us to monitor subtle changes in vitals that are predictive of real trouble and respond accordingly."

Furthermore, says Zhang, because this data would be stored cumulatively, researchers may use the data to develop analytics that look at outcome trends for large populations of patients and provide new ways to improve patient care. Like the Boston pothole app, data gathered over time in a larger setting can help provide important information useful to improve real-life experience.

Dr. Peter Morris, chief of the Division of Pulmonary, Critical Care & Sleep Medicine at UK HealthCare, embraces Zhang's ideas, saying big data capture and analysis is crucial to process improvement in ICU management. Morris and Philip Eaton, DNP, director of Nursing -Medicine Service Line, will apply Zhang's techniques in the ICUs to simultaneously improve health outcomes while lowering the costs of expensive ICU care.

"If we can use data to, say, predict how flu season will spread and peak in Kentucky, we can anticipate the ICU resource allocation needs — for both staff and equipment — and have all the necessary personnel and equipment on standby and ready to deploy as needed," said Morris. "The same is true for day-to-day operations: we have learned patients admitted to the Medical ICU from outside hospitals peak between 2 p.m. and midnight, so we have adjusted staff levels to meet that demand."

Morris ticks off other opportunities: GPS locators on gurneys, for example, might help track wait times for imaging and other testing. "We need to look at ourselves every minute of every day to get patient care optimal and cost-efficient," he said. "Big data analytics will help us get the right care to people at the right time."

Media Contact: Laura Dawahare, Laura.Dawahare@uky.edu