The healthcare industry has always generated vast amounts of data, but only recently have practitioners discovered ways to convert patient information into strategies for early intervention, enhanced healthcare delivery, and predictive medical care.
The timing couldn’t be better. According to advisory firm McKinsey & Company, healthcare expenses now represent 17.6 percent of GDP – nearly $600 billion more than the expected benchmark for a nation of the United States’ size and wealth. By slicing and dicing data sets, the healthcare industry stands not only to improve patient care, but also significantly reduce costs.
Just ask Dr. Tim Buchman, director of the Emory Critical Care Center at Emory University School of Medicine. Buchman is currently spearheading a streaming data analytics research project in Emory’s intensive care unit (ICU) that’s helping healthcare providers analyze reams of bedside monitor data in real time.
Typically, bedside caregivers are inundated with dozens of continuous inputs and readouts charting everything from heart physiology and respiration to brain waves and blood pressure. Unfortunately, following every squiggle, beep, and number displayed on each computer screen is a significant challenge for today’s time-strapped doctors and nurses.
However, by combining IBM’s real-time streaming analytics system with Excel Medical Electronics’ bedside monitor data aggregation application, Emory is able to capture data in motion, identify patterns in physiological data and alert healthcare providers to serious complications like sepsis, heart failure or pneumonia. More than 100,000 data points per patient per second are collected and analyzed in real time, enabling clinicians to make split-second medical decisions.
“Current-generation ICU monitoring systems are designed to simply present data in the moment and then have it vanish from the screen,” said Buchman. “You’re looking at about a six-second snapshot.” Real-time data analytics, on the other hand, “presents an opportunity to create situation awareness in real time for the sickest patients in the hospital,” he said, adding that early detection and intervention can lead to huge cost savings for the healthcare industry.
Real-time data analytics isn’t the only big data tool being used by the healthcare industry to improve patient care and cut costs. Geospatial data is also finding its way into the hands of healthcare practitioners and researchers.
The Louisiana Department of Health and Hospitals, for example, recently teamed up with geographic information systems (GIS) software vendor Esri to map epidemiological issues, such as babies with low birth weights. Traditionally, researchers have relied on anecdotal evidence to flag hot spots for poor birth outcomes. However, using Esri’s GIS mapping software, the LDHH plugged in 354 points of data for every live birth in Louisiana – information captured by Louisiana’s Vital Records Registry — and combined this exhaustive data collection with the mothers’ residential information.
Rather than simply group incidences of low birth weight by basic markers such as ZIP code, Esri’s sophisticated algorithms identified clusters among locations and then generated a map based on this data. For example, by crunching low birth weight records with geospatial data, the LDHH discovered correlations between low birth weight rates and crime-riddled neighborhoods.
“What GIS data helps us do is isolate areas [of low birth weights] once and for all using something more than anecdotal information,” said Ryan Bilbo, GIS manager at the LDHH. “Now we can quantify how many births occurred there and how many are poor birth outcomes.”
By flagging areas of low birth rates, preventative healthcare measures can reduce the number of high-risk births, thereby cutting healthcare costs. Yet Bilbo said the challenge is anonymizing personally identifiable information, such as a mother’s residence, for public consumption without obfuscating the findings. After all, having to conceal a mother’s location won’t do much to help policymakers determine which areas of a city are in the greatest need of medical assistance.
“That’s the key,” said Bilbo. “It’s about taking that personal health information and converting it into something that’s aggregate without doing damage to the data.”
Buchman agreed that privacy and security challenges lie ahead. “As we move forward in our nation, we are going to have to find ways to share data safely,” he said. “Sensitive data in other domains, like law enforcement, can be shared safely across widely dispersed entities.” To protect patient privacy, Buchman said, his research project’s data is stored on site.
Another challenge is getting healthcare practitioners up to speed on today’s big data tools. Reading analytics systems and geospatial renderings requires technical skills that may be outside the purview of your average healthcare practitioner.
“We have not done the job that we need to in medical education to educate our young men and women about biomedical informatics,” said Buchman. “It’s a knowledge domain and a skill set that we’re behind in.” Nevertheless, he added, “it doesn’t in any way make these tools any less useful.”
Cindy Waxer, a contributing editor who covers workforce analytics and other topics for Data Informed, is a Toronto-based freelance journalist and a contributor to publications including The Economist and MIT Technology Review. She can be reached at email@example.com or via Twitter: @Cwaxer.