At Penn Medicine, a powerhouse combining University of Pennsylvania's esteemed medical school and extensive hospital system, CIO Michael Restuccia and Brian Wells, associate vice president of health technology and academic computing, are driving an ambitious big data agenda with the help of good IT governance.
Part one of this SearchCIO Innovator profile reported on two big initiatives: the technology behind PennOmics, a big data and analytics effort to deliver personalized care to cancer patients, and a new centralized high-performance computing cluster that allows Penn's hundreds of researchers to analyze large data sets. Part two of the profile delves into their work on data analytics and Restuccia's recipe for effective IT governance.
Restuccia and Wells are using analytics to track and improve operational performance. The IT work began in the health system's finance department, where dashboards were built that helped CFOs follow the money -- from analyzing which departments were making or losing money down to which individuals within those departments were high consumers or producers of revenue. "This was our entrée into analytics," Restuccia said.
Very quickly, however, under the auspices of the chief medical officer and vice president of quality, the data analytics work moved into patient safety and quality of care. Using tools from SAP, Oracle and IBM, the IT organization built dashboards that track patient safety and care in individual departments, right down to the nursing units, and measure performance against norms.
"We have come up with the series of dashboards that indicate around 140 quality metrics across the health system, and that's delivered to our hundreds of clinical managers every three days," Restuccia said. The data is exhaustive, covering infections of all kinds, patient satisfaction, the time between a diagnosis and when medicine is administered, mortality, and more.
When IT doesn't have to know nine different analytics tools, it allows us to spend more time with our end users to define their needs and requirements.
CIO, Penn Medicine
As with any business intelligence project, there were the usual challenges of working with users: from pinning down which areas should be covered and defining terms (e.g., what an infection is), to creating the right look and feel for the dashboard. Even harder to manage was the hue and cry from the clinical community when the data started rolling out -- the data was wrong, it couldn't be their nurse station, these were not their employees.
"It was a huge challenge each time we rolled out a dashboard in a new area. Now we are prepared to justify the data," Restuccia said.
The dashboards allow clinicians to drill down to the patient level to pinpoint, for example, when and where an incident that rendered a low patient-satisfaction score happened. The data is there, and -- most important -- the medical quality professionals in charge of the program own the data, with IT there as a "second line of support for technical problems," Wells said.
Three Cs of good IT governance
One of the most gratifying dashboards IT has built, Wells said, is a real-time monitoring system used to track leukemia patients in a clinical trial that involves extracting a patient's white blood cells, genetically modifying them to combat the disease and putting the modified cells back into the patient. The whole process is tracked, from fusion dates to cell manufacture through the rough couple of weeks that patients endure after being reinfused with their modified cells. "It's one of the few examples where we are taking real-time data out of the health system and linking it up to the clinical trial research and providing this dashboard at the same time," he said. "Generally, most patients are cured and come out the other side."
More on big data governance, analytics
Out with the 'HiPPOs,' in with the data scientists
Governance lessonsfrom a healthcare CIO
Tapping the collective IT IQ to drive innovation
These IT-enabled breakthroughs wouldn't happen without the Senior IT Governance Council and its backing of what Restuccia calls the three Cs: common systems across the enterprise, centralized IT and working in a collaborative manner with the end user community.
"They pay dividends on a daily basis," Restuccia said. "When IT doesn't have to know nine different analytics tools, it allows us to spend more time with our end users to define their needs and requirements. And at the end of the day, IT is a service organization."
This was first published in November 2013