"Hospitals in general have a lot of different data sets, and it's easy to extract data if you're just looking within that set," says Steve Maxwell, the chief administrator for the department of emergency medicine at the University of Michigan Health System. Most organizations have the ability to find the average length of stay within a department, the average turnover time in the surgical department or the demographics of the hospital's patients, he says.
However, bringing these data sets together can provide a more comprehensive picture of not just one patient's experience but of the efficiency and efficacy of the organization as a whole. By employing SurgeryMetrix software from ArborMetrix (founded by a University of Michigan faculty member), the University of Michigan Health is able to aggregate data from disparate silos to get that more complete picture of the organization's operations. "SurgeryMetrix enables us to cross different systems and look at data that we previously couldn't link," says Mr. Maxwell.
One of the most recent analyses conducted by the University of Michigan Health System was in attempt to improve operating room utilization. Using SurgeryMetrix, officials aggregated information that let them compare a surgery's scheduled time to its actual time, the procedure's outcome and the total expenditures.
"It let us say, 'Because a physician is faster or closer to their expected time, do we see a better outcome in terms of readmissions or complications?'" says Mr. Maxwell. "And further, if we're looking to optimize cases coming into the system, do we want to say one cohort of patients should be driven to a certain subspecialist… and would [those subspecialists] be faster or have lower complications if they just did that procedure?"
Currently, SurgeryMetrix is being used in several surgery departments at the health system to provide leaders and physicians alike a glimpse into surgeons' and departments' overall performance. The software will also play a role in allocating time slots in five ORs currently under construction, helping the health system ensure the new spaces are used as effectively as possible.
ArborMetrix software is also being used by one of Mr. Maxwell's colleagues, Michael Gaies, MD, a pediatric cardiac intensive care physician at the University of Michigan C.S. Mott Children’s Hospital. Five years ago, Dr. Gaies led a team of researchers in establishing an international consortium of pediatric intensive care units to share data and develop a common clinical registry with which to study quality and outcomes. The resulting group, the Pediatric Cardiac Critical Care Consortium, is continuing to collect and analyze relevant data to improve care quality, now aided by another ArborMetrix product, RegistryMetrix.
There are similar data registries around the country with goals similar to PC4's, says Dr. Gaies, who also serves as PC4's executive director. However, these other registries "don't provide actionable, timely data — ArborMetrix allows us to access data in real time and see how initiatives are performing and institute local quality improvement efforts," he says.
After just eight months of use, Dr. Gaies says use of the software has already provided insights into complications most likely to occur in certain age groups and generated many hypotheses about where participating providers can focus early and preventive efforts to reduce complication rates.
"One of the most notable things was seeing the frequency with which patients who are taken off breathing machines go back on," he says. "It was really high in small babies. This tells us we've not quite figured out how to determine if [a baby] is ready to be taken off… we've known it was a problem, but this gives us a number to focus on."
Both physicians see the potential of ArborMetrix software to change care delivery at the University of Michigan as well as throughout the industry as a whole. "We're just beginning to scratch the surface" of the potential of having all this data brought together, says Mr. Maxwell. "It was so frustrating to know there was a problem and have the data but not have it all linked… this [software] definitely makes my job easier."
More Articles on Big Data:
How Big Data Can Help Reduce Costs: 6 Scenarios
How to Succeed in Big Data Without Really Trying: 4 Steps From Booz Allen
Survey: In Analytics, Data Variety More Challenging Than Volume