In an AACC session titled “Big Data Solutions that Unleash the Power of Your Lab for Patients, Payors and Providers,” Sonny Varadan, MBA, of Nichols Management Group discussed the challenges and benefits posed by the incorporation of big data into laboratory practices.
With the introduction of big data—a term used to describe any data set so vast or complex that traditional data processing tools are no longer useful—the purpose of analytics has shifted from simply providing information to guiding optimization. As the field of analytics has deepened in sophistication and complexity, the discipline has moved from processing historical data to shedding light on the future. The four phases progress as follows:
- Descriptive – using data to describe what happened
- Diagnostic – leveraging data to diagnose why an event or outcome took place
- Predictive – building data models to predict what will occur next
- Prescriptive – extrapolating data to determine how to generate a certain outcome
Unlike business intelligence, which is used to find answers to questions that already exist, big data can help reveal questions users didn’t know they needed to ask. Big data provides a foundation for business intelligence, making the two approaches complementary rather than redundant.
In some cases a lab’s existing clinical data warehouse and tools may be sufficient to process big data, provided the data set in question is a modest size. Any effort will require access to computing power in the form of server forms, memory, and CPUs. More importantly, labs need to invest in data scientists or applications that have the time and resources to perform the statistical analysis, data mining, and retrieval processes necessary to identify trends, figures, and other relevant information.
Documented cases have already made the case for tapping big data to streamline processes and save costs. The Kaiser health system translated data from its clinical repository, comprising 9 million patient records, into $1 billion of savings in reduced office visits and lab testing. The University of Pittsburgh Medical Center aggregated information from 200 different sources to identify patients at risk for kidney failure based on barely discernible changes in their lab results.
With healthcare organizations increasingly called on to do more with less, Varadan says, big data can help laboratories decrease costs, increase quality, communicate with patients in a more timely manner, offer providers personalized care, and collaborate with payors for quality outcomes.