Experts Weigh In on Big Data Tools
Everyone needs to take big data seriously.
The marching order comes from a report from the McKinsey Global Institute. Big data refers to colossal data sets made up of data collected from the past combined with the data now being recorded at a rapid rate due to new technology that tracks and stores it like never before. More than $300 billion in value could be created in the U.S. each year if the health care sector creatively and effectively harnesses its power, and health care expenditure could be reduced by 8%, according to MGI. As certain as we are that big data will achieve big things, until recently, researchers and innovators haven’t been able to say with much certainty how they’re going to do it.
During a panel discussion on big data at the Health 2.0 conference Tuesday, Indu Subaiya pointed out that although big data has been available for many years, developers are just starting to come up with tools that help us make sense of it. “This year at Health 2.0 I think we’re beginning to see technologies really for the first time doing that intelligent mining, archiving, presenting and visualizing of this information,” Subaiya said.
There’s a calling for health 2.0 companies to build upon and create new technology so that we can meaningfully use data. “If there’s one word that captures the whole area, it’s ‘opportunity,’” said Paul Wallace, senior vice president and director of the Center for Comparative Effectiveness Research at The Lewin Group. Wallace sat on the panel Tuesday that aimed to discuss how big data will transform health care. What Wallace and other panel members ended up doing was laying out really good advice for innovators about how they should approach the development of these kinds of tools, devices and applications that will derive meaning, and ultimately, value from data.
Here is a summary of each panelist’s vision for big data tools which they say will help us:
The majority of what we know about what works and what doesn’t work in health care is brought to us by academic researchers. While past research has taught us a lot, the information that comes out of best practice studies is limited because it is so general. Tools to help us analyze big data should help research to get specific. So the questions big data can help us answer aren’t only about what works, but what works for a particular patient, Wallace said.
Learn from the everyday
Experiments aren’t just happening in clinical trials anymore. We’re helping to run experiments all the time as we generate data just by going about our everyday lives, said Carol McCall, Principal at Health Analytics Innovations. The good news for researchers is that we’re starting to capture the data now, and big data tools will help us learn from daily events. “What we need is a new kind of analytics that allows us to learn at scale from our everyday experiments,” McCall said.
Do what we can’t
Humans are limited when it comes to analyzing big data, but supercomputers will allow us to sift through an inconceivable amount of data to test one hypothesis. Colin Hill, CEO & President of GNS Healthcare described a machine that can comb through a data set of trillions of drug claims to find 250 occurrences of adverse drug reactions, a relatively small number. That kind of identification is beyond human capability, but once specific events are found, humans can go in and do the analyzing, interpreting and validating that computers cannot.