A Breakdown Of The 'Big Data' Era
A new health trend, 'Big Data' can be found in a variety of health and science specialties. Dr. Andrew Su, an associate professor of molecular and experimental medicine at the Scripps Research Institute, describes the concept, tactics and purpose behind researchers gathering and analyzing data.
On what it actually means
"People probably characterize Big Data by three properties. The first, and most probably obvious, is by volume. The sheer volume of data we are able to generate these days. The second piece is velocity, how quickly data is being produced and also how quickly we expect that data to be analyzed. More and more we get into the idea of real-time analytics. The third characteristic we hear about is variety. You’re not just thinking about data of one different type. You are thinking about data from many different types."
How it is used in genomics
"The Big Data that I think about most is the idea of molecular Big Data. This is molecular profiling of us, of humans, in the context of health and disease. In that sense you can think about DNA sequence being the most common form of Big Data. So the Human Genome Project run by the National Institutes for Health was able to produce the first draft human genome sequence in 2002. And that was the culmination of 20 years of work. That same amount of work, generating one human genome sequence, is now something that can be done in less than a week."
How it is used by health systems through electronic records
"Certainly we have a sense of medical informatics. If we think of all the health care data we have access to from patient visits to lab tests and things like that. We have personal sensor data, where more and more people are tracking their activities or tracking their heart rate, their fitness and so forth. That’s certainly a source of Big Data."
How it could help improve the future of health
"On the plus side, there’s a really nice study I like out of Stanford University. They analyzed their health records looking for adverse effects for different drugs. They studied Viox, this drug that was actually pulled off the market because people who took Viox had an increased risk for heart attack. So what they did was they retrospectively mined all the health data they had in their system. They found that if they just looked at who took the drug and then within the several months after that who had a heart attack, they were able to essentially detect this adverse effect in their health records much earlier than the FDA was and the voluntary reporting system that we typically use to detect these effects. So that potential is a real big positive in terms of electronic health records."
Dr. Andrew Su is an associate professor of molecular and experimental medicine at the Scripps Research Institute in La Jolla, California.