Backwards as it may be, supercomputers in healthcare are starting with death. At Beth Israel Deaconess Medical Center in Boston, there is an on-site supercomputer which takes information from 250,000 patients over the past 30 years. The machine takes vitals of each patient every three minutes and can track trends and abnormalities. It also has the ability to predict a patient’s chance of death in the next 30 days with a 96% accuracy rate according to Dr. Horng, the team leader for research and development of the computer. Despite this grim use-case, patients should be optimistic.
Housing supercomputers in hospitals is a logical next step in the era of big data. Few industries generate more data than healthcare and even fewer have as much to gain from data analysis. With the help of supercomputers, hospitals will be able to make smarter decisions about treatments, identify rare diseases, and recognize early warning signs which could save time, money, and lives.
IBM’s supercomputer (and Jeopardy! champion) Watson is being used in 14 hospitals around the country and Mount Sinai Hospital in New York City have built their own. The machine at Mount Sinai, named Minerva in honor of the Roman goddess of wisdom and medicine, took data from 30,000 patients and created a map of those with Type 2 diabetes and opened up the possibility that there are more types of diabetes that we know little about. It’s not hard to extrapolate and imagine how much of our knowledge is incomplete.
One of the most promising areas of development is DNA analysis. Being able to parse through thousands and hopefully soon millions of patients’ genetic code with the speed of a supercomputer could lead to enumerable breakthroughs in the medical industry.
As more hospitals adopt supercomputers, the amount of data will grow exponentially. And as the number of data points increase, the more accurate the computer models will become. Together, the entire medical community could create a cloud-based database of patient data from around the world that could help prevent potential epidemics or identify and treat at-risk patients earlier than ever before.
All of this data is only useful if it’s organized. Image-enabled electronic health records (EHRs) is the next step towards this attainable goal. Using IT infrastructure, it’s now possible to have all of a patient’s medical information, from blood tests to growth charts to MRIs, in one place. DICOM Grid gives hospitals the ability to house every piece of patient data, including images, in one place. On a micro level, this gives doctors all important information about a specific patient in one place which will make eliminate inefficiencies and help improve diagnoses. On a macro level, all of this information on the cloud can be added to a database and compared to other similar patients to advance our medical knowledge.
In many industries, big data is a buzzword with little substance and little practical use. That is not the case in the healthcare industry. Collecting, organizing, and analyzing all of the available data will have unlimited benefits. It’s time for supercomputers to #PutData2Work.