Big data is the big game these days. And by all accounts, business is booming. Analysts like IDC state that the big data technology and services market is growing at a whopping 31.7%—or about 7x faster than all of IT.
Part of why its growing is because it is has become so real-time and user friendly, its almost addictive. We are making discoveries so quickly now that it is feeding an insatiable hunger to find out more, and to use that information for good.
Reflecting on this past week in just the healthcare market, it is easy to see how much we are finding out and how big of an impact big data is making on our everyday lives.
Studying Asthma, Flu and Bronchitis in a Day
Last week, we posted results of a 24-Hour Data Science Code-a-Thon hosted by Kaiser Permanente last May. At this event, several companies, including Pivotal, spent one day using the latest and greatest Hadoop technologies to study respiratory conditions. Although there are other stories for sure, our experience alone is example enough of just how rewarding data science is these days.
In just 24 hours, using already established data sets and the talents of just 5 of our big data experts, the Pivotal team was able to describe 3 data stories and write two applications to help physicians. Without specific expertise in the pulmonary healthcare field , they were able to:
- Link geographic areas with greater than expected asthma prevalence to higher ozone levels for extended periods during the summer.
- Find that 17% of the asthma patients do not show up to the pharmacy to pick up prescribed medications, and are then 13% more likely to have an asthma related hospitalization.
Admittedly, more study is required in each of these areas. However, these kinds of real-time rewards in data science are giving scientists quick fodder to justify further spending and research across the board. And with the ability to discover conclusive facts in one day, science should expect to drive further study more efficiently and produce results that impact our lives faster than ever.
Predicting Outbreaks
Also in the news this week, IBM teamed up with some researchers from Johns Hopkins University and the University of California at San Francisco to study how to predict outbreaks of dengue and malaria.
Each year, dengue infects 100 million people and, thanks to global travel patterns, has even recently has shown up in Texas and Florida. Malaria infects over 200 million people and costs Africa alone over $12 billion for the cost of illness, treatment, and premature death.
Now, thanks to their efforts we have the chance to identify outbreaks early, to better contain its spread, and allocate resources more effectively. Using an open-source modeling application dubbed the Spatio Temporal Epidemiological Modeler (STEM) created by IBM, researchers will have open access to use any kind of data and quickly correlate it with disease data. This application is planned to be released to the public on Oct 15, but in building out the application, researchers found they were able to see how changes in local climate and temperature affected the spread of the disease. Now they can use that data to figure out where the next outbreaks will be.
Making Hospitals Proactive & Fixing Our Healthcare Problem
Mount Sinai Medical Center is teaming up with the original data scientist from Facebook, Jeff Hammerbach, to prevent hospitalization and lower healthcare costs. With $3 trillion in healthcare spending a year, it is important to remember that each hospital bill is paid by procedure—meaning right now, hospitals only make money if you get sick. This is a flat-out backward economic dynamic that many believe explains the root of our healthcare mess in the United States.
Part of a larger experiment that is overseen by the federal agency in charge of Medicare, Mount Sinai is joining over 250 US doctor practices, clinics and other hospitals to harvest their data for good. If they succeed, we all stand to win.
They are aiming to use big data to predict when you are going to get sick, and then use that information to better advise you—saving you additional pain and healthcare costs.
The pilot study at Mount Sinai targeted reducing readmission rates was total success, reducing readmission rates by a half. To do this, Hammerbach and team built a computer model that incorporated disease, past hospital visits and even race to determine a patient’s risk. The results would provide a list of high-risk, chronically ill patients that needed more follow up. The hospital staff could then diligently call and keep track of these patients and work to avoid letting them get so sick they required hospitalization again.
Essentially, this is what the family doctor used to be able to do when he did house calls and remembered your name. Now doctors see a different patient every 12 minutes, and it becomes impossible to keep track. Technology helps those doctors, nurses and all their staff to work together to keep up and keep you healthy. (Note: I am pretty sure that 12 minute statistic is from Grey’s Anatomy, so I am not sure if its true, but it’s close)
And it doesn’t stop there for Mount Sinai. Clearly, they see the future of big data across all areas of their business, including research. Together with data scientists like Hammerbach, they are also using genetics and geography to do some interesting research on diabetes. Essentially, they are looking for “hot spots” of disease outbreak and then applying genetic research to just those areas to better profile what genetics are involved and target better guidelines for doctors to follow for each patient. Its both helping to refine research as well as creating more tailored medical therapies for patients.
I am sure it doesn’t stop here for medical science. There have probably been more that I missed. But I know one thing for sure—I can’t wait to see what’s in store for next week. This stuff is so addicting!
Editors note: This post has been updated to more clearly that the Code-a-thon had results published last week, but the work was done in May. It is assumed all three of these stories had work that was done in previous weeks or months, however their results were all discussed in the past week.
