|By David Tishgart||
|October 6, 2012 01:00 PM EDT||
Data is quickly becoming one of those certainties in life, like death and taxes. It'll always be there, and like the Once-ler's Thneed factory from The Lorax (sorry, I have kids), data figures to keep on biggering, and biggering, and biggering and biggering.
More data means more knowledge, greater insights, smarter ideas and expanded opportunities for organizations to harness and learn from their data. Banks, retailers and even government are embracing big data, but while IDC estimated the big data market at $2.2 billion in 2011, only 6% of that investment came from health care.
On the flip side, a 2011 report from McKinsey Global Institute suggests if health care in the U.S. used big data to drive efficiency and quality, the potential could be more than $300 billion in value every year.
So big data investment by health care is small and growing, but the potential is significant. To realize that potential requires data to be secured and protected at all times, assuring information accuracy and integrity. A major concern with big data systems is their inherent lack of security. A typical NoSQL data store lacks a number of key security features that are available in traditional databases or provided by a third-party security vendors. This is going to be a big issue and potential barrier to entry for big data moving forward.
Consider this case of an urban health care facility just outside of Washington D.C., where the emergency room was experiencing an alarmingly high rate of returning patients. To determine the root cause of the situation, researchers sifted through data collected from more than 300,000 ER visits. By correlating seemingly unrelated information, they were able to surmise that the length of stay of a patient was a key factor in determining whether they would make a return trip to the ER. Now doctors can determine the likelihood that a patient will need to be readmitted to the ER and tailor their follow-ups accordingly.
Just imagine if this data had been skewed either by a rogue insider with access to the data or by a malicious actor outside the hospital. Researchers would have come back with an entirely different view of the problem and perhaps a less effective solution.
This underscores the importance of securing big data through a layered approach that employs firewalls, authentication, patch and configuration management, antivirus and event monitoring tools.
Ultimately, the safest thing a health care provider can do to maintain data integrity, limit access to sensitive material and enable HIPAA-HITECH compliance is to encrypt all data at rest. By encrypting data, storing the keys in a separate, secure environment and enforcing tight controls governing who (or what) can access the encryption keys, organizations can create a hardened barrier around their sensitive data.
In the event of a device theft - currently the most common type of data breach in health care due to the high number of mobile devices storing unprotected health records - encryption ensures data cannot be read by unauthorized parties, while access controls restrict data from third-party vendors like cloud or SaaS providers.
To secure regulated HIPAA data stored in popular big data stores such as Hadoop, it's important to use a Linux encryption tool that offers the aforementioned features and does not impact the performance of the rapid-fire MapReduce queries that make big data technology so valuable in the first place.
The bottom line is, there needs to be some middle ground where patients feel their protected health information is secure, while hospitals and research organizations have the access and ability to conduct big data analyses that improve the quality of the care they're providing.
After all, as my friend the Once-ler once might have said, good health care is what everyone, everyone, EVERYONE needs.