Using data analytics to help improve health care provided by the Kaiser Permanente health system wasn’t the only topic IEEE Senior Member Michael S. Johnson talked about in our interview for my “Better Health Care Through Data” article. As director of utility-care data analysis for the Kaiser Foundation Health Plan, in Oakland, Calif., he also discussed some of the challenges he faces in working in this burgeoning field and how IEEE can make a difference.
What are some of the issues you are encountering?
At this point in the evolution of big-data technologies, it is honestly very difficult to tell whether one is reading a serious article about it or a marketing piece. There is rarely little difference between the two. Kaiser is not atypical in the sense that as we have dipped our toe into these waters. We find ourselves in the uncomfortable position of having to seek out unbiased, unvarnished, and unhyped technical advise about the pros and cons of these kind of technologies but having our only source of information be vendors. That’s not unusual for a new technology, but it’s a state of affairs that shouldn’t last very long.
Another concern is the state of research, which has evolved into a system where negative results almost never get published. This really hampers research in a fundamental way because it actually biases research in the long term. I think the same thing is certainly happening with big-data technologies—the only stuff that gets reported are the successes.
How can IEEE help?
IEEE historically has played a role in migrating beyond that because it has become a source of unbiased information about a lot of technologies. Its journals, conferences, and other events, like the 2013 IEEE Computer Society’s Rock Stars of Big Data conference where I was a speaker, have become sources of information about important technologies that are not tied to specific vendors or the marketing arms of companies.
An interesting exercise for IEEE to conduct in one or more journals would be to solicit its members to submit case studies that are anonymized (because nobody likes to admit they blew it) from those who have either failed, really struggled, or spent more money and time than they anticipated. I think that is a potentially valuable thing IEEE can do and would be possible in part because engineers are typically willing to do that. People in other disciplines are not comfortable in talking about failure but engineers are. That’s one area where IEEE can make a contribution—trying to balance the literature by helping people understand why some projects fail and what we can learn from their mistakes.
What can IEEE do to foster more efforts by organizations, governments, and other entities to use big data to improve their processes?
Right now there are still a lot of organizations sitting on the sidelines. People are extremely uncomfortable and it’s not even the financial investment. These technologies aren’t that expensive relative to the other things companies purchase. It’s much more about the human investment. To even pilot big data in a large company in a serious way you have to redeploy some of your scarcest resources, which are experienced people who know your data. They are already doing useful things with your data but you need to pull them off what they are doing and put them on a project that is pretty speculative. That represents a fairly substantial decision on the part of companies. They are really uncomfortable making that decision when the only thing they can do is compare the brochures they get from various manufacturers and technology companies about how great their technology is.
IEEE has an important role to play here too because it could foster the adoption of big-data technologies just by becoming a neutral third party that can help people understand them and get unvarnished information about successes and failures.
Read more articles about big data from our special report on the topic.