Big Data & Analytics Heroes

Boris Bershadsky

Director, Outcomes Analytics, UVN

Boris Bershadsky, director of Outcomes Analytics at UMN declares that “more progress is needed to handle massive amounts of information in real time.” See what other thoughts this week’s Big Data and Analytics Hero shares with us on how big data and analytics are affecting healthcare and the world at large.

What is the market still missing for big data and ­analytics to really deliver ROI?

Health Information Technology (HIT) is famous for not computing ROI in general. Relative cost of HIT as a component of total cost of care is unknown. As a result, effects of technologies traditionally associated with "big data and analytics" are unknown also - ROI cannot be measured or evaluated. This is very sad because from the social perspective HIT and patient care pull money from the same bucket (zero-sum game) and without knowing their relations HIT systems cannot be optimized.

Should tomorrow's generation acquire analytic skills no matter the degree? Why or why not?

It depends how we define the term "analytic skills". If analytic skills are viewed as an ability to apply statistical methods, then NO. If they refer to logically consistent statistically valid descriptions, then YES. The next "lay" generation should know the methods limitations but not the methods themselves. Currently these two approaches - methods and methods limitations - are not separated.

Do you think big data and analytics will handle the data growth in 10 - 15 years or would we need another shift in technology? Why?

Technological foundation is already here - smart phones/tablets, cloud computing and in-memory calculations. For the next 10-15 years it should be sufficient to handle massive data in real time. More progress is needed to handle massive amounts of information in real time. Logical models -- how to properly convert massive amounts of data into useful information supporting real-time decisions -- are more difficult. This direction can be defined as operational data warehousing - it should enable real-time application of state-of-the-art learning machines in distributed environment with hundreds related sources of data. The next shift should happen here.

View all of our Big Data & Analytics Heroes here on the Hub, and look out for next week's Hero: Edward Kwartler, director of advanced analytics at Liberty Mutual Insurance.