Big Data and Healthcare: Corrolate, Root Cause, Operationalize & Disrupt

I have to say that it’s been massively fun lately at EMC talking about the impact of Big Data Analytics on vertical markets. It has been fun learning from Chuck‘s experience / scars associated with advocating for disruptive change in the IT market.

Recently, Chuck blogged about the emerging ACO’s and potential disruption that information and cost/benefit aligned markets can provide to markets like healthcare.

I commented on his blog, and wanted to share the insight with my audience as well:

Thanks for your insightful “state of healthcare” review. I find that there is one more facet to healthcare that makes the data/information analytic landscape even more compelling, and that’s an interesting book by David Weinberger, Too Big to Know.

Dr. Weinberger brings forward the notion that books and traditional learning are discontinuous. And that there is emerging, due to the hyperlinked information universe, a massive ecology of interconnected fragments that continually act with the power of positive and negative reinforcement. The net is that traditional views, or in David’s vocabulary “Facts” are largely based upon constrained reasoning, and that as the basis of this reasoning changes with the arrival of new facts, so should their interpretations.

There are a number of entities sitting on top of 30years of historic clinical data, the VA, certain ACO’s and certainly academic medical centers that can chart many of the manifest changes in treatment planning and outcome, but may themselves, because of their constrains present significant correlation. But just the same, many not be able to discern accurate causality due to a lack of completeness of the information base – maybe genomics, proteomics, or the very complex nature of the human biological system.

The interconnectedness of the clinical landscape is of paramount value in the establishment of correlation, and derived causality, and that with the increase of new information, traditional best practices can be [and needs to be] constantly re-evaluated.

I believe that, this lack of causal understanding, due to highly constrained reasoning, has lead to many of the derogatory statements about todays outcome based care model. As todays outcome measures are based more on correlation than on causality, determining causal factors, and then understanding the right control structures should improve the operationalization of care. I further believe that, as the availability of substantially more complete clinical information, across higher percentages of populations, will lead to improvements in these outcome measures and the controls that can affect them. The net result of getting this right is improved outcomes at decreased costs, and effectively turning the practice of medicine into a big data science with a more common methodology and more predictable outcome. In effect, full circle, operationalized predictive analytics.

The opportunities for market disruption are substantial, but there remain high barriers. One constant in being able to exploit disruptions is the right visionary leadership who has the political capital and will, but also is willing to be an explorer, for change is a journey and the route is not always clear. Are you a change agent?

Join me at the 2nd Annual Data Science Summit @ EMC World 2012

Leave a Reply