Feb. 28, 2016
The current focus and media frenzy over big data in healthcare leaves many clinical and management professionals at a loss to understand practical applications. With the emphasis on value based purchasing and outcomes benefit, how does big data analytics create return on investment, cost savings or enhanced quality? An emerging company, Ayasdi, provides a new model for data analytics and gives practical examples of their success. The burgeoning field will dramatically alter our delivery of healthcare.
The traditional creation of a relational database often involved primarily claims which could be queried regarding relationships of the information. For instance what was the age distribution or payer mix of patients having a CT scan of the head or which doctor orders the most CT scans of the head. The inherent difficulty of the analysis is twofold. The data set often does not contain the integration of the electronic medical record and pertinent clinical information. More importantly the user interface required a query of the "relevant " question. The biggest concern is that you do not know what you do not know! What is currently being devised are sophisticated algorithms which topographically map the data and identify previously unrecognized trends and relationships.
An excellent practical example involved the data analysis at Mercy Hospital system who looked at their total knee replacement data which represented $50 million dollars in revenue to the institution. When they examined length of stay (LOS) at their multiple hospital sites, the algorithm identified a small group of physicians who, based on some early research results, were using pregabalin preoperatively. Remarkably the drug indeed significantly reduced LOS. The system savings were dramatic.
This simple previously unrecognized highly desirable outcome was unearthed through big data, resulting in better care, shorter hospitalizations, happier patients and costs savings.
The additional applications of this big data analysis are protean including identifying best practices, clinical outcomes and finding efficient clinical pathways. Physicians no longer need to decide best practices or seek a third parties benchmarks but can be presented with their own data as well as millions of other patients to create safer , better quality more efficient health care delivery. Identifying best drugs, diagnostic testing and preferred devices for specific diseases, symptoms and procedures becomes easy. Performance measures both clinical and financial are readily available.
By identifying what we could not otherwise know, big data analytics are poised to revolutionize health care delivery.
Nicolas Argy, MD, JD