Wednesday, April 25, 2012

#ALGORITHMS: "IBM'S Watson Conquering Expert Medical Diagnoses"

The use of artificial intelligence (AI) to quickly sift and sort through millions of medical records, report, journal articles and other resources is repurposing the gaming-oriented Waston AI for making instant diagnoses of diseases that traditionally require weeks of visits to specialist, with no guarantee of success. R. Colin Johnson

Here is what IBM says about its progress at repurposing Watson for medical diagnoses: Memorial Sloan-Kettering Cancer Center (MSKCC) and IBM have agreed to collaborate on the development of powerful solutions built upon IBM Watson in order to provide medical professionals with improved access to current and comprehensive cancer data and practices.

MSKCC oncologists will assist in developing IBM Watson oncology applications utilizing patient’s medical information and synthesize a vast array of continuously updated and vetted treatment guidelines, published research and insights gleaned from the deep experience of MSKCC clinicians. This will provide an individualized recommendation to physicians and provide users with a detailed record of the data and evidence used to reach the recommendations.

Oncologist will have access to relevant cancer care information in order to personalize diagnosis and treatment plans for their patients. Creating an outcome and evidence based decision support solution will help doctors create an individualized course of action for patients based on current evidence. Physicians will be able to confirm a hypothesis or seek alternative ideas based on confidence weighted responses.

In general, healthcare organizations are challenged with unlocking the vast and growing stores of patient information, clinical data, research, and market insights to make more informed decisions and improve patient outcomes. Medical data is doubling every five years and it is estimated that 80 percent of medical information resides in an unstructured data format like nurses notes, clinical reports and doctor dictated comments. One challenge is to provide a complete view of all the available data so that trends and patterns can be determined and acted upon.

Ready for Watson is a program designed to prepare and accelerate client development toward a Watson Solutions deployment. The program identifies steps to be taken along a progression path and the associated enabling technologies to consider. One such complimentary technology is IBM Content and Predictive Analytics for Healthcare which provides healthcare organizations a means to improve patient outcomes. Natural language processing is paired with predictive root cause analysis providing a unique forward looking view of the data.

In the case of Seton Health Care Family, they had more than two million patient contacts a year with nearly a dozen patient records per contact, some records were electronic, but many hand written notes. Clinicians required an aggregated view of the patient to drive informed decision making. Seton Health Care Family was able to mine the unstructured data by using natural language processing and search technologies to produce actionable results regarding re-admission risks for congestive heart failure patients.

IBM Content and Predictive Analytics for Healthcare empowered Seton Health Care Family to turn insight into action for improved patient care and built a critical foundation for the future work planned for a full Watson deployment.

Operational and clinical insights can be attained by following the path to Watson.
Further Reading