ARTICLE FOCUS:

• Watson today
• WellPoint pilot
• Potential design tool

Imagine a physician's tool that could evaluate in minutes, or even seconds, data from connected health devices and implantable devices, as well as the patient's complete medical history and all available medical literature, including medical records, texts, journals, research documents, and even ongoing clinical trial results. Then, this same tool would take its analysis and suggest possible diagnoses for the physician to consider.

Such a tool recently appeared on national TV in the form of IBM's Watson, the talking supercomputer that beat out the all-time "Jeopardy!" champions.

Today, IBM researchers are enabling Watson to take on the role of an assistant, equipping it with vast amounts of medical information, clinical trial data, medical research, and even data about patient populations. IBM Watson's breakthrough understanding of natural language and relationships among complex text-based terms could significantly enhance the delivery of care, but could it one day be relevant to the device industry?

Consider the evidence
The amount of medical knowledge is doubling every five to seven years, according to various estimates. One quarter of all medical errors involve misdiagnosis or delayed diagnosis. IBM Watson has the ability to take in vast amounts of medical information and guide physicians in decision-making by analyzing the structure and wording of questions, and in a matter of seconds formulate an answer with the highest level of 'confidence' of being correct.

Most clinicians believe in practicing evidence-based medicine, which is a central theme of recent changes to US healthcare laws. But it’s very difficult for a caregiver to extrapolate the results of medical research to the special case of an individual. Watson could be trained to go through FDA records of adverse reactions to drugs to see if there were reasons a patient shouldn’t get a common prescription. The new technology could also review the medical histories that patients accumulate as they age.

A clinician doesn't need technology like IBM Watson to treat a patient with high blood pressure, but it might help if the patient is an 80-year-old diabetic with prostate problems, because he may not be able to take certain medications. The computer system could backstop a doctor with suggestions and warnings.

While Watson could never replace a doctor, it could serve as an invaluable tool for doctors to use. IBM Watson not only retains millions of pages of information, it has the ability to analyze a conversation with a patient and note critical pieces of information— such as drug interactions or family histories— and ask for additional information that might have been overlooked.

It can then combine this information with data from connected and implantable devices and medical literature to determine possible diagnoses—complete with documented “reasoning” and probabilities. Or it could request additional, seemingly unimportant information needed to test hypotheses.

WellPoint, one of the nation’s largest insurers, will begin to pilot the system with doctors and nurses early next year for utilization management and in oncology trials. By helping reduce erroneous diagnoses and recommending the most effective treatment options, the insurer expects to help keep its members healthier.

Working with Columbia University Medical Center and the University of Maryland School of Medicine, physicians and medical researchers are contributing their medical expertise and research to this collaborative effort. But as IBM Watson’s role in healthcare begins to emerge, it is important to examine if or how this same technology could enhance the commercial and clinical value of medical devices. This is something we are currently examining as IBM researchers and our partners look at ways this and similar technology could be used alongside data gathered by medical devices to provide patients a constant feedback mechanism to improve therapy.

A recent study by the IBM Institute for Business, “Value, Fade or Flourish: Rethinking the role of life sciences companies in the healthcare ecosystem,” indicates that in order to excel in coming years, a new breed of medical device manufacturer will begin to emerge. These companies will focus not only on the device, but on the patient as well, by packaging wellness and care support along with the device, and even selling management services with their devices directly to consumers.

Against the backdrop of these industry changes, a great deal of research is underway to determine the future applications for IBM Watson to enhance the value of medical devices. The ability to ingest and learn from data delivered by smart devices and sensors has the potential to dramatically accelerate the speed with which doctors can arrive at an accurate diagnosis or find effective treatments. Factor in Watson’s analytical speed and reach, and doctors could dramatically improve the quality of care provided to patients.

Going one step further, analyzing and sifting through massive amounts of structured and unstructured data as it streams in from multiple external and implantable devices could open the door to better predictive analytics. This kind of insight could be used to predict infections or potential device failure days in advance of current methods by sensing and responding to minute changes in the device or in the human body. For example, while oftentimes infections associated with an implant cannot be detected until the patient develops a fever, by analyzing and sensing data as it streams in, predictive analytics could potentially be used to spot problems before they occur.

IBM Watson can also analyze insurance claims, health benefits, and clinical data to suggest which devices could be most effective for a particular patient and at the same time covered by a patient’s insurance plan. As a result, IBM Watson could one day play an important role in recommending devices as it guides physicians in the practice of comparative effectiveness. By ingesting medical device literature, clinical trial studies and comparing it to patient data, it could identify the best option among many therapeutic options— be it a medical device, prescription drug or even surgery— based on individual patient needs and proven results from others with similar characteristics or co-morbidities.

Additionally, IBM Watson could be used to analyze patient-specific data in real-time and aggregate it with other health data to personalize treatment or educational information patients receive about their device. More personalized education could in turn help patients make better choices to support their own health. The technology could also be used to guide patients and caregivers by posing questions and answers when reading device data or for alerting physicians when device data reaches a certain threshold.

Potential design tool
For medical device design, training and clinical trials, IBM Watson could be used to generate the right hypothesis, identify the most appropriate population or target data and then weigh the relevance and credibility of data to guide manufacturers in innovation and decision making. For example, it could leverage vast amounts of information on previous design work, understanding what materials or designs worked, and which ones did not for consideration in new designs and materials selection.

Still in the early stages of development and trials, IBM Watson could be used to support the medical device community in many ways as researchers work to fine-tune the technology and continue to build upon its knowledge base. For this we are looking to the medical device community for expertise, guidance, and suggestions on how deep Q&A technology and the ability to provide deep differential diagnoses, improved treatment decisions, and the ability to analyze device data in real time could be harnessed to improve the value of medical devices.

As device manufacturers reach a crossroads and rethink their business strategy in an effort to entrench their position in the healthcare ecosystem, it might be time to look beyond traditional technology and consider how a new generation of computing such as IBM Watson might play a role in the device market.