HIMSS: Can Dr. Watson help manage 1.8 zetabytes of medical data?

LAS VEGAS—Physicians’ most significant challenges include processing information and bringing those data to the point of care, Nick van Terheyden, MD, CMIO, Nuance Communications, said during a Feb. 22 session at the annual meeting of the Healthcare Information and Management Systems Society (HIMSS). van Terheyden outlined how IBM Watson’s Deep Question and Answer (DeepQA) technology might assist with the process and improve healthcare delivery.

van Terheyden launched into the session by identifying a few key drivers in healthcare. He noted:

• Physicians are challenged by an overwhelming amount of information. The amount of information published in studies and journals has created a gap between physicians’ knowledge processing capacity and their knowledge processing requirements. The trend is expected to continue; 1.8 zetabytes of data, primarily unstructured, are created annually and information doubles every five years.

• Sixty to 80 percent of lifestyle diseases, such as diabetes, cancer and cardiac events, are preventable. Early identification and lifestyle modifications are key. van Terheyden noted that heart disease is first identifiable in toddlers.
• The lag between research evidence and clinical practice has reached 17 years.
• About 5 percent of autopsies reveal lethal diagnostic errors missed by physicians.

A final trend, van Terheyden said, is the move toward care teams rather than individual providers. A patient may have dozens or even hundreds of providers involved in diagnosis and treatment.

Enter Watson

Although it is highly unlikely that the human brain can keep up with the estimated 80 hours of weekly reading required to stay on top of medical research, Watson technology can sift through 200 million pages in three seconds.

Following Watson's Jeopardy debut, IBM is partnering with a handful of providers to determine how to link that processing capability with physicians at the point of care.

“Watson DeepQA provides a series of technologies to process unstructured data and interpret them to create a framework of knowledge at the point of care,” van Terheyden said.

The Watson development team continues to refine the technology to meet the needs of healthcare providers and expects to offer a service that analyzes multiple concurrent data, ranging from patient symptoms to clinical research.

One of the major challenges is how to incorporate Watson intelligently into clinical workflow. Watson needs to be able to pull information from the EMR without asking or disrupting the physician, and use that information to provide additional useful information, such as prompting questions to rule out unusual diagnoses.

Essentially, Watson may be refined into a series of guidelines and prompts. The long-term goal, said van Terheyden, is to create a state-of-the-art system that provides contextually relevant information and evidence-based medicine and presents it in a meaningful way to clinicians at the point of care.

It is an ambitious goal with multiple challenges, which include the ambiguity of human language and the complexities of interacting with existing silos of data.

While Watson may be needed stat, a more gradual process is likely, partially because the most significant challenge confronting Watson may be determining the optimal way to structure the interactions between technology and humans. That is, Watson should not interfere with the relation between the physician and patient.

For more about Watson, please read " Q&A: 'Doc' Watson Reports for Duty" in Health Imaging.