IBM's Watson and Nuance's Dragon Medical

Earlier this week, after the Watson computer defeated human champions in a three-day match on the popular TV show Jeopardy, its creators and a Massachusetts software company prepared to sell a medical version of the smart machine.

IBM has an agreement with Nuance Communications Inc. of Burlington to sell Watson-based products to health care providers. Nuance already offers voice-recognition software for a variety of applications. The companies are pitching forthcoming technology that will, among other things, allow medical providers to describe symptoms orally and get diagnostic information in return.

Watson technology will enable Nuance to add artificial intelligence to its offerings. And that's where this blog comes in. I've been talking about these two topics, albeit separately, for some time now.

Watson's ability to analyze the meaning and context of human language, and quickly process information to find precise answers can assist decision makers, such as physicians and nurses, unlock important knowledge and facts buried within huge volumes of information, and offer answers they may not have considered to help validate their own ideas or hypotheses.

For example, a doctor considering a patient’s diagnosis could use Watson’s analytics technology, in conjunction with Nuance’s voice and clinical language understanding solutions, to rapidly consider all the related texts, reference materials, prior cases, and latest knowledge in journals and medical literature to gain evidence from many more potential sources than previously possible, thereby helping the medical professional to confidently determine the most likely diagnosis and treatment options.

For more information about the Watson computing system and the Jeopardy! challenge, click here.

On earlier posts to this blog, I discussed the semantic Web, ontologies, and speech recognition software from Nuance (and Adobe):

See, for example, my November 7, 2009 post "Vagueness, Logic and Ontology: Fuzzy Ontologies"

In traditional ontology theory, concepts and roles are crisp sets. However, there is a great deal of fuzziness in the real world.

For example, one may be interested in finding “a very strong flavored red wine” or in reasoning with concepts such as “a cold place”, “an expensive item”, “a fast motorcycle”, etc.

A possible solution to handling uncertain data is to incorporate fuzzy logic into ontologies. Unfortunately, these fuzzy ontologies have shortcomings – reasoners for fuzzy ontologies are not yet so polished as those for crisp (aka traditional) ontologies.

See, for example, my May 30, 2009 post on "Speech Recognition Software"

The trivial example of one-to-one translation given there can be extended to one-to-many translation: That is, in a matter of seconds, you could “program” Dragon Medical to type out a whole sentence in response to your speaking just a single word or code into the microphone. And, vice versa: you could “program” Dragon Medical to type out just a single word or code in response to your speaking a whole sentence into the microphone.

Strategies for Closed-loop Radiology Workflow with Speech-Driven Documentation and Communication

Commercially-available Clinical Radiology Support

Nationally, it is estimated that 20 to 30 percent of currently performed High Tech Diagnostic Imaging (HTDI) exams are medically unnecessary or inappropriate.

This is due, in part, because many HTDI exams are currently scheduled without the benefit of a standardized, evidence-based, decision support tool.

After a yearlong pilot program, in which more than 2,300 physicians from five Minnesota medical groups, five health plans, and the Minnesota Department of Human Services used e-Ordering (defined later) to order high-tech diagnostic imaging exams, it was found that the exams ordered with evidence-based decision-support technology had an increase in medical appropriateness versus orders initiated without it. The pilot also showed that using decision-support appropriateness criteria (defined later) in the physician’s office reduced patient exposure to unnecessary radiation, and contributed to a 0 percent increase in HTDI scans ordered in 2007 (following an 8 percent increase in Minnesota in 2006). N.B. During the time of the pilot, an estimated $28 million in healthcare cost savings was reported.

Going forward, Nuance’s RadPort, an electronic, evidence-based, decision-support (e-Ordering) solution is going to be used to support a statewide initiative to help ensure Minnesotans only receive medically appropriate high-tech (MRI, CT, PET and nuclear cardiology) diagnostic imaging (HTDI) tests.

Derived from the American College of Radiology (ACR) Appropriateness Criteria, and designed in conjunction with Massachusetts General Hospital, RadPort’s scoring methodology is continuously reviewed and updated by a panel of clinical and radiology experts.

RadPort combines the clinical information provided by the referring physician at the time of ordering with patient demographics to produce a “utility score” for the examination requested.

Click here for a discussion of appropriateness criteria and utility scores.

A high score indicates that the information provided strongly supports the use of the requested imaging test. A low score indicates that the request may not be appropriate. As a result, RadPort displays alternate procedure choices with corresponding utility scores of other relevant and appropriate exams.

In the event of a low score, RadPort provides clinical reference material designed to educate the referring physician so as to ensure appropriate ordering in the future.

Nuance Healthcare manages radiology as a single, end-to-end or closed-loop process. This approach gives radiologists and administrators a continuous view of the order from diagnosis through results process, which allows for analysis that can drive improvements at each stage.

The benefits of closed-loop radiology are evident: saving time, money, and resources, improving patient safety and care, and better clinical documentation.

Click here for an overview.

Click here for info on RadCube, a comprehensive, yet flexible, data warehouse for multi-dimensional business analysis and visualization tool. RadCube is another key application in the Nuance portfolio of imaging documentation and communication software.

RadCube merges OLAP and web architecture. This radiology-centric solution allows its users to perform highly complex clinical and detailed business analysis on your data in mere seconds using a simple drop-and-drag interface. By integrating advanced On-Line Analytical Processing (OLAP) techniques and Natural Language Processing (NLP) with one's unique institutional data, RadCube provides ways to visualize, analyze and optimize healthcare delivery to each patient. With RadCube, you can explore information from various business, clinical and research perspectives or identify trends, opportunities and potential issues quickly with familiar visualization techniques. You can analyze all operational aspects of your business including: Business Productivity, Quality Assurance, Utilization Management, Throughput Analysis, Communication Management, Evidence-Based Radiology, Advanced Data Mining Techniques, Trending and Forecasting, and Benchmarking.



{ click on image to see larger view }

Click here for info on applying “Six Sigma” principles to the radiology process.

Click here for an early article of mine on “Six Sigma”

A link to this article is also available near the top of "My [partial] bibliography" at the bottom of this blog.

Finally, click here for a brief introduction to i2b2, an open-source calable informatics framework that enables clinical researchers to use existing clinical data for discovery research and, when combined with IRB-approved genomic data, facilitate the design of targeted therapies for individual patients with diseases having genetic origins.

Epocrates Mobile Electronic Health Record

Health care delivery is now moving away from a specific location and into a virtual care space, where patient medical records and real-time clinical data will reside.

There has also been a revolution in access to clinical reference material, and physicians and other care providers are moving core scientific knowledge from books to apps. Today’s doctors are no longer what they can memorize.

Throughout this blog, I've been discussing many of the components of Electronic Health Record (EHR) systems that will likely be used in the larger systems installed throughout entire hospitals and regions. However, 50 percent of physicians in the US work in small or solo practices.

The Epocrates EHR is specifically targeting the solo practitioner and the small physician groups, which have different needs from the larger enterprise care facilities that so many EHR providers are pursuing. To that end, Epocrates says that it has "teamed up with a very well known” but still undisclosed practice manager partner, whose software will be integrated into the Epocrates EHR for calendaring, scheduling, appointments, billing and more.

The as-yet-not-released Epocrates EHR app will include the company's drug and safety content and will, they say, meet “meaningful use” and HIPAA compliance requirements.

There are offerings from other vendors such as Microsoft and Symantec too, but I've decided to look at Epocrates Mobile EHR because Epocrates was the first company to introduce a medical app for the iPhone. Furthermore, as seen in the image below, their app is now available for many of the popular hand-held smart devices.











Their fully functioning mobile EHR app will work both when connected and disconnected. It’s not enough to just have a mobile EHR, of course, and it is also a fully featured, web-enabled desktop application, too.

The mobile and desktop versions may have the same functionalities, but the different form factors will likely lead to certain applications being favored on each.

There is a functional parity between their mobile and web interface. The information will be delivered differently and the iPhone version will be used for some use cases more often than the desktop version and vice versa. These two platforms are complementary, so Epocrates is not necessarily "leading" with mobile.

Unlike some mobile “portals” to EHR systems, Epocrates offering is a native app with patient data stored in it.

One differentiator between the Epocrates mobile EHR app and some others is that theirs will be a native app and it will store patient data on the device. Users will interact with the EHR differently from a mobile device vs. a desktop client. The iPhone interface is not ideal for text entry, so they are looking for other ways to get information into the device. The iPhone interface will be used more for things like dictation, while the desktop interface won’t be used for that as much. From a task perspective, though, they are not looking to hobble the handheld interface in any way.

Click here to see new apps from Epocrates

Click here to see the Dragon speech-to-text application on a hand-held device

Epocrates is launching an EHR offering now because of the increase in EHR adoption because of the ARRA deadlines for stimulus fund incentives payments. These external deadlines are encouraging physicians to adopt EHR, making it an ideal time to enter the market.

There are also drivers from the technology perspective as well. It’s a good time to enter the market with a mobile EHR because more clinics have a high penetration of smartphone users. Also, 3G wireless networks now have higher bandwidth – enough to support these kinds of applications. Physicians rarely take note of [mobile] Internet connectivity issues anymore. Note: I've also talked a little about the less-ubiquitous 4G (WiMax) networks in earlier posts.

Epocrates plans to integrate the features of the iPhone, such as the camera, as well as dictation directly into the app, an interesting feature especially in light of the recent announcement by Nuance that they will be shipping a medical transcription application for the iPhone.

Epocrates will also utilize the iPhone’s Push Notification Service to alert doctors of important or timely information (and hopefully not overuse it). Given Epocrates’ depth of knowledge of pharmaceutical formularies, e-prescribing will be built in.

Although the mobile application will synchronize with the web application, it will continue working even when there is no internet connection. Data will be stored on the handheld device in a secure, encrypted manner and synchronize when a connection is available, a real issue for hospitals where there are many “dead” zones. And, there will be condition-based templates for easy entry of clinical information.

There will also be integration with a revenue-cycle platform so physicians can charge and submit codes, through a partnership with a “known” company. Epocrates anticipates they will achieve Certification Commission for Health Information Technology (CCHIT) certification by the time of release. The EHR will be delivered as a software-as-a-service (SAAS) model, meaning the physician will effectively lease, not purchase it.

This could be an attractive low-cost product for solo or small group physicians who do not have complex office staff EHR integration needs. In particular, for the physician who already uses their smart phone for many work activities, the potentially painless transition to using it as the primary interface into their office EHR may be very appealing.

Stay tuned.

Click here to see Epocrates executives at HIMSS 2010

Click here to see Andy Wiesenethal, Kaiser Permanente at HIMSS 2010

As an aside, I'd like to note that when Epocrates Online can be accessed with a browser on which Google's (or other) translation app has been installed, as shown in the image below.