EHRtv

This Video requires Flash Video Player

Dr. Eric Fishman with EHRtv speaking with Nick Van Terheyden.  Dr. Van Terheyden, the Chief Medical Information Officer for Nuance Healthcare

Category: Featured, HIMSS11
Date: March 2, 2011
Views:2,035 views
Information:

Dr. Eric Fishman: This is Dr. Eric Fishman with EHRTV. Today, we have the pleasure of speaking with Nick Van Terheyden. Dr. Van Terheyden is the Chief Medical Information Officer for Nuance Healthcare; obviously knows a lot about the speech recognition industry. So, Nick thank, you for taking the time to speak with me today.

Dr. Nick Van Terheyden: Thank you.

Dr. Eric Fishman: Nuance, obviously the largest company in the speech recognition industry, has made a number of announcements in recent past that are going to determine how speech recognition will be used in the future to some significant extent. They include yesterday’s announcement of cloud-based computing and obviously CLU, the clinical language understanding; your announcement with 3M recently about coding; and then obviously all of this combines into determining how physicians will interact with their computers going into the future. If we can spend a few minutes discussing each of those issues, that will be wonderful.

Dr. Nick Van Terheyden: That will be great. Love the opportunity and obviously it has been an exciting show for Nuance. I think we have had a number of very significant announcements, you know, the foundation of putting our technology and I think it expands much past just speech recognition but all of the adjunct technologies that support clinicians and in the delivery of good healthcare, high quality healthcare, reducing the overall resource usage to deliver the best possible care.

Dr. Eric Fishman: How long will it get would be until a physician using any of the dozens and dozens of electronic health records here in the hall will be actually using speech recognition on the cloud as opposed to a desktop version as this may be more common.

Dr. Nick Van Terheyden: In fact, the cloud-based computing or Nuance healthcare development platform was announced and in fact we have six partners who have already signed up and taken that integration solution to incorporate into their environment. I think what’s exciting about their experiences so far is of the order of the day to integrate speech into their applications. Part of the reason for that is that we have tried to make this as accessible as possible; one of the barriers to entry for speech in the cloud has been the difficulty of integration. We have created this solution that offers that integration that is as easy as possible, requires very little coding and certainly doesn’t require any of the weighty installations in whatever platform they happen to be working in. So, I think it’s going to be a very short order we are going to see those applications and we may even see some, you know, as a result some of the interactions we have had at …

Dr. Eric Fishman: Are there any delays inherent in the speech recognition? Obviously, both of us have been doing this for a long time and many years ago physicians would speak and the words would slowly march across the screen with a client server version or a desktop version; it’s almost immediate with good hardware these days. Is there any meaningful latency in the cloud at this point in time?

Dr. Nick Van Terheyden: I think that’s clearly the balance that people have to arrive at to say what’s the most appropriate for their use case and their experience and what they need. The advantage of that local installer is that you do have this instantaneous. We put processing power in people’s desktops that you don’t even have in the shorter ones, just taking off tomorrow. Well, yes we do because we would like to see it; and in the cloud, we can apply all of that but we still have the potential for latency and whether you have got tolerance for that latency will depend on your application and the demands that they have. Part of that will also be driven by the resilience of the networks that are put in place. Hence, large conferences, good case in point…

Dr. Eric Fishman: Yeah, a little bit more challenging.

Dr. Nick Van Terheyden: <> will push that.

Dr. Eric Fishman: I need you to talk about any departments that you have developed so far in this early stage of the cloud-based software?

Dr. Nick Van Terheyden: Yeah, we announced a couple as part of that initial press release – Calgary Scientific and E-Properties have taken the development platform – and then we are obviously working with some of the EHR vendors that at this point is in public knowledge, but you know, we are looking forward to that experience and I think they are too, they have been excited about that.

Dr. Eric Fishman: Now, let’s bring this up or down a level as you would like to and discuss it and talk about the clinical language understanding.

Dr. Nick Van Terheyden: Right.

Dr. Eric Fishman: How is that going to impact the physician in the exam room?

Dr. Nick Van Terheyden: Clinicians today capture the majority of the information in a narrative based form. I think we have seen that for the longest period of time.

Dr. Eric Fishman: Let me interrupt you and ask – clinicians, overall I would agree entire <> – would that also be your opinion with respect to clinicians who are using one of the more modern up-to-date certified EHRs and even in that environment they still…

Dr. Nick Van Terheyden: I think even with the structured data entry that we see with the EHRs that are looking for that process, whilst there are some that find that inappropriate way and a facile way, there are still many that don’t; and I think the challenge has been for many of the EHR companies to bridge that divide. You know, in any facility you have a spectrum of clinicians who have different methodologies, different ways of doing things. Well, clinical language understanding is how to bridge that divide; it provides another choice. For those that want to use data entry and use mouse or keyboard or, you know, even a tablet…

Dr. Eric Fishman: I guess I would interrupt for a moment. I recently did the survey of about a thousand pediatricians and multiple answers were permitted, but I was shocked that almost 70% say they type a portion of the data entry.

Dr. Nick Van Terheyden: And my follow up question to that would be “Why do they type?” Is that because they don’t have access to some form of speech technology; and if they did, would that still encourage them. You know, I wouldn’t discourage typing because I think in some instances of that clinical interaction, when you are talking to the patient, talking to a machine can be intrusive. So, typing is, you know, a multifunction type of activity. But for those 70%, if they had the opportunity, would they say “Gosh, I prefer to do that or it might be easier some of the time”, and I think the answer is always a spectrum of choices and making sure that they are available, but essential to that when you capture that narrative is the ability to extract out those discrete elements; what in many instances in these solutions is asked is that the physician entry select them from a dropdown list. Clinical language understanding takes that to a new level that says “We have the narrative, we can extract the facts, we can tag those facts against a controlled medical vocabulary and ontology that is referenced depending on the domain that we are in – Rx known for drugs, no meds, CT and indeed even ICD9 and ICD10. So, we are starting to prepare for the ICD10 transition because obviously that’s going to hit us and, you know, if we believe the deadline, a firm deadline would be October 2013.

Dr. Eric Fishman: Approximately 5 years ago, speaking to John Shegerian. I asked him point blank if he believed that clinical language understanding – may not have used that phase – would be available within 2 years and he demurred a little bit and he said “Well, if you would give me 5 years, I think I would say yes.” Do you believe there is something which is working but still needs a meaningful amount of research, development and just time to improve the accuracy sufficiently to be clinically viable in almost all instances?

Dr. Nick Van Terheyden: It depends on where you are in a document, which elements you are trying to extract. So, it’s more accurate and <> there is less accurate and all those; but I think much like the experiences speech recognition and our ability to get that out into the market place, one of the key transition points to the clinicians was instead of turning around and saying “You must use this. So, here is the frontline solution, use it live or die”; providing that choice, let’s say “You can use it, but if you don’t like the result that comes back, you have the option of sending it to somebody else for review.” Clinical language understanding has exactly the same model that says instead of compelling the physician to review the results, they can use a back-end process. Now, they lose some of the functionality because now you don’t have the opportunity for real-time feedback.

Dr. Eric Fishman: Does that back-end process usually or always utilize human beings <>?

Dr. Nick Van Terheyden: Yes, in exactly the same principle as we do in speech recognition tied into all of the same solution set that we have and for us that makes sense, you know, for some clinicians that are willing to; and what it boils down to is “Do I get value from this? If I get value, I am willing to invest the time. If I don’t receive the value, then I might not.”

Dr. Eric Fishman: Now, let’s bring this forward and talk about coding which is obviously a very expensive process in most physicians’ offices. How does this tie into that?

Dr. Nick Van Terheyden: Well, the recent announcement with 3M, pre-eminent experts in terms of the coding domain and what they have managed to achieve in the back-end process from clinical documentation improvement and all the tool set that they had as part of that. What we have done as part of that partnership and I think that’s what we are going to see over the course of time in the next several years with this show and others is the strategic alliance with companies that bring together experiences that don’t so much overlap but really contribute to a much, much better solution and that’s really the value of 3M and Nuance and it is 3M has this experience in that coding, documentation improvement and Nuance has that interaction with the clinician, the ability to capture clinical documentation in the most efficient way using all of the technologies. Bring those two to them and what you do is you apply 3M coding knowledge at the time of the clinician capturing that information. So, we turn that computer-assisted physician documentation. What they do is they provide feedback to the clinician that is offering him an opportunity to expand or provide more detail. Clinicians know what they owe and they don’t always document everything, and the biggest challenge that they face is remembering all of the elements that they are required to incorporate to satisfy all the downstream requirements for billing, re-investment and so forth. They get courses, they get education, but mostly they get pestered with continuing emails and requests from the documentation improving quality folks, sometimes days later, very challenging to remember. CAPD offers that process real time at the time of clinical documentation. So, when the physician says “heart failure”, we provide – and what I really like about this is the discrete nature of it, small question mark highlight that says, you know, there’s something here; you can choose to ignore it, choose to offload it to somebody as you do today; but if you would like to open that, it will offer you the choice that say “You said heart failure; did you mean systolic, diastolic, systolic and diastolic?” which intuitively as a clinician you have a view and you say “I know it because I actually documented it from a clinician reimbursement to an appropriate level of coding, we need that information and it needs to be tied back to the document. And as you think about ICD10, underlying experience is “we have got this massive expansion of codes, we need more information, and we are all ready.” So, that’s already preparing the physician. So, you get better quality documentation, you improve the overall coding process and remove some of the inefficiencies and require less staff at the back end, and I think the residual benefit what I call CBE or Continuing Billing Education for a physicians that says every time you do this, we are going to give you a little nod and say “Hey, did you mean this?” After two or three times, you are not going to do it anymore; and if the rules change, we will educate you but we are going to <> you again so that you learn as part of that process and that’s the best possible education that you can get and the least intrusive.

Dr. Eric Fishman: Nick, this already sounds like the future, so I feel a little awkward asking you – what does Nuance see in the future?

Dr. Nick Van Terheyden: Okay. So, you know, you set me up for a great opportunity to talk about probably one of the big buzzes at hands and certainly based on the experience of the IBM presentation that took place on Monday that was, you know, knee deep in the isles, Nuance announced its relationship with IBM to take the Watson technology with mutual investment over the course of the next five years to say “How come we bring that jeopardy experience, impressive as it was, into the medical domain?” Maybe I have a bias as a clinician; either way I looked at the jeopardy experience a number of months ago when they started promoting it and my immediate response was “How fantastic would that be for a clinician to have this knowledge base to help me process all of this massive clinical information that I have been challenged to keep up-to-date with so far and offer that in some form and in some toolset that allows me to make sure that I don’t forget anything? I am guided through some process with the patients so that I give the patients what they expect – best possible care, most up-to-date care at the time of the consultation” which is not always what happens through no faults of the folks that are they delivering.

Dr. Eric Fishman: I stopped practicing as did many physicians in this hallway, but it would be a fun time to be doing at least from an electronic standpoint in the future.

Dr. Nick Van Terheyden: Very exciting time and you know recognizing that is a little bit of ways out, but being part of that innovation and having seen the developments that, you know, massive quantum leap that IBM was brought to bear in terms of understanding means that we can truly bring medical knowledge into the domain of really understanding. So, it’s very exciting time but I am just lucky to be part of it.

Dr. Eric Fishman: Nick, thanks very much.

Dr. Nick Van Terheyden: Thanks very much.

Dr. Eric Fishman: This is Eric Fishman EHRTV. We have been discussing with Dr. Nick Van Terheyden, Chief Medical Information Officer of Nuance Healthcare, and the future of speech recognition, coding, clinical language understanding and other features in healthcare. Thank you.