Dissecting Digital Health — with Christoph Lehmann
According to Google Scholar, Chris Lehmann’s academic citation tally is 4,655. But he is far from an ‘ivory tower’ academic, Chris is passionate about informatics making real differences in people’s lives - so much so, that he founded the journal of Applied Clinical Informatics.
This is the full transcript of the podcast Dissecting Digital Health with Dr Louise Schaper, interview with Professor Chris Lehmann, President Elect of the International Medical Informatics Association.
[01:15] Opening Remarks by host Dr Louise Schaper
[01:28] Today’s guest, Chris Lehmann introduces himself as president elect of the International Association and professor of biomedical informatics and paediatrics at Vanderbilt University. Louise tries to discover how Chris learned about informatics, and he shares his experience creating a patient simulation.
[03:29] Chris talks about his views on why automation through IT will make care safer and better. As an example, he tells Louise about a web-based TPN calculator that reduced errors by 90% and went viral in the hospital. This snowballed into many other changes implemented in the hospital for greater cost and time effectiveness.
[8:12] Louise asks for Chris’ thoughts on why the field of healthcare struggles with wider application of systems. Chris talks about the proof that health information technology’s benefits comes from home-grown systems, rather than commercial systems, which are behind the home-grown systems in many ways.
[11:45] Louise references House [TV show] as an analogy. Louise talks about the potential capabilities a computer would have in clinical decision support. Chris talks about current and effective diagnostic support systems and Louise asks Chris to go into more depth on specific examples.
[16:36] Louise asks Chris about his predictions on when the implementation of greater diagnostic support systems will take place. Chris thinks that the way we do treatment today will be seen as barbaric in 20 years’ time. Louise and Chris discuss the crucial need for everyone working in healthcare to have some training in informatics.
[18:48] Louise moves the conversation to IMIA (The International Medical Informatics Association). Chris explains IMIA’s goals to be a reliable resource for health informatics information, as well as provide cross-border networking opportunities for informaticians.
[21:52] Louise brings the discussion to a close by asking Chris to give his words of advice for those who want to be involved in the field of health informatics. Chris talked about the first steps for getting involved, from attending conferences in order to understand the current dialogue within the field, to physically participating in the conversation.
[23:56] Closing remarks
Opening Remarks by host Dr Louise Schaper
[01:15] Louise: Welcome to another episode of Dissecting Digital Health. I’m your host, Louise Schaper from HISA, and I’m sitting here with a beautiful view, looking out over the Sydney Harbour, and I’m sitting with…would you like to introduce yourself, guest?
Today’s guest, Chris Lehmann introduces himself as president elect of the International Association and professor of biomedical informatics and paediatrics at Vanderbilt University. Louise tries to discover how Chris learned about informatics, and he shares his experience creating a patient simulation.
[1:28] Chris: Yeah, sure. My name is Chris Lehmann. I am president elect of the International Medical Informatics Association and a professor of biomedical informatics and paediatrics at Vanderbilt University.
[1:40] Louise: Sounds impressive. [Laughter]
[01:41] Chris: Thank you.
[01:43] Louise: Well, we’ll talk about your current roles in a bit, but I’d be interested to learn a bit more about your journey. So, in high school and university, did you know what informatics was? Or did you fall into it like so many of us?
[01:56] Chris: Yeah, I think I definitely fall into the category of falling into it. During my paediatric residency, I was friends with one of our LAN administrators, and he had this idea that we should do a patient simulation. So, we created, when the internet was young, when the web had barely existed, all the backgrounds were still gray…
[02:20] Louise: Okay, so this is like mid-1990s?
[02:22] Chris: Yeah, mid-1990s, 1996, around that time, I created, together with him, a patient simulation that used a natural language interface. So, you could type in, “Where does it hurt? Does the pain radiate? Do you take any medications?” and the patient would understand the questions and answer them. So, I subsequently went to an AMIA meeting, and because I also played the patient, people in the hallways went, “Oh my God! I know you. You’re the interactive patient. I’ve seen you in your underwear!”
[02:56] Louise: [Laughter] So you were famous for your underwear? [Laughter]
[03:00] Chris: I was famous for my underwear, and I thought, “Man, this informatics thing is not so bad. Maybe you can actually get some infamy out of that.” Actually, I followed up on that with a paper that I wrote about 10 years later that had the title, “Managing Pornography-Seeking Behavior in an Online Dermatology Image Atlas,” which also got me quite [Laughter] some reputation.
[03:26] Louise: It’s a different strategy for citations, but it’s working for you…
Chris talks about his views on why automation through IT will make care safer and better. As an example, he tells Louise about a web-based TPN calculator that reduced errors by 90% and went viral in the hospital. This snowballed into many other changes implemented in the hospital for greater cost and time effectiveness.
[03:28] Chris: It worked quite well. As I told my mother back then, for two weeks when you searched PubMed for pornography my name came up first, so I had achieved everything I ever wanted to achieve.
I had a great time with the interactive patient, but then I realised rather quickly when I was standing in the NICU one day and I saw a resident overdose a patient by a factor of 10, I realised that I had made that same mistake before and I realised that education can achieve only so much.
Automation and standardisation that can be offered through IT is actually the way to go to make care safer and better.
And, because I really was traumatised by overdosing the patient, the patient had aneuria and I was sweating for days until kidney function returned, and I really felt a need to do something about this. So, I picked the most complicated problem I could find in my hubris, and decided I was going to fix the ordering of parental nutrition, which was notoriously difficult. It’s the holy grail in paediatric pharmacy. There’s so many ingredients that go in there, there’s so many dosing rules you need to know, items interact with each other, you have osmolality and things you need to consider.
So, I built a web-based TPN calculator that emulated the existing ordering form that we had, and I was able to reduce errors by 90%.
[05:00] Louise: 90%!
[05:01] Chris: 90%.
[05:02] Louise: How long did you work on that project for?
[05:04] Chris: [Laughter] It’s really embarrassing. The first version that we put out was done in four weeks. I did it with a nutritionist, and a pharmacist, and myself, and we had it done in four weeks.
[05:18] Louise: Wow.
[05:18] Chris: We introduced a new type of error in the process, so we had to revise it. None-the-less, the second version had a 90% error reduction.
[05:28] Louise: Oh my God. Fantastic. What has been the outcome from that early research that you did?
[05:33] Chris: So, I learned a couple of things from that. Number one, I only developed this thing for the NICU, and within a month it went viral in the hospital and they were using them in all paediatric floors.
I learned that what I had to do was I had to build something that was directly within the workflow of physicians, that didn’t change it, and that provided value for them, something that was in it for them.
So, in this case, the ordering went from 10 minutes, where they had to do all of the calculations by hand, to two minutes, and they loved that. I reduced their workload. I also realised that you try something simple, pragmatic, there’s a good chance of being successful. So, since then I have developed a lot of decision support, especially around medication management, but another five years later I realised providing safety and quality and reducing errors gets you only so many accolades. When your bosses really pay attention is when you save them money. So, we developed an antimicrobial stewardship program, and saved the children’s hospital $400,000 in the first year alone, and saved patients from getting unnecessary antibiotics. That was a real success. We’ve done things along these lines ever since, because if I can tie safety application together with cost savings, it is a win-win for everybody involved.
We just published a paper on therapeutic interchange where we trigger at the point of ordering to providers the PNT committee, the pharmacy and therapeutics committee, suggest that you use another agent. Would you like to switch to that? As a result of that, we saved for the 45,000 members of the Vanderbilt Health Plan almost $250,000 in a quarter. So, if you think about this on a larger scale, there are an enormous amount of savings that can be had, switching from IV medications to PO medications on patients that are on a diet in a hospital has enormous cost implications.
So, my journey was, oh, this can get you accolades, this can make care safer, but this can also make your bosses happy when you save money in the process.
Louise asks for Chris’ thoughts on why the field of healthcare struggles with wider application of systems. Chris talks about the proof that health information technology’s benefits comes from home-grown systems, rather than commercial systems, which are behind the home-grown systems in many ways.
[08:09] Louise: Yeah, that’s always good. Keep your boss happy. But see, some of the things you’ve just described, so, your first motivation around improving patient safety, and then seeing what can come out of that in terms of the benefit for patients, for clinicians, and for health care administrators as well, they’re sort of basic rules for informatics, and every informatician would be like, “Yep this is the way to make things work.” You’ve obviously had an innate ability to be able to achieve those results. Why do you think that we still struggle so much in more wider applications where we’re getting these things wrong?
[08:43] Chris: So this is an excellent question. We are entering a very interesting time. As was said in a paper not too long ago, we’re entering the dangerous decade. The implementation of EHR has taken off. We have more and more users dealing with commercial systems, and unfortunately,
the evidence that health information technology helps comes actually from the home-grown systems, where people are nimble, where there are short life cycles, where there is immediate modification based on user feedback.
These systems where we have the opportunity to tinker within our organisations, where we have control over HIT.
We are now entering a period where we have these commercial systems entering the market and taking a big chunk of the market space, and as a result of that, there are a lot of those advances that we have made are not necessarily implementing these systems, or they’re implemented in a fashion that they’re not usable or difficult to use. As a result of that, we may go into commercial systems, and
commercial systems usually means that we have stepped back in the quality and the amount of decision support.
We just did an analysis of the home-grown system at Vanderbilt we have, and looked at 3,000 knowledge artefacts for decision support. 3,000 in our system. When we go to a commercial system later this year, we will have a lot less decision support and it will take us years to make modifications to get to the same level of sophistication as we have been.
[10:38] Louise: So what can you say about why that decision was made? Looking at it just on that level, it seems like a step backwards, but obviously I’d assume the hospital is not going to make a backward step, so obviously there must be other drivers for them.
[10:51] Chris: Yeah, that’s an excellent question. So, unfortunately, a lot of the existing home-grown systems are built on technology that is no longer supported. There are functionalities in commercial systems that have big importance and advantages for hospitals, such as billing, such as being able to report meaningful use measures that are robust and working well, there’s the advantage of being able to put multiple institutions on the same platform and have the ability to exchange data that way. So, there are lots of advantages of these commercial systems. But the level of sophistication of decision support is still years behind a lot of those home-grown systems.
Louise references House [TV show] as an analogy. Louise talks about the potential capabilities a computer would have in clinical decision support. Chris talks about current and effective diagnostic support systems and makes the concerted point that CDSS must be a part of physician workflow in order to maximise their effectiveness.
[11:42] Louise: Yeah. On the clinical decision support, I know when I was just really a baby in health informatics and was reading about CDSS — there was a popular television show, an American show called House that was on television at that time, and the premise of the show…did you ever see House? Hugh Laurie is the doctor who takes way too many Vicodin pills.
[12:04] Chris: Oh, yes. I’ve seen maybe two episodes of it.
[12:09] Chris: His bedside manner leaves much to be desired.
[12:11] Louise: That is true. That is true. But, the premise of the TV show, because I used to use that as an analogy when I used to give public talks back in the day. Anyway, the premise was, because clinical decision support was still a very new concept in those days, the idea was that by 2016, which is the year we’re in now having this chat, that clinical decision support would be everywhere. So, in the television show, the idea is that somebody has a very, very complicated medical scenario going on and House is super clever, which is why people put up with his crappy bedside manner, and he and his team..they’re effectively diagnosticians, so they work out what is wrong with this person.
So, everyone knew it, it was a very popular show, so I used to talk about clinical decision support to non-clinicians saying, “Well, you watch that show…” and
the idea is that you have computer systems that would be able to prompt you what questions to ask of your patients and what tests to do, and it would also make sure, not only helping you in the diagnosis, but also helping in treatment as a really fantastic tool that clinicians could use.
But, we still find in 2016, and I never would have thought this back in the year 2000, that clinical decision support is still so rare in our systems at a global level. What is holding us back on that? That would be such an amazing tool.
[13:27] Chris: So this is a really good question. So, we ask people are they an above average driver, 90% will say they are above average drivers. [Laughter] Physicians are all above average diagnosticians, okay? So, there’s a good amount of [Laughter] hubris in their skills. There are a number of actually quite excellent diagnostic support systems. The problem with them is that even though there’s evidence that shows that they help physicians make the right diagnosis earlier, they usually do not fit in the workflow.
They’re not integrated in the workflow. It means you need to leave what you’re doing, go to another computer system, work with that. Now, we’re making progress on that. Things like the info button where you have a real-time decision support that’s delivered based on your needs, that is probably one of the first effective integrated ways of dealing with that.
[14:41] Louise: Could you talk a bit more about the info button for those who haven’t seen it?
[14:45] Chris: So, the concept of the info button, and I think it comes originally from Columbia University from Jim Cimino, is the notion that based on the patient and provider contexts, you have the opportunity to request additional information that would be helpful in the case that you currently are dealing with. So, that can be information about the medication the patient is on, information about differential diagnosis based on symptoms. So, it’s pretty much unlimited what additional information you can make available to a provider based on the situation that they’re in, or a patient for that matter.
[15:27] Louise: Right.
[15:28] Chris: So, I think having something that’s integrated in the workflow is critical. I think that we will see more of it, we will see more web services of decision support which is way more efficient. If you look at
immunisations for children, for example, it’s incredibly complex, it changes every year, so why should we change the decision support system in 6,000 hospital electronic health records every year if we can change it in one place and deliver it as a web service?
[16:02] Louise: Yeah.
[16:03] Chris: So, we will see more of that in the future. The fact that the marketplace is choosing fewer and fewer vendors will make that integration easier, so there will be more companies that will be able to provide layered over decision support on these commercial vendors, and as a result of that we’ll have better information for clinicians and patients in the future.
Louise asks Chris about his predictions on when the implementation of greater diagnostic support systems will take place. Chris thinks that the way we do treatment today will be seen as barbaric in 20 years’ time. Louise and Chris discuss the rising professionalism of the health informatics field and the crucial need for everyone working in healthcare to have some training in informatics.
[16:31] Louise: Yeah. That sounds good. Okay, done. So, when’s all of this going to happen, Chris?
[16:37] Chris: [Laughter] I think I’ll be busy for the rest of my career. We will think of the things we are doing right now as so primitive. I don’t know if you ever saw that Star Wars movie where they go back in time to San Francisco and Bones is in the hospital and this patient is going through dialysis, and he is saying, “How barbaric!” He gives him a pill and the guy is cured.
Well, a lot of the things we’re doing right now, we will consider them primitive and barbaric 20 years from now, and they’ll laugh at what we’re doing 20 years from now and 50 years from now.
[17:15] Louise: Yeah, okay. Well, I’m looking forward to that day as I’m sure very many of our health informatics and clinical colleagues are. What else do you see as…okay, so if we’re having another chat in 10 years, when are we going to see big changes for you and I to talk about them? Should we schedule this meeting in another five years, 10 years’ time?
[17:33] Chris: I think the big thing in our field will be that we will see more of a professional development in our field. We will see more certification going forward.
[17:47] Louise: For individuals as well?
[17:48] Chris: For individuals who are working in this field. People who started this, and they did it because they were geeks, they were interested in it, they had a vested interest in improving something, but a lot of people who started in informatics in the ’60s, ’70s, ’80s were not trained in clinical informatics. We are now seeing in the United States training programs popping up, we see certification in informatics for other disciplines, nurses, pharmacists. I think the quality of people who are working in this field can only improve as a result of that. I think that will be the biggest difference, and
we need to make sure that these efforts of improving the quality of the workforce are supported worldwide.
Louise moves the conversation to IMIA (The International Medical Informatics Association). Chris explains IMIA’s goals to be a reliable resource for health informatics information, as well as provide cross-border networking opportunities for informaticians.
[18:43] Louise: I couldn’t agree with you more on that. So on the worldwide agenda as well, can I ask you a little bit more about IMIA? The International Medical Informatics Association, you’re the incoming, or president elect. You take office in August next year, is that right?
[18:55] Chris: That’s correct.
[18:57] Louise: So, for the listeners at home who haven’t come across IMIA before, tell us a bit about IMIA and its mission and why we should care.
[19:06] Chris: IMIA is an association of associations, so HISA, AMIA, COACH, large medical informatics associations become members of the International Medical Informatics Association who has several roles. One of them is a convener, to bring people together for exchange of information for education and outreach. It’s also an organisation that can provide expertise on an international level, so it can be a resource to the United Nationals or the World Health Organization. It’s further an organisation that allows the sharing of informatics principles and lessons learned across borders. It provides networking opportunities for informaticians worldwide. But, most importantly, IMIA is a place where we try to generate best evidence, where
we can provide guidelines and practical resources that can be used by informaticians and implementers worldwide.
We collate and publish in places like the Yearbook of Medical Informatics that the organisation publishes online.
[20:26] Louise: Which is a fantastic resource, and I think actually in the show notes of the podcast I’ll put a link to the IMIA Yearbook because it’s been open source now for a couple of years, hasn’t it?
[20:36] Chris: Yes. We were fortunate enough to convince the IMIA board that this resource should be open access so it can be used by informaticians everywhere in the world, including resource-poor countries. We think the value of it is too high to hide it behind a pay wall.
[20:54] Louise: Excellent. And you’ve got exciting things planned I guess for IMIA for the next couple of years. We’re going to see each other again I guess in China?
[21:03] Chris: We will see each other in Xiamen, China, beautiful resort town (note, since this recording was made, MedInfo 2017 venue has been changed to Hangzhou, China)
[21:08] Louise: Oh, is it?
[21:11] Chris: Yeah, I was told that it has a beautiful beach and their facilities are supposed to be wonderful, so I’m looking forward to a bit of R and R after the conference, as well.
[21:21] Louise: RnR, okay, because last meeting of course was in San Paolo and I took the opportunity to go to Rio as well and had a bit of RnR. I hadn’t thought of doing the same thing in China. That’s what we should do then.
[21:33] Chris: Yeah. I think a lot of people will use the opportunity to come through Hong Kong, which is an hour and a half away by flight from Xiamen, so that would be a lovely stop in between, have your suits tailored and head off to the meeting.
Louise brings the discussion to a close by asking Chris to give his words of advice for those who want to be involved in the field of health informatics. Chris talked about the first steps for getting involved, from attending conferences in order to understand the current dialogue within the field, to physically participating in the conversation.
[21:46] Louise: [Laughter] Alright. There you go, so we’ve had all sorts of fashion advice as well today on the podcast.
So final question for you Chris, and thanks again for taking the time to chat to me today. So, for people who maybe they are clinicians, maybe they are researchers, people who actually want to get involved in the field of health informatics and aren’t yet, or they’re just sort of around the edges, what would be your advice as the best thing to do? What you’ve said today in our chat is you really do see the future workforces being integral to the future of healthcare and I could agree more. So, if people are wanting to get involved, what should they be doing?
[22:22] Chris: So, first and foremost, you have to feel into the field and see if it’s really what you are looking for. So, I think IMIA member organisations put together great conferences across the world.
The first place to go would be to go to one of these conferences, listen to what is being presented, try to get an idea for where the field is heading and what people are doing, try to meet some folks.
But then, if you decide this is the right thing for you to do, then the question is how do you gain the expertise? There are lots of opportunities, you know, there’s always the train yourself and learn by doing it methodology, but there are great training programs worldwide that give the opportunity to learn from informaticians in academia all over the world. There’s the opportunity to become involved on a local level in your hospital and committees that work on clinical decision support. So,
I think the most important thing for you to do is join others and raise your hand and volunteer to participate.
This is the way to become part of the field, to gain expertise, to learn, and then ultimately also a way for you to share your knowledge that you’ve collated over the years.
[23:47] Louise: Fantastic. Alright then. Thank you very much for your advice, I appreciate that Chris, and thanks again for having a chat with me and I’ll see you again…well, later today, but I’ll definitely see you again in Xiamen next year.
[23:58] Chris: I’m looking forward to seeing you in Xiamen.
[24:00] Louise: Thanks. We’ll speak to you then. Bye.
[24:02] Chris: Bye.
MedInfo 2017 http://medinfo2017.medmeeting.org/en
The Dangerous Decade Coiera E, Aarts J, Kulikowski C. J Am Med Inform Assoc. 2012 Jan-Feb;19(1):2–5. doi: 10.1136/amiajnl-2011–000674. Epub 2011 Nov 24. https://www.ncbi.nlm.nih.gov/pubmed/22116642
ePrescribing: Reducing Costs Through In-Class Therapeutic Interchange SP Stenner et al. Appl Clin Inform 7 (4), 1168–1181. 2016 Dec 14. https://www.ncbi.nlm.nih.gov/labs/articles/27966005/
IMIA Yearbook of Medical Informatics http://imia-medinfo.org/wp/imia-yearbook-of-medical-informatics/