Dissecting Digital Health — with Blackford Middleton

Blackford Middleton is an explorer of digital frontiers. He is a giant in the world of health informatics. His 30+ years as a physician and health informatics innovator has led him to conclude that healthcare is ripe for digital disruption.

Subscribe — Soundcloud, iTunes, GooglePlay

This is the full transcript of the podcast Dissecting Digital Health with Dr Louise Schaper, interview with Dr Blackford Middleton, Chief Informatics and Innovation Officer at Apervita.

Guest: Dr Blackford Middleton, Apervita
Host: Dr Louise Schaper, HISA
Tweet Louise
@louise_schaper Tweet Blackford @bfm
Production: This podcast is produced by Ivan Juric


Show Notes

[01:34] Opening Remarks by Dr Louise Schaper

[01:44] Today’s guest, Blackford Middleton, introduces himself as an explorer of “digital frontiers”. He talks about his parents’ influence and role in his life as inspirations for his career path.

[04:08] Blackford shares how he discovered informatics, and talks about his beginnings in epidemiology before medicine. Eventually he found an interest in the management of data which led and inspired him throughout his 25 years of work in clinical informatics.

[07:23] Louise asks Blackford to share how his interests in the field have evolved over the last 25 years. Blackford talks about prospects of differential diagnosis systems, big data insights, as well as precision medicine. The two share their ideas on how technological support will mean more time and energy left to focus on the more human aspects of clinical work.

Blackford addresses the question of whether the patient or clinician will be the main beneficent of technological advancement. Blackford argues that the patient will likely be the most empowered and goes on to give an example in Boston when decision support tools were given directly to patients. Louise and Blackford delve into the possibilities of improving care when patients have greater autonomy and direct access to their health information.

Blackford dives deeper into the topic and brings in the idea of what has been coined as a “cognitive funnel”. He discusses the area of intellect where a computer would be more effective than humans. Louise and Blackford discuss the error and breakdowns that occur in medicine during the handover from computer control to human control.

Louise wraps up the conversation by having Blackford impart his wisdom to potential newcomers in the field. Blackford talks about the great and growing area of opportunity for qualified personnel in informatics and shares his optimism regarding improvement in healthcare informatics.

[17:16] Closing Remarks


Full Transcript

Opening Remarks by Dr Louise Schaper

[01:34] Louise: Welcome to today’s episode of Dissecting Digital Health. I’m your host, Louise Schaper, and I’m sitting here with the wonderful…would you like to introduce yourself?

Today’s guest, Blackford Middleton, introduces himself as an explorer of “digital frontiers”. He talks about his parents’ influence and role in his life as inspirations for his career path.

[01:44] Blackford: Blackford Middleton. Good morning. It’s a pleasure to be with you, Louise. I’m very happy to be sharing this time with you and exploring digital frontiers.

[01:53] Louise: Oh, I like it — exploring digital frontiers. Have you always been an explorer Blackford?

[01:58] Blackford: Actually, pretty much. You know, it’s interesting. I’m the son of a physician and a nurse, my father was a bit of an explorer. He was an allergist/immunologist back in the day when allergy was sort of emerging from snake oil and pseudoscience to a much more scientifically-oriented discipline and immunologically based. I thought he’d already carved out that career and done very well, so I thought I would try to do the same in clinical informatics.

[02:25] Louise: So, your parents, and particularly your father, were they your inspiration?

[02:30] Blackford: In many ways, yes. My dad certainly was a terrific investigator and clinician and professor, mentor, teacher. He really was that triple threat which can be so much fun to pursue ideas and do so with groups of folks, both learners and colleagues who are motivated to try to push the envelope and make discoveries to make things better.

[02:57] Louise: So, conversations when you were growing up around the dinner table — were they intellectually stimulating? Did your parents challenge you, or were you just sitting there talking about reality TV?

[03:08] Blackford: Thank God it was pretty normal. [Laughter] A good mix I’d say, but interestingly, as I went into medical school and became more familiar, of course, with the field and kind of what my father had been thinking about, we were then able to have richer conversations about his pursuits and scientific endeavors and their clinical application, which was fun. Before, of course I just didn’t understand it all, but he had been very interested in sorting out the mast cell and its role in the allergic response and histamine release. He was one of the real pioneer investigators in that space, and there’s a book now that he began, which posthumously is known, after him, Middleton, Allergy Principles and Practice.

[03:56] Louise: Oh, fantastic! Okay, then. Alright, so everyone at home listening on their devices, wherever they might be, in the trams of Melbourne or walking around the lakes of Chicago, they should have a look at that book, then.

[04:08] Blackford: Indeed.

Louise, Blackford & ePatient Dave @ Irish pub in Chicago

Blackford shares how he discovered informatics, and talks about his beginnings in epidemiology before medicine. Eventually he found an interest in the management of data which led and inspired him throughout his 25 years of work in clinical informatics.

[04:08] Louise: So cool. Well, okay. So, how did you discover informatics?

[04:13] Blackford: So interestingly, you know, after university, I went to study epidemiology before medicine, and I became enamoured with just the management of data. Of course epidemiology training back then really was data science, and remains to this day, even though we have new names for a lot of this, as data scientists and the like, but

it was striking to me what one could do from an epidemiologic frontier in clinical epidemiology to look for patterns of disease, of course, and make associations and what not.

Then, I went to medical school and went to fellowship at Stanford, and began to become very interested in clinical informatics, partly because it would supply the data with which we could then do clinical epidemiology and association studies and the like. So, the first step was to get the data right, building EMRs and data management systems and of course 25 years later, we’re still working on that. [Laughter] It’s not quite a done deal.

[05:17] Louise: Okay, well, I’ve got my Tardis here right now, we can jump in and go back in time. I’m going to talk to Blackford Middleton 25 years ago. If I asked you then, where do you think the field of informatics will be in 25 years, what would you have said? Are you surprised by where we are now?

[05:35] Blackford: Yeah, great question. I guess 25 years ago, finishing fellowship, at Stanford in particular there was great hope for artificial intelligence. As a medical student…

[05:46] Louise: 25 years ago?

[05:48] Blackford: Oh yeah. And really, there was a hay day there, and then a valley of despair, in a way. Now, AI is back, but the Knowledge Systems Laboratory at Stanford and Ted Shortliffe and others doing this really great pioneering work on inference and expert systems and my own fellowship work was in that area. We built a differential diagnostic system that could create a differential diagnosis based upon an input of patient data, and I thought, as a junior physician, that this was the cat’s meow, because the knowledge-base seemed to be so difficult to master and of course every physician aims to do their best, but mistakes happen.

Oftentimes, we don’t have access to the relevant knowledge, not to mention access to the relevant data.

So, I thought that was the cat’s meow, and I thought that we would have expert systems assisting decision-making in clinical encounters within years, maybe a decade.

Instead, of course, we had to focus on the transactional infrastructure that is electronic health records, and your listeners and our countries have made extraordinary investments and endeavours to automate clinical practice with electronic health records, so now we’re kind of back at it again. Now that we have the install-base with electronic health records, what can we do now with the growing data resource availability and how can we apply analytics at the point of care in population health contexts to improve decision making and outcomes.

Louise asks Blackford to share how his interests in the field have evolved over the last 25 years. Blackford talks about prospects of differential diagnosis systems, big data insights, as well as precision medicine. The two share their ideas on how technological support will mean more time and energy left to focus on the more human aspects of clinical work.

[07:23] Louise: So, what you were interested in 25 years ago, how has it evolved?

[07:31] Blackford: Well, for sure now, in those early days, very simple decision support systems and drug selection or drug interaction or differential diagnosis systems. The latter actually never panned out well, but now with the masses of data and increasing availability of online knowledge resources, for example the IBM-Watson effort of course to digest and draw to be able to infer upon the world’s literature is extraordinary and big data insights as well. How can we learn from 10, 20, 30 million record databases about expected outcomes or do scenario planning, if you will, at the level of individual clinical decision making? Which drug is best for you today, based upon the world’s experience?

[08:20] Louise: And precision medicine, too.

[08:22] Blackford: And precision medicine, yeah.

[08:24] Louise: So, when would you say then…I’ve got my crystal ball here, I’ve got my Tardis, we’re all set. How long do you expect, given all of your experience and knowledge in the field, before we reach that point where our relationships in healthcare are actually, I think they’ll be richer as a result of the technology because it will make things easier for clinicians and for patients and we can get back to the business of real interactions with human beings, but we can use the technology as a tool to get the decisions that are needed. How long before you think we have that?

[09:03] Blackford: I think an optimistic projection might be in five to ten years, we’ll see radical transformation of what the EMR looks like and that

EMRs may become much more of a true cognitive assistance tool and alleviate a lot of the burden that is now placed on the clinician to touch and transact with the record and data entry and dealing with decision support alerts and fatigue and the like.

I think as computing improves, for the clinical encounter, hopefully be more of a computer in the background, a resource available to the clinician without necessarily having to have a direct connect through the keyboard, so to speak, with better speech translation or documentation, natural language processing, and the like. So, I think that’s coming.

Blackford addresses the question of whether the patient or clinician will be the main beneficent of technological advancement. Blackford argues that the patient will likely be the most empowered and goes on to give an example in Boston when decision support tools were given directly to patients. Louise and Blackford delve into the possibilities of improving care when patients have greater autonomy and direct access to their health information.

My first thought, though, when you asked the question, was is it going to be the provider who will be most empowered and most enabled by these technologies as we go forward, or the healthcare team even, or will it be the patient, the citizen, the client, if you will, that with these devices in our pockets, and more self-monitoring, and citizen-science as we heard about here in our keynote at AMIA, will I be generating data and managing many, many more decisions myself as a citizen for my own care and well-being? I think that frontier is as exciting, in some ways, to empower the patient and whatnot.

We did a couple of experiments in Boston back in the 2000s timeframe where, for example, we would give decision support directly to the patient. Interestingly, many feel that patient engagement is the holy grail and activating the patient is critically important, and we would agree with that.

75% of the time when a patient in a randomised controlled trial is given the opportunity to deliver a diabetes interval history to their primary care doc, they would do so.

So, that’s pretty good, a pretty good uptake. On the physician side, 75% of the time when that interval history came in, a journal from the patient if you will, as we called it, 75% of the time the docs would look at it. The most interesting finding in this trial was that those patients certainly became activated and engaged, but more importantly, the physicians became more activated.

When that relationship and the exchange and the exercise of co-management of diabetes became reality and feasible, the physicians were much more likely to manage the patient’s diabetic meds, for example, in a more appropriate way.

[11:51] Louise: Interesting. The Society For Participatory Medicine and Open Notes — they say that when you give patients access to this information by and large they do not just have a tool where they can be empowered, they actually do, and it makes it easier for them. I know that sometimes doctor will — because there’s a whole “Doctor Google”, but once you actually really do give patients information about them, not just a random search on the internet, that’s when I think it really changes that interaction of getting that information into the hands of patients. Do you agree?

[12:32] Blackford: I would totally agree. I think we see it already. It does raise some interesting questions. How do we make sure we’re managing uncertainty appropriately at a level? I know around my dinner table, occasionally at night, questions will arise and everybody will consult Google or your favourite search engine and the way we reason across information now is changing, and I think that’s an important thing to consider as we think about how we might use AI in health care. Do we take a search result as fact, or how do we qualify these findings online to make sure that we’re getting something that is relevant to me or my decision making or my context?

Blackford dives deeper into the topic and brings in the idea of what has been coined as a “cognitive funnel”. He discusses the area of intellect where a computer would be more effective than humans. Louise and Blackford discuss the error and breakdowns that occur in medicine during the handover from computer control to human control.

There’s a great old famous paper by Marsdon S. Blois in the New England Journal, approximately 1980, I think, where he describes a cognitive funnel. At the open end of this funnel is where humans reason most effectively because we can handle uncertainty better, we can handle the edge conditions, if you will, of reasoning, in any context. The narrow end of the funnel is where he suggested computational machines would reason better over calculations, for example, and really sort of narrow decision making context. I think an interesting question that will come today, comes to us now, is how do we manage then this interaction between human reasoner and cognitive aid, because the hand-off may be occurring many times as one proceeds through the cognitive funnel.

It’s noted, actually, in the airplane literature, transportation literature, that hand-offs from the autopilot to the pilot in the 747 is one of the most dangerous times.

[14:23] Louise: Well, that’s been shown in medicine as well.

[14:25] Blackford: Indeed.

[14:26] Louise: That where most of our communication breakdown happens is during that handover.

[14:29] Blackford: Indeed. Without even any inference going on, just hand-offs. Yeah.

[14:33] Louise: There’s just so much information to impart, and people do their best, right? I guess that’s what we’re passionate about, is improving the tools that clinicians have at their disposal to make their jobs easier. That’s really what it comes down to, I think.

Louise wraps up the conversation by having Blackford impart his wisdom to potential newcomers in the field. Blackford talks about the great and growing area of opportunity for qualified personnel in informatics and shares his optimism regarding improvement in healthcare informatics.

I know we don’t have a lot of time today and you’re going to have to run off to another very important meeting, so I just want to ask you a final question, and maybe next time we see each other we can have a longer conversation about your views and your history in informatics. My final question for you today is for people who are listening who are interested in getting into the field of informatics or maybe they’re quite new to healthcare, maybe they’re actually already in healthcare, but what are the next steps for them? What sort of advice do you have?

[15:24] Blackford: First, I think the demand for qualified personnel is just going to be booming in informatics broadly. It won’t be only at the professional level, if you will, and it won’t be only at the clinician side of the equation if you will. We believe strongly at AMIA of course that qualified professionals are important for the workforce to expand and augment the workforce, but it’s not just the docs. We have a chief of board certification for the physician clinicians who have board certification and a primary specialty to receive sub-specialty certification.

[15:58] Louise: Which is great. America’s the only country where that’s actually happening.

[16:02] Blackford: We’re also working on the professional certification, if you will, for everybody else, all of the non-MDs. So everybody else can receive a comparable level of certification in it and demonstrate their expertise, but actually at the Baccalaureate level and here at AMIA we have high school scholars who come…

[16:21] Louise: That’s so cool!

[16:22] Blackford: …and do their projects and make presentations, and the innovation is incredible. So,

I would encourage, I would relay to folks that the field is booming, that we’re still at the advent, even after these 30, 40, 50 years of work.

In 1957 there was the first artificial intelligence in medicine conference at Dartmouth, and even after this much time,

we’re still, I feel, at the beginning of this transformation where we can bring information and knowledge and information communication technologies to bare at the clinical encounter and improve the way we deliver medicine and lead our lives as individuals.

[17:02] Louise: Fantastic. The next 25 years, thinking we talked about 25 years ago, so I think the next 25 years are going to herald significant changes in the way we do medicine. Informatics is the place to be!

[17:13] Blackford: It is, indeed. [Laughter]

Closing Remarks

[17:16] Louise: [Laughter] Absolutely. You won’t get any argument on this podcast. Alright, thank you so much for your time, Blackford. I’ll let you get off to your meeting and we’ll chat again soon.

[17:22] Blackford: Thank you so much, Louise. I look forward to the next time.

[17:24] Louise: Thank you. Ciao!


Some references to check out

Clinical Judgment and Computers — Marsden S. Blois, Ph.D., M.D. N Engl J Med 1980; 303:192–197 July 24, 1980 DOI: 10.1056/NEJM198007243030405 http://www.nejm.org/doi/pdf/10.1056/NEJM198007243030405

Ten Commandments for Effective Clinical Decision Support: Making the Practice of Evidence-based Medicine a Reality. David W. Bates, MD, MSc, Gilad J. Kuperman, MD, PhD, Samuel Wang, MD, PhD, Tejal Gandhi, MD, PH, Anne Kittler, BA, Lynn Volk, MHS, Cynthia Spurr, RN, MBA, Ramin Khorasani, MD, Milenko Tanasijevic, MD, Blackford Middleton, MD, MSc, MPH. J Am Med Inform Assoc (2003) 10 (6): 523–530. DOI: https://doi.org/10.1197/jamia.M1370. Published: 01 November 2003


Contact Us

Suggest a guest via dissectingdigitalhealth‘AT’gmail.com
Want to learn more about digital health and health informatics — join HISA: Australia’s Digital Health Community www.hisa.org.au