John Halamka, CIO, Harvard Medical School
John D.
Watch on YouTube
49:21 
John D. Halamka, MD, MS is Chief Information Officer of the Beth Israel Deaconess Medical Center, Chief Information Officer and Dean for Technology at Harvard Medical School, Chairman of the New England Health Electronic Data Interchange Network (NEHEN), CEO of MA-SHARE (the Regional Health Information Organization), Chair of the US Healthcare Information Technology Standards Panel (HITSP), and a practicing Emergency Physician.
Dr. Halamka completed his undergraduate studies at Stanford University where he received a degree in Medical Microbiology and a degree in Public Policy with a focus on technology issues. While at Stanford he served as research assistant to Dr. Edward Teller, Dr. Milton Friedman, and presidential candidate John B. Anderson. He authored three books on technology related issues and formed a software development firm, Ibis Research Labs, Inc. Additionally, he served as a columnist for Infoworld, technical editor of Computer Language Magazine and technology consultant to several startup companies.
In 1984, Dr. Halamka entered medical school at the University of California San Francisco and simultaneously pursued graduate work in Bioengineering at the University of California, Berkeley, focusing on technology issues in medicine. During medical school and graduate training, he continued his business activities and developed Ibis Research Labs into a 25 person software consultancy, specializing in medical and financial information interchange. Ibis was sold to senior management in 1992.
Dr. Halamka served his residency at Harbor-UCLA Medical Center in the Department of Emergency Medicine. While at Harbor-UCLA he was an active member of the information systems team and developed a hospital-wide knowledge base for policies, procedures, and protocols. Further, he was instrumental in creating an online medical record, a quality control system, and several systems for medical education. His research focus during residency was building automated triage tools for patients infected with HIV.
View John Halamka's Links:
- Blog (Geekdoctor.blogspot.com)
- Explanation of his personal health records
- Continuity of Care Document
- Non-interoperable Electronic Health Record
In 1996, Dr. Halamka joined the faculty of Harvard Medical School and continues to integrate his knowledge of medicine and technology focusing on the use of the Internet to exchange clinical patient data. His research includes security / confidentiality issues, scalability issues, and implementation of standards for exchange of administrative and clinical information. As a clinician as well as researcher, Dr. Halamka uses these tools to improve the care of the patients he treats in the Beth Israel Deaconess Emergency Department. He is also an active teacher, lecturing on both medical and technology topics to the students, residents, and faculty of Harvard and MIT.
As Chief Information Officer at Beth Israel Deaconess, he is responsible for all clinical, financial, administrative and academic information technology serving 3,000 doctors, 12,000 employees and one million patients. As Chief Information Officer and Dean for Technology at Harvard Medical School, he oversees all educational, research and administrative computing for 18,000 faculty and 3,000 students.
As Chairman of NEHEN he oversees the administrative data exchange among the payors and providers in Massachusetts. As Chief Exchange Officer of MA-SHARE he oversees the Regional Healthcare Information Organization (RHIO), which develops clinical data exchange efforts in Massachusetts. As Chair of HITSP he coordinates the process of electronic standards harmonization among stakeholders nationwide.
Michael:
(00:03)Healthcare, that’s what we’re talking about today on episode number 82 of CXOTalk. I am Michael Krigsman as always joined by – you know in truth, I lie in bed at night thinking about the superlatives that I can say about my glorious co-host Vala Afshar. Mr. Afshar how are you?
Vala:
(00:25) Thank you very much. What a great introduction.
Michael:
(00:28) Nothing for the best but the best.
Vala:
(00:32) Speaking of the best, Michael we truly have an extraordinary guest today. We do, please.
Michael:
(00:37) We are joined today by Dr. John Halamka and I hope I’m pronouncing your name correctly John, who is the CIO of – it’ a long list. John, why don’t you introduce yourself briefly and tell us the list of where you’re CIO and what do you do.
Dr. Halamka:
(00:57) Right, so these days I oversee the technology infrastructure at Beth Israel Deaconessand it’shospitals. There are five hospitals in three locations, urgent care ambulatory and a lot of Eastern Massachusetts and expanding footprint for healthcare and we’re also an accountable care organization. Three petabytes of data, 22,000 simultaneous users, three data centers, clouds, lots of mobile devices.
(01:24) I also oversee a variety of planning activities for the health and information exchange in the commonwealth for Massachusetts. I’ve done a variety of things across the commonwealth in the last 20 years and in the Obama administration to help oversee some of the standard creation for health care. I am an emergency physician, a professor at Harvard and a farmer.
Michael:
(01:45) And you are also CIO at Harvard Medical School.
Dr. Halamka:
(01:48) Well I’m a professor at Harvard Medical School these days. I recently handed the office of the CIO at Harvard Medical School to a research focused CIO who had built out super computing infrastructure. So you can only do so many things in a 168 hour week.
Vala:
(02:05) Dr. Halamka, you were named the 50 most influential technologist of the past 50 years, and Beth Israel was named the number one healthcare. So how does a doctor and a practicing emergency physician become a CIO?
Dr. Halamka:
(02:25) Okay, well this is a little complicated, but when I was 12 years old – see, we’re going back to the beginning. My parents we’re in Law School and I was a latchkey child in Southern California at the height of the Aerospace industry. Before there was Silicon Valley, before there was dot com, there was Boeing, Hew, PRW. Those guys got integrated circuits, early microprocessors. They tested them for military spec and if they didn’t pass, they sold them to surplus stores for a dime a pound.
(02:59)So I was riding my bike as a 12 year old around surplus stores picking up early chips that were $500 a piece for a quarter. And so shot my analog and digital logic, early microprocessors and with popular electronics and published that Altair 880 article, I was there building the first Altair.
I was the first undergraduate at Stanford University to have a computer because I built it. and so hence this dual technology and medicine path has followed since from 12 onto 52.
Michael:
(03:37) And you actually are a practicing emergency room physician?
Dr. Halamka:
(03:40) I do 600 consults a year these days, oddly I’m an expert on poisonous mushrooms and plants and so multiple times a day I will get a telemedicine consultation, appearing on my iPhone will be a picture of something a two year old has just eaten and the question, life or death, what do we do?
Michael:
(04:05) I don’t know how to respond to that.
Dr. Halamka:
(04:07) the answer is don’t buy an Audubon field guide and go hunting for mushrooms.
Michael:
(04:13) Okay, so later on let’s talk about mushrooms and you also have a practicing farm, so let’s come back to that. But tell us about your role as the CIO, tell us what does that actually mean, what do you do.
Dr. Halamka:
(04:34) It’s fascinating because most people laugh in the tech industry and are given a job for what- 18 month, three years and I have been the CIO at Beth Israel Deaconess for almost 20 and you would say, well why? Well it’s because the nature of my job has radically changed pretty much every two years.
(04:53) That is in 1996 a lot were writing code. I was creating personal health records and health information exchange and hey this was over the internet. In 1996, the internet are you nuts!
(05:09) and then over the last 20 years have gone from writing architecture to doing strategy, to doing you know, how do we do a merger acquisition cloud security which is now one of the areas I have a primary focus.
(05:24) My day-to-day is more about figuring out how you can run Agile enterprise with so many single locations and merger activity while keeping the data secure and private. It’s very very different in writing code and do the basics of what I had to do in 1996 making the plumbing work.
Vala:
(05:48) Surviving two decades of being a CIO at one of the most prestigious healthcare organizations in the world must mean that, you know you don’t make any mistakes, error free IT organization?
Dr. Halamka:
(06:04) Oh yeah. So of course there are two ways you can get through a crisis. You can hide it, in which case when they find out you end up on page one. Or, you can invite the press and a party to watch the decision making in the stress and making them understand at deep level, the issues. Then you end up either on page 37 or you end up as a Pulitzer Prize winning article that might get published in a major trade journal.
(06:33)So I’ve always chosen the latter, which is, oh, my network collapsed, why? What was the impact on the patients? How did I keep my job? I’ve invited the press in. I brought in the Boston Globe, I’ve brought in ICO Magazine, they watched us through the problem resolution.
(06:54) When I’ve had issues of security, we disclose immediately, the issue to federal and state regulatory agencies, and then I go on the road saying to CIO’s, you haven’t been hacked, you just haven’t looked.
(07:09) You know we’re very very honest about the mistakes you’ve made and somehow the industry forgives me because of my honesty.
Michael:
(07:18) That’s a very interesting point so is it literally the case that the industry forgives you for because of your honesty or are there other kind of related dynamics that are going on with that?
Dr. Halamka:
(07:34) Well there is a certain Sigmund Freud, oh my, he was the one that got tagged with the privacy issue! Phew, dodged that bullet. So people feel a camaraderie when you are out there on the frontline and suffering they know what they could have suffered an try to make it an industry issue rather than a personal issue.
(07:55) But I think there’s a certain humility to your point. Mistakes are made every day. Mistakes are learning experiences and I don’t actually say, who is at fault I ask, what was wrong about our process, our strategy, or structure that enabled that mistake. I never shoot the messenger. So maybe that attitude makes people understand we’re all imperfect and if we can learn from each other, that probable not a fire able offence.
Vala:
(08:24) So with this you know, hyper connected world and everyone’s mobile, and you know, a physician in your hospital loses his or her phone, whose fault is it.
Dr. Halamka:
(08:36) So it sounds like you’ve read my job description!
Vala:
(08:41) All point to a doctor Halamka.
Dr. Halamka:
(08:44)My job description reads the following. Top technology position, accountable for everything. CISO makes policy and technology choices and if the CISO fails, the CIO is accountable.
(09:02)So yes, every lost laptop, every socially engineered lost password, every decision made by a vendor that results in a breach rolls up to me.
Michael:
(09:16) So we have a question from Twitter, and we tend to give priority to the audience on Twitter because they are our constituency. And Mike Shappelle says he uses the care group case study in his course each semester, would you provide some prospective 10 years on. I guess 10 since it was written.
Dr. Halamka:
(09:44) Absolutely, the case study reflects on this collapse we had in 2002. And you guys being Boston based you may know that as mergers occurred in the 90’s and say you take two money losing institutions and combined them into one, substantial profits don’t result
And so we had a problem with a variety of merger and acquisition activities. Margins were low and capital was tight and the capacity to invest an infrastructure was that my capital budget for 2001 was zero. It was literally zero.
And what you have imagine this hyper growth of bandwidth consumption, the appearance of evil actors on the internet and a capital budget of nothing and that combination was disastrous and ended up with everything in propagation and errors on the network which resulted in total denial of service for about a day and a half across the hospital and required some network re-architecture.
Michael:
(10:51) I remember that by the way. I live in Boston and I remember people walking down the street talking about that actually.
John
(10:57) Yeah, so what did I learned. Well first the manual was right, let no urgency go unused. And so if my capital budget was zero and then a disaster occurred, suddenly $5 million appeared and I was able to create the infrastructure necessary to keep patient records safe.
(11:15) So that was fascinating, you know there was a sense of urgency and (unclear 11:20) that I could have lost my job but the outcome was actually quite positive, it was stronger networked and better governed.
(11:28) Infrastructure is something often we forget. How exciting is the wires in the walls. Who is the advocate for the wiring in the walls, no, no, no I want this sexy new MRI machine. I want the Google Glass, the iPhone, the Apple watch. It’s only as good as the infrastructure in the walls.
So my job every year is to advocate to that hidden infrastructure, the storage, the servers, the data centers, the power, the networks that are behind the scenes and so that’s been my role. And my biggest observation is the blind spots that I had for over the years. Did I in 2002 know how to engineer from scratch a layer three network that was resilient to all the evil actors on the internet? No. you see the problem was I didn’t even know what questions to ask. And so what I’ve been very careful about is to surround myself with people smarter than me and to ensure the blind spots that I have are filled by others.
Vala:
(12:31) Are these others CIO’s in your network when you say surround yourself with folks that are equally smart. Are these other CIO’s?
Dr. Halamka:
(12:43) The answer is it depends. So we have Massachusetts something called the CIO forum, where all the CIO’s gathered on a monthly basis and share their experiences. So we have people who have gone as early adopters to certain technologies and said, oh don’t go there or go there. So I learned from my peers.
(13:02) Also it’s really important that I am not the guy who is the smartest engineer in the room and why. You know, you want to employ those people who are going to have that deep knowledge of, well what storage platform should we be on.
(13:18) You know, we’ve gone from the metrics to VNX to ep modes to Isolon, we are a vendor neutral kind of organisation going to whatever solution is most appropriate for the time, and I delegate that to very smart engineers and made sure that we have the right engineers to make such decisions.
Vala:
(13:37) In terms of learning from miss-staffs, do you think there is just unrealistic deadlines, are we moving too fast. Let’s take the healthcare.gov as an example. I mean I know that every miss-staff is an opportunity for us to learn and improve, but is this a technology thing or a scheduling unreasonable training and other constraints that lead to deal mistakes and failures.
Dr. Halamka:
(14:05) Sure, so I tell my staff – well I tell them many things, like there is no problem that can be blamed on IT, you know of course that’s important to know. Like Ebola we pretty much cause that.
(14:19) But I also tell them, you should never go live based on a deadline, you go live when the product is ready or the people are ready to use the product. And yes, it summarizes as follows. If you go live to early no one will ever forget. If you go live late no one will ever remember, and if healthcare.gov was delayed six months who would have cared.
Vala:
(14:50) Nobody
Michael:
(14:54) Yeah, but they of course they succumb to that temporary political pressures and I think that this issue – I have actually written a lot and I’ve studied a lot about failed IT related projects. And the interesting thing about it is most of the time it has very little to do with the actual technology itself, and has everything to do with the context around it. Including the decisions, the management decisions and the political pressures that management feels that drives them to make certain shortsighted decisions like go live based on a date, even if everybody or some people know that the system is simply not ready.
Dr. Halamka:
(15:39) Right, and so what of course you have learnt from healthcare.gov is what was the political fallout for going live today that was politically appropriate – pretty bad, almost devastating.
(15:53) And so having done hundreds of go lives in my life, I know there is no naysayer, there is no pressure, there is no political issues important enough to go live to early. So that’s fine, take this morning I just referred another go live. I said not November 3, December 8. The people, the stuff around the software will be ready then.
Michael:
(16:18) But how does a CIO get himself or herself into the position where they have the confidence to and influence together to make those decisions and run the risk of pissing off all the various people going to blame them for all the things that could happen.
Dr. Halamka:
(16:43) The simple answer is remember I’m a Harvard professor in the world of Harvard faculty. What’s the value of formal authority, oh that would be nothing right. I have no authority of any kind. I have informal authority, I have the capacity and the influence. So when people say, you know actually we have watched the last hundred go lives and they have actually gone okay. So if he says we better delay on this one, well you know past experience is a good predictor and we probably should.
Vala:
(17:15) Operational excellence over a sustained period.
Michael:
(17:17) So Kim Stevenson, who is the CIO of Intel is a very good friend of CXOTalk and she has a pyramid that she uses to work a model of IT. And at the bottom she talks about operational excellence, which basically means if you want the organisation to trust you which is one of the things I think you are talking about, then you must develop a track record of delivering what you say you are going to deliver – actually following through at the base level.
Dr. Halamka:
(17:54) Right, and so he is one of the things that I learned back in that 2000 care group network outage case. You are feeling so nervous and you know that in that core router configuration everything is going to light up and it’s all going to be fabulous. And so you tell all of the people in one hour it will all be back, and then seven hours later you realise oh my god, the problem is a whole lot worse than I first thought and you just shot your credibility.
(18:26) So I learned time and time again of what you need to do is under promise and over deliver.
Vala:
(18:33) That’s great advice. What are some of the primary mandates that you and your organisation must respond to.
Dr. Halamka:
(18:39) Okay, so the Obama administration has a $20 billion medical record plan and this is the Affordable Care Act and the America Recovery and Reinvestment Act, which forces us to follow thousands of pages of regulations of processes and support of the federal program. And if you do it right you get a stimulus, if you fail to do it you get a penalty. So that has actually driven a whole lot of my industry for the last four years or so.
(19:14) And then we have some very exciting projects like ICD10, now I do want to point out, so this is not a chicken injury. It’s a guinea fowl injury and you as a doctor need to code that species, because otherwise it’s fraud and abuse.
(19:33) ICD10 is 170,000 code including to for bird related injuries, and don’t forget eaten by orca, burned while jet skiing and fell through jet engine.
Michael:
(19:49) There is also one, fell through jet engine again.
Dr. Halamka:
(19:52) That’s a twin encounter exactly.
Michael:
(19:55) That’s a twin encounter, falling through a jet engine.
Dr. Halamka:
(19:58) You never know and so we have had to implement that federal mandate, and then (the omnibus rule?) $1.5 million fines per incidents for privacy breaches, and that could be a lost laptop.
Vala:
(20:15) Sorry 1.5 million for one incident.
Dr. Halamka:
(20:18) Correct
Vala:
(20:22) So why is healthcare data more costly or more valuable than any other data can you explain why the 1.5 million.
Dr. Halamka:
(20:52) Okay,so let’s say I want Michael’s Social Security number, $.25. Suppose we want his Visa card one dollar. Suppose I want his target Visa card $1.50, because it turns out the target breach includes the store that you bought things at and therefore I can fly below the radar screen of all of the vulnerability checking if I bought something at the store that he was buying at.
(21:00) Now I want to buy his entire medical record because you know I don’t have health insurance he does. In fact I hear he has that great Blue Cross plan, one of those Cadillac plans. I had a heart transplant and all I have to do is pretend to be Michael, and suddenly I get a heart transplant for free. What would I pay for a free heart transplant? On the black market about $100-$200 to buy Michael’s medical record including his insurance information.
(21:30) So this is a Willie Sutton issue, you want $.25 for those security numbers or $200 million medical records – go where the money is.
Vala:
(21:40) That’s incredible.
Michael:
(21:43) So what’s the role of IT and your role as CIO in all of this.
Dr. Halamka:
(21:51) Right so here’s my challenge, the affordable care act tells us we should now look at wellness rather than thickness. We should keep you out of the hospital and out of the emergency department. So you know, I think you are wearing a Fitbit right now, so you’re Fitbit should send your activity data over to healthcare, where then you should trust my app so I can upload your activity levels and look to see if you are having a decreased activities of daily life, so I can make an intervention. It could be depression, it could be weight loss, it could be diabetes – who knows.
(22:28) So I could get more data with your consent and give it to more people while at the same time, the privacy of every byte of data of everywhere it goes.
Vala:
(22:40) Is this part of the meaningful new stage two process.
Dr. Halamka:
(22:43) Right so the meaningful new stage two necessitates the sharing of data across multiple caregivers so that we coordinate care, population and health care management and it requires the security analytics so that I am keeping that data safe. And it requires that I share the data with the patient and their families.
Michael:
(23:04) So is this portable data – I have a friend that is working on a project for New England State regarding data portability between various care providers so that the patient can move and the data will flow with them.
Vala:
(23:22) I saw in your one of your recent blogs you talked about the past year using meaningful use at Beth Israel, so I want to just quickly, so unique patients, 740,000 unique documents or 8.3 million problems, 9 million procedures 8 million results 37 million, medication 6.8 million – I mean this demonstrates an incredible amount of inter-offer ability and visibility for your organisation to achieve all of this, and is this common throughout healthcare organizations or is this unique to Massachusetts and what you have been able to achieve.
Dr. Halamka:
(24:02) So Massachusetts does leave the way in inter-offer ability. It’s art, we are a non-profit medical state with nonprofit insurers and we all love each other and they all work together because we believe the pie is a fixed size. Payers, providers are all part of the same revenue stream we had better figure this all out together and in other states somehow hospitals compete with each other’s and payers and providers don’t like each other. So we just have a very good culture and that culture has enabled us to build infrastructure, so since 1997 we’ve had health information exchanges of the last couple of years since the state government has taken over our operation of some of this infrastructure. It’s just cloud hosted, transport, security layers, indexes, directories and that kind of thing.
(24:56) So yeah, we have a good enabler, but at the same time this whole federal policy at all of this stimulus we’ve been talking about has motivated people to start sending data across organizational boundaries. So it’s a combination of technology and policy and enablers. New England has just been blessed with all three climbing.
Vala:
(25:16) You noted in your blog and this was reference to your father in law as an example. My father in law has demonstrated to me that the meaningful use policy outcome of understanding of his care and respect for his privacy preferences. Is this ultimately leveraging technology to improve outcome from a patient’s point of view?
Dr. Halamka:
(25:17) Well so let me give you my father-in-law example. Six weeks ago he had a stroke, and by the way he has written consent that I can talk about this publicly, so there are no issues here of privacy and his challenge was like many of us, his records were at a primary care doctors office, a community hospital and an academic health Centre.
One ran Epic, one ran Meditech, one ran Homebuilt and the challenge was delivering care to him that was respectful and appropriate required merging all the data from three dispirit sources in real time to make clinical care decisions.
(26:17)Now I’ll give you a ridiculous example, I mean this is CXO and you know, being a CXO is awful. You know, I needed to check-in to McLean Hospital for a month ‘vacation’ and you know, James Taylor checked into the psychiatric facility many years ago, wrote Fire and Rain and it’s a bucolic spot.
(26:41) I also needed at this time at the Betty Ford clinic you know, too much cider there and oh, did I mention my sexually transmitted disease, HIV status and the domestic abuse.
(26:52) Well those things that I’ve just mentioned are completely fictitious of course, but they illustrate the nature of the privacy of the medical record. My preference might be share my Beth Israel deaconess data which is my flu shot, my allergy dependent to penicillin and my glaucoma, but don’t share my McLean mental health, my Betty Ford substance abuse or my Fenway community HIV records, that’s my choice.
(27:21) So what we have tried to do in this state is that outcomes depend upon care coordination, but maybe you don’t want your dermatologist to know about your mental health condition – then that’s okay, you know we’ve built the infrastructure to support that.
Michael:
(27:37) But how do you manage that, how does that get handled because how do you know what to a particular patient is something that is positive or negative. It depends on the context of that patient doesn’t it?
Dr. Halamka:
(27:53) So here’s what we do. At every registration of you in our facilities we present you a one-page, six grade level, easy to read, what are the benefits of your data are being shared across the community with caregivers for your benefit. Let me give you a multiple choice exam.
(28:12) You decide, do you want to share or not. You know, it’s totally up to you, you actually have to affirmatively state what your choices are sign it. And then we put that in electronic form in a message that goes up to the state. And the state uses it in a fascinating way.
(28:31) If you say yes then we put in a record at Epic level of who you are and the fact that you have a relationship with an institution and that you have opted in to disclose that relationship.
(28:40) If say no we use it to delete any history with your relationship with that organisation so that if the database was hacked, it’s not going to say John Halamka, Maclean psychiatric – no. There will be no evidence of any kind that I was ever there. The granularity is at the level of the institution, it’s at this point at the limits of technology.
(29:11) We can’t say we want three meds share those, two meds don’t, the problem is share that one but not that one – you opt in at an institution or location level.
Michael:
(29:21) I’m one of your consumers because I live here in Boston and my doctors and so forth are all associated with Harvard Medical School. And I remember those questionnaires and I remember thinking to myself, what kind of tricky agenda do they have in their minds asking me these questions.
Dr. Halamka:
(29:44) Right so our state laws in Massachusetts chapter 305 and 224, stipulate that we must ask patient permission before sharing a byte of data. Now other states have very different laws. We’ll share data about you until you tell us to stop, so the nature of the agenda that you see is purely to comply with our state regulations.
Michael:
(30:09) Let’s talk more about the data. Because we have been talking about the data but data as with many other industries is transforming healthcare. Maybe you can elaborate on that a little bit.
Dr. Halamka:
(30:25) So let me give you another personal example and again consent don’t worry no issues with Hippo. My wife was diagnosed to a stage IIIA breast cancer in December 2011. The genomics of her tumor were heard two negative or estrogen positive progesterone positive and at the time she was a 50-year-old Asian female.
(30:49) Now, wouldn’t it be interesting to say of the last 10,000 Asian females tumor like this, how they treated, what was the outcome. Did they get sicker, did they get well, what were the side-effects and then say, I will ensure that she gets the medicine that seems to provide the best outcome for 10,000 people like her.
(31:14) That is what we call a learning health care system, as opposed to today’s healthcare system it takes on average 20 years from an innovation from one hospital to defuse throughout the country.
(31:28) We were able with my wife to take all the data and all the Harvard hospitals and do the query and find the medication that would be most effective for her, and she is totally cured and everything is fine.
(31:40) So that’s sort of a big data, although I’m not sure what big data is – we have 3 petabytes so it’s not that big example…
Michael:
(31:49) This much data is big data.
Dr. Halamka:
(31:50) There you go! Yeah, we were able to treat her optimally using that kind of analytic across multiple institutional data sources.
Vala:
(32:01) Do you have data scientists that work in IT or who do you partner with to do this type of analysis?
Dr. Halamka:
(32:09) And so actually I actually have a whole department that focuses on that. This is tricky stuff because let me give you another personal example.
(32:20) When I was two years old, my mom give me pink medicine for an earache and saw that I had two red dots on my stomach and she said, you’re allergic to penicillin. So for 50 years I have had a medical record, allergy to penicillin observed by mom.
(32:42) What is the value of that data? Should I in the case of a life-threatening meningitis not get penicillin because my mom, 50 years ago thought that it might cause harm? So our data scientists have to be able to look at structured data and understand the prominence, the validity, the certainty, the quality and integrity of that data.
(33:04) And the challenge of course is that 50% of the data in an electronic healthcare record is not structured. So what do you do with something like, the patient whose mother had breast cancer is doing fine. Right, if you are going to Google search on that and you find patient and breast cancer.
(33:23) Or here is another example, if any of you have read any radiology reports, I love radiologists, they are some of my best friends. But their favorite breakfast food is the waffle and their favorite plant is the hedge. Right, so they will write a report like this. Michael has a totally normal chest x-ray, but there is slight haziness on the left that could be a fingerprint smear or he is dying of cancer.
(33:50) Right, and how does a NLP, Natural Language Processing system deal with negation, ambiguity, you know either or. So that’s where again a data scientist has to come in and say, you know this natural language processing stuff is kind of okay for some things, but we are not going to guarantee you that what it interprets is accurate.
(34:16) Remember, Watson thought Toronto was a US city!
Vala:
(34:20) We have a question from Twitter from Frank Scavo, a technology analyst and one of the top – so yes, one of the top technology analysts and Mr. Scavo asks is there an open source project around opt in healthcare data sharing.
Dr. Halamka:
(34:38) So what we have done in the Massachusetts Commonwealth is brought together all of Allah payers and providers and had a common set of policies and educational material developed by the state, a set of processes that are all common. So I mean, open source generally implies as it was a new license or some such thing, but I would say this is sort of in our community it is shareable IP that is used by all.
(35:07) I haven’t specifically seen an open source project, so he is one that I would hypothesize that we need. You guys have used XML and you’ve probably found that you can almost represent anything with XML, how about the consent assertion markup language. CAMEL kind of goes with HIPPO you know these are like animals, right.
(35:33) And what we could do is create a transportable open source consent for a preference in numeration that you can carry around on your iPhone. It’s a good idea and it has been done yet to my knowledge.
Vala:
(35:49) Hopefully we’ll have start-up folks…
Michael:
(35:51) Yes we need start-up folks, and actually sent a tweet to marketing a few minutes ago, saying he needs to connect with you.
Michael :
(36:04) So what about, god I’m trying to think of what is the most interesting, there is so much that I want to ask you.
Vala:
(36:12) Before the show you mentioned you have 1600 emails a day, well Michael what I’m interested to know where do you spend your time? Is it technology issues, is business issues, or is it healthcare issues. That’s just a tremendous volume of email on a daily basis..
Dr. Halamka:
(36:21) Right, so we are at this point in healthcare where – remember healthcare has lagged as an industry in the adoption of technology and its maturity of automation, and the acceleration of our work and that’s the 20 billion that has gone into the industry from the Obama administration is such that we are having to deal with the heady issues that the airline industry dealt with 10 years ago.
(36:54) so that 1600 emails represents the, oh my god, what should our policy be on this? What should be our technology on that? It’s just unbelievable exciting times and acceleration in my space
Michael:
(37:10) You know, it’s really interesting when you were talking earlier about your wife’s cancer and happily that is now in the past and the ability to search among across different hospital databases. The other day I had a conversation with Richard Spires, who is the former CIO of Department of Homeland and he was saying precisely the challenge of federal CIO is this siloing of information, very much as you were describing that you overcame. So maybe can you talk a little bit more about this notion of siloed information and collaboration going across departments and the things that you’re doing to help overcome these silos.
Dr. Halamka:
(38:03) Well sure, so in the past in a fee per service world getting data from hospital to to hospital would be like sharing competitive secrets, oh my, you’re going to steal my patient. I my patient because more procedures, more care is more profitable.
(38:22) But in this world of the affordable care act and in an ECO world we are actually paid for keeping patients healthy. And if we don’t share data and we have silos we’ll all go out of business. So the alignment of incentives to share data has finally caused us to break down the walls of the gardens.
(38:38) Now it is hard because I called something high blood pressure, Vala calls it hypertension, and you call it elevated blood pressure – we’re all talking about the same thing, your blood pressure is 150/90.
(38:54) And so we have had to as a country try to build the vocabulary standards, semantic in offer ability so that the concepts are recorded in different IT systems are actually understandable as you transport them from place to place.
(39:09) Now here’s the problem, if you look at the natural format that is used in automated teller machines, yeah what does it have, 20 variables. I think I know what a dollar is, I think I know what a date and time is, and I think I probably know what a bank routing number is.
(39:24) What’s a fever, so hey Vala, you feeling hot, Michael are you feeling chills, you know I don’t know – what’s a fever? And so the problem is so much of medicine is all these ambiguity and every doctor will be trained as an apprentice and record information differently. So this is our challenge, it’s creating common language, creating a capacity to send it silos and firewalls and competitive locations and then keep it safe.
(39:59) Because the other problem is that if I send you’re data to someone do we have trust and what if they compromise it, I’ll be held accountable. It’s tough.
Vala:
(40:09) Which technology do you think will be the most important in the coming years. We hear Internet of things and wearables and may be predictive analytics and algorithms using big data. Do you have a sense of you know if you fast forward CXOTalk five years from now what would we spend the 45 minutes talking about.
Dr. Halamka:
(40:52) Okay, so you of all people have her the acronym SMAC tossed around as social mobile analytics and cloud. Well on healthcare is often social cloud analytics and mobile – SCAM. You know that there is there are companies whose products are compiled on telli-point, which is a very powerful development language.
(40:55) But let stroll down on those technologies in where they could be real and important. Although it wasn’t precisely true, you saw on the news recently that doctors and nurses in Texas did not communicate around a patient’s who travelled to west Africa. Now why, well the nurses know and the doctors know they are in separate places, and in fact every doctor writes their own notes. What we in fact need is Wikipedia for healthcare, so why not have doctors, nurses, social workers, pharmacists all creating a group authored notes that represents the status of the patient based on everyone’s opinions and interviews.
(41:32) And that note represents the singular and not multiple source of info on that patient for that day. Then, underneath the note we have a Facebook wall – not really a Facebook wall but my gist is it’s the events that happened to you, Oh, here is a blood-pressure measurement. Here was something that your parents decided was important. Here is something that came in from a lab result or whatever.
(41:55) So in one place you are seeing group authored facts and then incidents and events below it. That’s a social documentation method that has never been used in healthcare. So let’s try something like that.
(42:09) Now what about mobile. So mobile yes, as you point out you know here is a device which if I were it identifies my pulse, my activity level, my sleep. It sends Bluetooth of energy to over here and then I can choose to make it available to my caregiver.
(42:30) That idea has got power. If you have congestive heart failure and you gain weight 7 pounds in a weekend, you’ll probably end up in an emergency department on Monday. We can double the dose of your meds and make an intervention because we were able to find out that you had a change of variants before you decompensate.
(42:49) So mobile increasingly getting data from the home.
(42:54) Analytics, we need to be looking for gaps in care. We need to understand what we need to do to keep you well by using protocols and guidelines and looking at data from multiple sources and figuring out where to fill in the gaps.
(43:05) Cloud, the hospital systems have often been written in the 1990’s aero technology, but they are migrating to Cloud hosted and mobile friendly infrastructure so we are going to be more agile. You’re going to be able to spin up an electronic health record at lower cost and much more quickly than we can today, and the hundreds and millions of dollars per site, and the years of work is going to be radically reduced.
Michael:
(43:35) Okay, we have just a few minutes left and have several questions, so let’s do a lightning round with you where we are going to ask you a few questions from Twitter and if you can respond pretty quickly, how is that?
Number one, from Sheri Reynolds, how I knew payment models influencing healthcare IT?
Dr. Halamka:
(43:59) We used to spend a dollar on IT and Blue Cross would benefit handsomely. Now, with the idea of Obama care and global capitated risk, every dollar I spend on IT and it reduces cost through quality results in I having more cash – I benefit, he who invests is he who benefits, that’s the big difference.
Michael:
(44:22) Okay, by the way I hope there are some reporters out there who are listening (lost broadcast)
(44:31) Another question from Execture interactive, how does (at John Halamka) find time to practice medicine. How do you find time to be a doctor and doing all of this?
Vala:
(44:44) But before you answer, please talk about union farm because they need to know about that bit and it will make your answer even more.
Michael:
(44:57) So yes, tell us about Unity farm and then tell us how you find the time to practice medicine.
Dr. Halamka:
(45:01) Right so three answers to that is being a CIO in 2014 is not a job it’s a lifestyle. It’s not as if there is any separation from life and work and it is all just continuous and its goes around the clock and it is 365 days a year no matter where I am in the world, my office is where my laptop is. So it’s the nature of the game.
(45:20)Now I come home after whatever work day and it could be 12 or 14 hours and I have a farm with 100 animals 55 different subtypes of apples, 11 different types of shiitake mushrooms, six kind of oyster mushrooms and guinea fowl, ducks, chickens etc. and I shovel manure.
(45:41) And keep in mind shoveling manure and being a CIO are very different activities. Very different indeed. So sure, as an emergency physician I have learned that I sleep when it’s convenient, you know, three or four hours a night and I have done that since about the age of 18.
(46:01) So may be the secret cancer is I’m 52 chronologically, but 65 by waking hours.
Vala:
(46:08) Okay, so before you ask one last question, for those of you interested in apple cider a recent blog by Dr Halamka, I just want to share the distribution here because in fact it’s very important. Dr. Halamka picked 147 pound of apples and made heart cider with the following distribution – very important. 8% crab apples, 56% Macintosh, 21% honey crisp, and 15% Macon, can you please talk to us about how you came up with the distribution.
Dr. Halamka:
(46:45) So here you take your typical apple and so here is a Mack, and you have a new Mack and I have a new Mack. Each apple has a characteristic of being sweet, tart, a stringent, an aromatic and the problem of course is that if you drink cider made from crab apple your tongue would curl. But if you drank cider made from just sweet apples it would be (coit? 47:11)
(47:11) So the trick is to after batch, after batch, after batch get the experience of the right combination of those four characteristics to have something pleasing. I tend to like having a little bit tarter cider, so I choose apples and a method of fermentation that makes a cider compare with food.
(47:33) And I’m a vegan, so maybe that is the other secret and I don’t spend all of my time digestion meat and dairy, you know is work and vegetables for me.
Michael:
(47:44) And your farm has a website and so you actually sell produce and so forth to the public right?
Dr. Halamka:
(47:53)We do and we have 100,000 bees, so we’ve got honey, mushrooms and you’ll find over the next couple of years as the farm continues to expand and our commercial quandaries will get greater and at the moment we are mostly regional in just surrounding farm stands around eastern Massachusetts.
Michael:
(48:13) Okay, well this has been a very fast CXOTalk.50 minutes and it feels like we’ve been talking for an hour and we can go on for another few hours.
Vala:
(48:27) Thank you so much for spending a Friday afternoon with us, an incredible honor for me and Michael.
Michael:
(48:35) You have been shown number 82 with Dr. John Halamka, who – this has been an awesome show so if you haven’t seen the whole thing I really urge you to go back and watch the replay. I’m Michael Krigsman with my what – friendly, he is friendly. People say he is the nice co-host and I’m the evil co-host.
Vala:
(48:59) Not many people say that.
Michael:
(49:00) You don’t say that.
Vala:
(49:01) No I don’t say that.
Michael:
(49:04) And that’s it, we’re done it’s been great talking with you John Halamka, thank you so much for joining us today and everybody thank you for tuning in and we look forward to seeing you again next time. Bye bye.