Dr. John Halamka joins us to discuss Promoting Interoperability Final Rule and what are hospitals expected to do about the Social Determinants of health factors in their communities. This plus a pragmatic look at IT.
Bill Russell: 00:08 Welcome to This Week in Health IT where we discuss the news information and emerging thought leaders from across the healthcare industry. This is episode number 35. Today we talked to the, to the Geek Dr. Dr John Halamka CIO for Beth Israel deaconess about the final rule for promoting interoperability, social determinants of health, and some pragmatic questions around health it. This podcast is brought to you by health lyrics. Health systems are moving to the cloud to gain agility, efficiency, and new capabilities. Work with a trusted partner that has been moving health systems to the cloud since 2010. Visit health lyrics.com to schedule your free consultation. My name is Bill Russell, recovering healthcare CIO, writer and advisor with the previously mentioned health lyrics. Uh, before I get to our guests, I want to make a quick note so everyone is aware of the great resource for your it teams. This week in Health IT has a youtube channel, easy for me to say with great insights from industry insiders, short segments, and complete episodes all curated for easy access. Every week we add another seven short videos and the entire episode, uh, over there. So right now we’re at around 300 videos. Check it out today, at this week in health it.com/video and share it with your colleagues. So today’s guest joins us from the home of the first place. Boston Red Sox, John Halamka, CIO for Beth Israel deaconess. Good morning, John. Welcome to the show.
John Halamka: 01:29 Well, well I happened to be here and remember I am also a red stock at red sox and medical allstar. So you know, I can send you pictures of me on the fenway big board.
Bill Russell: 01:39 Wow, that’s pretty. Uh, that’s pretty impressive. So you, it is amazing when you go from, uh, from ballpark to ballpark how much healthcare advertises and his is a part of the baseball program. So Beth Israel is, is pretty connected with the red sox.
John Halamka: 01:55 We are the official hospital,of the Boston Red Sox. There you go. And the closest emergency department. So if you are hit by a foul ball, you come to see me
Bill Russell: 02:05 and I’ll tell you what I hadn’t been to a couple games this year. Um, those, those standards appear to be getting closer and closer to the field. I mean, it’s, it’s not uncommon I would think for, uh, for people to get hit by a foul ball these days. I do take my glove now because some of those foul balls are a come screaming into the stands.
John Halamka: 02:25 Well, as you might imagine, we’ve done the analytics and the analytics. This is actually true. Looking at the cohort of those who visit Fenway Park over the last 10 years, what is the highest risk factor for traumatic injuries? Answer, wearing a yankees tee shirt.
Bill Russell: 02:44 True fact. That’s not a surprise. I, um, I, you know, short side note, I went to a playoff game at dodger stadium and uh, and it was the cardinals against the dodgers. Im a huge cardinals fan I wore my Cardinals Jersey we had to leave after the eighth inning, a because we came back and scored six runs against Clayton Kershaw and uh, people were saying things to me that I’m glad my kids were not around to here. And my buddy, who I was there with looked at me and said, we should probably leave now. And so we left. We left in the eighth inning, uh, out of, out of safety reasons. So that, that is a, that is a legitimate health risk. Wearing the, uh, the rivals shirt to a baseball game. It is true, uh, you know, the last time you were on, we didn’t get to talk about this but Tell us a little bit about, uh, Beth Beth Israel deaconess. You’ve been there awhile. Uh, you know, what’s, what’s the footprint, what’s your focus? What are you guys doing?
John Halamka: 03:43 Sure. So I have been at Beth Israel deaconess since 1996 as an emergency physician and as an it leader since 97. And over the course of time people say, well, wait a minute, how could you stay in one organization for that many years? Well, think about it. Back in 96, 97 machine learning was impossible. Um, it was hard. Stability, reliability, storage networks, very challenging. So those first five years were all about building infrastructure, the predecessor to today’s cloud. Um, and then, oh, security became an issue and then years of security and keeping our systems and data integrity good. Oh, and then how do I think about apps and how do I think about emerging services and integrating those and inoperability to look at air five years long has been actually a totally different focus. So now what’s next? Well, you can imagine healthcare gets better by getting bigger.
John Halamka: 04:42 So they say it seems to be the trend throughout the country that mergers and acquisitions, community hospitals coming together with academics, independent physicians joining practice groups, that sort of thing. Well, in Boston’s same things happen, so we have a 5.5 billion dollar merger of it is from Beth Israel deaconess and Lahey Clinic, Mt. Auburn, New England baptist all coming together and that will pose a set of interesting interoperability challenges. How do I coordinate care across 7 million people and do so in 450 sites of care that are going to be running several different EHRs. And so it’s more than just a few data elements for meaningful use. It’s actually ensuring that you’re getting quality and safe care at the right time, at the right cost us just fresh merged institution.
Bill Russell: 05:35 Right. So, uh, you know, just, just a quick side note question on that. So are you going to have the same ehr across your acute care facilities at least, or is that not going to be the case
John Halamka: 05:45 so think about this, regardless of your brand loyalty, epic, Cerner, Meditech, Athena, eclinicalworks, when a merger and acquisition happens day one, you’re going to have heterogeneity, right? And so you can imagine Lahey clinic, epic based, so it’ll have a cloud of epic, um, my community hospitals, all meditech based. So cloud metatech, Beth Israel deaconess self built cloud of Beth Israel deaconess self built. So you can look for the next couple of years and you know, that you’re going to have to deal with not a 100 different Ehr but three and you know, data sharing across those three. And then time will tell based on how the market evolves, will that become two will become one. It’s a little early to know.
Bill Russell: 06:38 Yeah. That’s interesting. So the, the, I would love to just go into that for the next half hour, but, but uh, but we’re going to talk about some other things. So could actually. The last time you were on, we talked, you shared about the gates foundation, the work you’re doing in a, in Africa it. Can you give us an update on what’s going on there?
John Halamka: 06:58 Sure. So the challenge in South Africa is $65 million people, 16 percent of the population is HIV positive. And the challenge of coordinating care across what is a very heterogeneous country, right? There’s urban, there’s rural, there are issues with infrastructure. Network bandwidth is expensive and slow and identity management. Who are you is actually a challenging question because names are misspelled. Workers move around. So with the gates foundation, we took the, the process of care delivery and broke it down into several what I’ll call API or core functions. So core functional one, who are you? So we’ll do identity management and we’ll do it based on bio metrics because name, gender, date of birth, math doesn’t work. So well issue somebody’s identity cards, hard to know. But biometrics, if I say I’m going to take, you know, fingerprints, I scan, you know, retinal or Iris, a palm vein geometry or whatever is the bio metric of the future, but build a system by which I can link your data by bio-metrics.
John Halamka: 08:05 That’s an interesting infrastructure. So sort of API number one is a general bio metric infrastructure and we’ve deployed that in the right to care clinics and now can tag your HIV laboratory data to you. So you just walk into a right to care clinic and it says on, here are the last five viral loads that you’ve had showing your medication is working very well or not. Those are from one problem to is how do I share the data with the patient, right? As we’ll talk about. I’m sure there’s this increasing trend in the U, s of patients getting access to their own data, their notes, etc. Well, hey bill, do you have an iphone? Ten will imagine that in South Africa my lowest common denominator is the nokia flip phone you had in 1997 running on a gsm network maybe. And so We’ve had to create a medical wallet for the patients that runs on a feature phone over very low bandwidth and that’s something we’ve deployed.
John Halamka: 09:12 We did a lot of usability testing and keeping a good number of folks in South Africa on the team really helped us what the needs assessment was. And then the final question is how do I deal with population health data aggregation and look at variations in care quality and understand trends. So what we’re working on currently is, is how do you expand what our early work on biometrics and this medical wallet to something that’s going to help for countrywide population health analytics. And that of course could be machine learning and it could start in South Africa and scale to other countries. So what’s the platform? So we’re starting to think through that.
Bill Russell: 09:54 Awesome. Well, I’m looking forward to, uh, you know, just continuing to get updates on that. That sounds fascinating. And we’ve talked a little bit the last time of why, you know, those things elude the United States health system for privacy and different political reasons and whatnot. Uh, and sometimes it’s, you know, the, the, the environment, the cultural environment, and the political environment will give you the opportunity to do things in africa that you couldn’t do here and hopefully prove, prove the concept out, get the, uh, get the statistics and bring it back, bring some of it back into the staes will be interesting. Um, so on our show we do two segments. We do a in the news,
John Halamka: 10:35 we have clearly
Bill Russell: 10:44 Did I lose you. Are you still there? Oh, I’m sorry. You know, This, this internet thing, sometimes it’s not as reliable. we don’t have a quality of service across this line. So, um, alright, so let’s, let’s, let’s move on because we do have a lot to talk about. So on our show, we do two segments. We do in the news where we each pick a new story and discuss and then we do sound bites, which is a series of about five questionS that, uh, I, I want to pose to our guests. So you hAve picked the, uh, probably the most important story of the week, uh, which is the, a cms finalized and promoting interoperability rule. So I’ll let you kick it off with the, uh, with the first story and if you could summarize it for us and we’ll chat about it.
John Halamka: 11:31 interOperability program. So important. So remember as chair of the hit standards committee, I was there as meaningful use stage one and two rolled out. They were wonderful in that they built a floor for functionality but they got a little cumbersome. And why did that happen? Because the meaningful use program, with it’s stimulus and it’s penalties and it certification was so effective. It became a policy vehicle. And so I’ll make this up. It’s not exactly right, but if the fda says, oh, we want to track medical devices, I know we’ll put that in meaningful use and the cdc says, oh, we want to track ebola. Oh, we’ll put that in meaningful use. And cms is, oh, we want 20 new quality measures. We’ll put that in meaNingful use. So by the time we got to the end of stage two, it got to be very challenging to figure out what you’re measuring, who does what and how much time it would take to the doctor to even do it.
John Halamka: 12:27 So what promoting interoperability program says, let’s scale back and think about just the few things we want to do really well. Create some really clean measures and let’s offer partial credit. It’s right, it’s not on or off black and white. If you as a hospital or doctor’s office are making progress and your trajectory is good, that’s fine. So what it says is, oh, let’s pick e prescribing an opioid, a what we’ll call interventions to reduce the opioid crisis. And so that’s sort of point one, it’s you prescribing and electronic prescribing controlled substances query of the prescription drug monitoring program, verify opioid treatment and some are bonus and optional. And again, partial credit is okay, so that’s their point one, point two referral management. Although we had in stage two, the idea of I send a summary to you, there wasn’t really the incorporation of that summary and it’s closed loop referral, so you’ll get referred to a cardiologist.
John Halamka: 13:31 You see the cardiologist, the cardiologists never tells your pcp at the plan. Well how useful was that? So again, there’s going to be a bi directional data exchange and incorporation, closing the loop and partial credit for progress along the way, providing patients access to their data, including notes and do that via api. So anytime an app comes knocking, the data is sent to the patient, that’s all Good. And then public health, SyndroMic surveillance immunizations, case reporting, public health registries, clinical trials, those sorts of things. So instead of saying what there’s 100 measures and all these different complexities, it’s four with partial credit, a 100 points possible if you get 50 out of those 100 points, no penalties. So this is really streamlining the program focusing assault. I said I think it’s a very good approach.
Bill Russell: 14:25 We actually covered this in a previous episode when they proposed it, but it’s a here’s a handful of things. So ehr reporting period minimum of a continusous 90 day period in the calendar year of 2019, 2020, 2015. Eahr cert is what’s required rule finalizes a new performance based scoring methodology, which you discussed a little bit, which meant to be less burdensome. Cms is finAlizing a new eve eve describing, measured, really focused on opioids. Um, changes. Uh, let’s see. Oh, and the changes to the measures of which again, you’ve talked about removing a total of 18 measures and de duplicating 25, which is huge. Uh, and then the other one which is interesting, require hospitals to post cost information on the internet, uh, in a machine readable format. Clearly we’re not covering everything. It’s 2,600 pages. You have, you had a chance to read it yet
John Halamka: 15:24 of course in detail and realize that 2,600 pages, like any of these regulations, 90 percent of it is preamble, right? So that’s, you can actually get everything you need to know by going into the appendices and looking at the tables because it’s really justification for why they did what they did.
Bill Russell: 15:46 Yeah. So it’s interesting. This is interesting to me, but uh, let me ask you this question. So these things generally come out mid cycle. Nobody has a financial calendar that ends in august and starts on September 1st. And uh, so this, this is hitting beth Israel right now. This is the final rule. Um, give us an idea of what process you’re going to go through a to generate the projects necessary to be compliant and to uh, to do the things you need to do to move this forward at beth Israel.
John Halamka: 16:15 So here’s the, Don Rucker at onc recognizes this and this is why they put in that 90 day 2019 requirements. So here’s what I mean. So here it is september of 2018, but you don’t have your certified ehr in place until october one of 2019, right? Because you have a 90 day evaluation and as long as it’s in by october one, 2019, you can get your 90 days. So all, I mean it turns out our fiscal year is october one to September 30th. So this hitting in august turned out to actually be good for us because I was then able to program all the interventions into the fy 19 budget and I actually don’t execute on any of this stuff until fy 20. So that for us was okay.
Bill Russell: 17:11 You can get ahead of it. You have gotten ahead of this and in reality we’re using the 2015 cert Most ehrs are there. Of course you have a homegrown one. So I guess for you to do.
John Halamka: 17:23 Yeah. And so we of course were the first ehr certified back in the meaningful use stage one is in all the functionality for the 2015 survey already. We just haven’t gone through the process. So that’s fine. It’s now in the fy 18 budget as a, the staff time necessary to take what is existence software and to go through the search process, which we’ve done multiple times already.
Bill Russell: 17:49 Yeah. So, um, so let’s, let’s talk about, you know, directionally what they’re doing here, what the onc is doing here, what cms is doing here. So you know, you go from administration to administration. How much does the change in administration change the direction and focus of onc, do you think,
John Halamka: 18:08 what’s fascinating is that that, and I’ve sort of bush, I’ve served obama and yeah, certainly stayed in touch with everybody who is on the current committee, the federal advisory committee in the trump administration and there’s a remarkable consistency and that is sure politics change but really the trajectory of it doesn’t so much so. A lot of the same people Who were at onc back in the obama administration are still there. Steve Posnack, John White, so they are diligently moving us along a rather consistent program and in fact sort of the theme of the current administration is less regulation, less burden, all the rest, but the themes of what we’re working on are fairly consistent, so I don’t feel like there was a revolution here. It feels like an evolution.
Bill Russell: 18:59 The only thing I see that’s a little slightly different in this is again a bent towards sort of free market. It’s a, we believe that access to information is, is critical. Not that interoperability is always about access to information, but this, this whole thing of, hey, let’s publish costs I think is fascinating because that’s sort of a free market mindset. If we start publishing costs, there will be transparency into how much something, uh, is, is, is going to cost somebody in how good the doctor is or how good the system is. And I think that’s one of the first steps for them in seeing this not as a universal healthcare program, but as a more of a free market program. We’ll. so we’ll have to see how this plays out. Um, you know, the leverS, the levers of the government pools do not happen overnight. They, uh, they generAlly take many years to play out. So by the time this starts playing out, there’s probably a new administration and, and change. And that’s one of the harder things to really, uh, for health system cios and health systems in general to sort of adapt to the constant change in the regulatory environment a cost money and cost resources and time. I mean, how, uh, I mean, can you talk to that a little bit? How do you, how do you prepare for that? How do you adjust for that?
John Halamka: 20:22 Sure. So governance is the key issue. So, um, I have since my earliest days as a cio had a guiding coalition of doctors and nurses and pharmacists and social workers and administrators who meet on a monthly basis to understand, well, what are the strategic imperatives, regulatory compliance imperatives, what it we need to do if there’s a sentinel event for safety and quality, how do we be impactful? so I have the hard discussion with them in 2009. I think. I’m sorry, but for the next five years, all of our business imperatives are going to have to be put on hold because we have icd 10 meaningful use, the hipaa omnibus rule and the affordable care act. And you must do them right? no choice, right? Because otherwise we’re all gonna go to prison. And so the governance groups that, you know, icd 10 is not a very sexy project, are you telling me we’re going to codify flaming waterskiis falling satellites and chickens hitting you on the hand?
John Halamka: 21:27 And I said to them, yeah, but must do. And so, but wait a minute, I have this pet project that is going to impact, you know, 25 doctors, sorry. And the governance folks gave me the air cover to focus on icd 10, those things that were, that exciting, but we’re most used. But so then here we are in this era as you described it, we’re out of that regulatory must do era and into a, oh, the private sector has to take the lead. You might have certain outcomes, you might have value based purchasing or whatever, but how you do it, it’s up to you. And my governance committee has now said, ah, we’ll fabulous. Let’s catch up on all the unmet needs of the meaningful use era. And oh, here a couple of strategic things. Oh, and by the way, we’ll make sure that what you do aligns with the future where we’re paid with risk contracts, but we’re going to do it our own way. And that’s okay. And as you point out, if five years from now we get back to a regular for a directed process, you’re going to need will help me through that.
Bill Russell: 22:37 Yeah, absolutely. So let me, let me take us to the next story and the genesis for this story to be honest with you, is a, I was looking at a conference that I’m getting ready to go to and someone was going to speak and they had this title, catchy title of why your zip code may be more important to your, uh, to your health than your genetic code. And I thought I remember that from somewhere. So I went to the all knowing google and uh, sure enough there was a harvard business review article. There was a, uh, uh, there was a and there was a huffington post article and that’s the one I remember reading way back in the day and it was circa 2009. And so I want to discuss that a little bit with you. Uh, you know, what’s changed since 2009. I know it’s becoming an evermore hot topic and boston’s an epicenter for some of this.
Bill Russell: 23:27 So let me, let me just summarize the article real quick. This is 2009 huffington post, why your zip code may be more important to your health and your genetic code. A Robert Wood johnson foundation did a study and they looked at social determinants of health and they had a few facts, and here are some of the facts. EvIdence suggests that medical care accounts for 10 to 15 percent of preventable early deaths. Some americans will die 20 years earlier than others who live just a few miles away because of differences in education, income, race, ethnicity, and where and how they live. College graduates can expect to live five years longer than those who do not complete high school. Middle income people can expect to live shorter lives than higher income people, even if they are insured, and people who are poor are three times more likely to suffer physical limitations from chronic illness, in other words, as it relates to your health are zip codes more important than your genetic code. So, john, you live in boston? It’s fairly fluent. Has its pockets but it’s fairly fluent and a progressive city with a universal healthcare. Uh, are the people of boston healthier today than they were in 2009? Let’s start there. Are the people in boston healthier today than they were in 2009?
John Halamka: 24:41 It’s a very complex question. Let me give you a preamble on an answer it directly. So, you know, I’ve been a vegan for 25 years and so I eat nothing but Plants, right? And so, no, I don’t eat eggs. No, I don’t eat fish. Those aren’t plants. Uh, and so what does that mean? Because I’m a vegan? Well, it means my cholesterol is 72. My blood pressure is 1:10 over 70 in my body. Mass index is 21 and b, everyone says you’re going to live to 100, but why would you want to? Uh, but so of course, shouldn’t it be rational that you should go to your insurance company and say insurance company, I get a safe driving discount for getting no tickets. How about a safe eating discount for being vegan and the insurance company says um, that’d be no.
John Halamka: 25:30 in fact, what we need is people like you to fund all the people who are eating fast foods every day, so thanks and so we haven’t in this country had an alignment of incentives to change some of the lifestyle issues that you mentioned, whether that zip code or what you eat or how you act, but what’s changing and why do I actually think things are getting better in boston, so not only, yes, we have this universal healthcare coverage so that there’s a penalty if you don’t have insurance, that’s good, but what we’re seeing is the move to value based purchasing then move to risk contracts with both upside and downside is so extensive in boston today, September, 2018, 80 percent of the reimbursement of beth Israel deaconess services is risk contracts or value based. So what does that imply? Well, that implies we better have an aco is looking at social determinants of health, right?
John Halamka: 26:33 If you are living alone and you don’t have access to the right food, the right transportation or an air conditioner, you’re going to be a high cost utilizers. So we actually need case managers and care navigators and visiting nurses to deliver care in your home to you that keeps you healthy because we are now at risk for your wellness. And so I spent $7 million last year on just building out all the infrastructure necessary for that care management care navigation, visiting home nurse service to keep people healthier in their homes as incentives have changed and therefore our business has changed.
Bill Russell: 27:16 Yeah, that’s fascinating. So risk, and that’s really true in the orange county market as well. Uh, you know, kaiser takes on the risk and we heard this when I talked to mark groves from intermountain, they have picked a zip code where they are going to assume risk and a sharp healthcare out of san diego a, they have a significant amount of risk and they, they act differently. I mean, they really do that assuming risk takes them away completely away from fee per service. So then their, their economic incentives are different. Um, so let’s look at this from two perspectives. One is, uh, you know, they’re not in boston, they’re not in Massachusetts. How can healthcare organizations, um, you know, fund, uh, is, is getting risk contracts the only way to really fund being a part of social determinants outside of the altruistic nature of wanting a healthier community or are there other ways to fund really these social determinants projects?
John Halamka: 28:17 Yeah. And it’s again, very complicated question. Um, so for example, um, I recently took a bunch of engineers from a, from an industry three bolt into much mix and the industrial engineers said about, you know, we want to create products and services for consumers that will help consumers manage their own health and so that I wanted to talk to a couple of patients. So they went over to this homeless gentleman and said, what is your favorite wearable? And he said socks. And so we have to be a little careful because there are the consumer driven products that will help with social determinants of health. But the most needy of the, our, our, our patients are unlikely to probably buy and manage their own health well. And so that’s where I think this idea of risk contracts, medicaid acos are going to really help with that. Um, there may be models in the future where a can have general healthcare and accelerate social determinants issues regardless of reimbursement to the provider, but for the moment it does seem to be very on the way a provider is reimbursed.
Bill Russell: 29:44 Yeah. So cvs is in your backyard and their, their, their pitch, if I understand it correctly from the conference where I heard a gentleman speak, uh, is really to step into that care navigator, be that first point of contact, be that, uh, uh, almost primary care physician, but helping them to navigate these complex health systems like yours, um, are you partnering with them or where do you see them fitting in? And how are they going to mesh into the environment that you described?
John Halamka: 30:13 So we have no specific partnership with cvs caremark, walgreen’s, et cetera. But we worked very closely with all these organizations in our ecosystem because with a risk contract, they say you need to understand inpatient and outpatient and ed and urgent care and sniff and pharmacy and all these other kinds of options. And make sure that their, the patients are given the right treatment at the right time, in the right place, at the right cost. And so why do I think that what cvs caremark or thinking is good? Well, so my quick story for you, uh, some years ago my wife was sitting at the dining table with my father in law. My father in law began speaking in word salad, raw total nonsense. He had no issues with cognition or movement. It just the words coming in and out of his mouth were random. And so what does she do?
John Halamka: 31:06 She calls me, I say, okay, he’s having a stroke and broke his area and it’s expressive aphasia. And what we need to do is get them to a ct scanner at a local community hospital to see if he has a bleed. He doesn’t need academic healthcare, he doesn’t need a tertiary referral center, he just needs a ct scan. And so we go to the lowest cost of care ct scanner that’s closest to our home. And then do a telemedicine telecare connection to a neurologist expert who interprets that and delivers the right care at the right cost at the right time. And he did totally fine. Now my wife called me. That’s not a very scalable model. If every patient has to call the physician and the family,
Bill Russell: 31:49 well, can you, can you share your cell phone number so that we can try to scale it and see how it goes.
John Halamka: 31:54 Perfect. And so the idea that there is this care navigator that directs you to the appropriate quality complexity and cost is really a needed function and some of that will be human, but I also have to imagine over time will develop machine learning models that will help a little with that kind of thing. and remember, machine learning is not going to replace doctors or anything. It’s going to augment care delivery so that everyone can practice at the top of their license and so, what do I mean by that? So here, let’s use an example that you can’t possibly see, but uh, so it turns out there’s a spot right there and that spot is brown and it’s circular and it’s flat. Is that skin cancer? It’s been there for 20 years. It’s homogenius in every way. No, it’s an age spot now. It turns out whether it’s google or amazon or a startup, you know, you can actually take millions of dermatological photos and suddenly put a probability on a novel photo, whether it needs a consultation or not. And so that sort of thing suddenly will help our dermatologists get the right cases to review and the pcps to know when to review and that kind of thing. So I just have to guess that cvs caremark and the like providing services augmented with guidelines, protocols and machine learning is a really good future approach.
Bill Russell: 33:24 Yeah. And then the, so the other question I wanted to ask you, I was with, I was with somebody this week and we were talking about this and they were um, they were saying, you know what doctor gives the most referrals in the, in the country. I’m like, well, it’s impossible to answer. He said, well, not really, but the doctor that gives the most referrals, if anybody in the country is google, dr google and we started talking about that. It’s like, you know, people will type things into their search bar that they won’t even tell their physician that they won’t even ask somebody else. And a lot of times, hey, my father or father in law is, is speaking a word salad. I’m their first inclination is going to be alright, I’m to, I’m just gonna hey siri, this is happening. What’s going on? Um, so how do you think about that? Oops. I said, hey, siri on my phone with um, but the, uh, how do you think about that as your trying to develop the next round knowing that in the boston marketplace, in your marketplace, people are going to consult google and what you want them to do is as quickly as possible get from google to a, uh, a qualified care navigator or physician within your system.
John Halamka: 34:38 Right? So here’s a couple of thoughts to that. And that is our philosophy at this point is that the ehr is a fine transactional system for compliance and regulatory, getting the bills out and that kind of thing. But is the ehr going to provide the level of innovAtion that we need to solve the problem you just mentioned? Probably not. So what are we starting to do? We started to create apps. Some are patient facing, some provider cases, but example turns out that we have 3000 doctors and you may know that I am the internationally recognized specialist on mushroom poisoning for every patient in the United States. bizarre as that sounds. I do 900 consultations a year.
Bill Russell: 35:27 Wow.
John Halamka: 35:31 Yeah. We in our app identify among our 3000 doctors. Sure there’s an orthopedist. But who’s the guy who’s the specialist on the right shoulder? You know, who is the person who knows more about mushrooms, whatever she got to the app. This is the nature of my sign or symptom. And it’s not a google search. It is actually a curated metadata driven way of directing you to the right care. So again, everyone can practice at the top of their license. So we’ve written that.
Bill Russell: 36:05 That’s interesting. I would just like to get. I realized that that’s an app way of doing it, but you know you have this.health domain. It would be great if that becomes the curated content where I can go and do a Search and I don’t have to worry about how good is this content, is it directed me in the right direction? How do we get thAt into more people’s hands? Actually, I. We could probably talk about this while I think it’s a fascinating problem. if people are going to google, how do we, if they’re going to google or bing or whatever they’re going to, how do we, how do we leverage that to better health across the board and I understand that the homeless person isn’t going to google and there’s other things around social determinants. I don’t want to get mail from people around this. I get it that it’s social determines is bigger than this, but there is a significant portion of people that they’re. The first thing they talk turned to is the internet and I think that’s what the health domain was sort of geared towards, is much higher quality data. IS that, is that your understanding as well?
John Halamka: 37:11 Yeah, it’s early though. Very early. And so what are we doing to your point? I think it’s, it is our responsibility to actually create alexa apps so that you could be in your home and say, my father is speaking in word salad, what do I do? And so at the moment we’ve created 30 different alexa skills and uh, everything from care planning, helping you understand when to take your medication, scheduling appointments and those sorts of things. So I’m hopeful that where we can get to a point where it’s ambient listening in your home, supported by cloud hosted decision support services and it gets as easy as, you know, alexa, ask bidmc x and then you get, because we’ve registered the keyword bidmc, the curated suggestion as how to navigate from there.
Bill Russell: 38:10 Exactly. Well, great. Uh, all right, so we’re going to move into our next section, which is a soundbite section here. I just tossed out a five questions and you’re going to give one to three minute answers. If you go over, there’s no buzzer, there’s no whatever, you know, we’ll just, we’ll. It’s more of a guideline than a rule. We’ll, we’ll just go down the path. So I’m, so I’m traveling, this is my east coast office, if you
Bill Russell: 38:36 will, and uh, hitting some clients this week and next week and I was in the room this week with two senior leaders on a healthcare it team that came in from outside of healthcare and it was fascinating, you know, very talented people from banking, from a manufacturing. It’s interesting where these people are coming from and they’re being brought in because of their specific technology skills and, and, uh, you know, new models that they’re looking at, those kinds of things. Do you think this is a trend that’s going to continue and what do you think the hardest thing for these people to really grasp about healthcare and the transition to healthcare, uh, is going to be when they come in from the outside.
John Halamka: 39:16 Sure. So I guess two thoughts to that is that if I were to ask, where do I want to partner with innovators, is it retail, is it banking and finance answers, it’s actually consumer. And so look at what google and amazon and apple and these sorts of folks are doing. They’re creating tools and technologies that at very large scale can empower a lot of interesting innovation. The challenge is is those organizations don’t have the healthcare domain expertise precisely. And so if they offer a cloud hosted machine learning service, it’s now my responsibility to make sure that it has the appropriate domain knowledge integrated into it. When I create an application that’s okay, you know, consumer companies go create the generic and then I will go create the vertical and using that tool
Bill Russell: 40:08 interesting. And that’s what they’re looking for when they partner with a health system. They’re looking for that deep domain knowledge. They, uh, and they recognize, I think now more than ever when you talked to a silicon valley startup, uh, or even one of these bigger players, they will say we need strong health system partners, a partner with physicians. And have those conversations there. Yeah. Um, so, uh, second question, this is more broader, so healthcare analytics getting a lot of, a lot of conversations around this and you know, because it can veer off at this point at this juncture in our history, it can veer off in a lot of different directions. It’s still predictive versus a retrospective and those kinds of things. But predictive, it’s getting more interesting and you’re looking at machine learning, you’re looking at ai and some other things. Um, so this might be another question that you answered with governance, but how have you structured your healthcare analytics practice? Um, to optimize it for Success and, and the changes that are likely to come here in the near future.
John Halamka: 41:13 Hey, of course, as you pointed out, is our requirements driven, which is in order to be successful running an aco, you need a set of benchmarking reports that are retrospective, that are looking at quality and cost and variation across different providers. what’s the trend is moving away from business intelligence and to machine learning. And as you suggest, I am going to take a patient today and based on machine learning techniques of 2 million patients before them, suggest what the right course of treatment should be and start to schedule the interventions I’m going to make. You have to pick your use cases a little carefully there, right? This is not replacing doctors with machine learning. That’s not it, but it’s saying aha, you need a surgery today. Actually look at 2 million patients like you and here’s who should do and how it should be done and how long it will take. And that’s the kind of interesting, not so much business intelligence because it’s more complicated than I’ve looked at your age and your ethnicity and your co morbidities and model it in a predictive way. Using a machine learning approach.
John Halamka: 42:24 We’ve got about a dozen projects that our governance groups have suggested are appropriate use cases for that approach. So your analytics, your analytics products are really bubbling up from all over the organization from a lot of different governance groups. There isn’t a analytics group per se. Well, so In a $5.5 billion dollar organization, you can imagine that you have a lot of stakeholders, so sometimes the stakeholders are the accountable care organization. That’s one set of analytics. Sometimes it’s the quality folks, sometimes it’s the compliance folks. So sure, all this ultimately bubbles into governance, but I believe in it, you know, very federated approach which is a delegate to the aco, what analytics they need. I mean, I can’t decide on their behalf and hopefully I’ve built this generic infrastructure of normalized data that is accessible via a variety of tools and the new machine learning capabilities that addressed these various stakeholder needs. Yeah, it’s interesting and I think you probably know this, but there’s so many different models out there around analytics and how it bubbles up and how it’s governed and data governance resides. And uh, it’s, it’s really, it’s one of those areas that’s fascinating to me. I’m going to keep diving into it on the show and see if I can on earth the different models that are out there. A third question, and this gets a little geeky, little more tactical, but where have you seen agile and scrum done effectively in health it?
John Halamka: 43:59 So a challenge of course is that I’m given every project that comes across the transom in it you got two or 300 different threads at anyone time. It’s really hard to do innovation when you’ve got two or 300 projects that are just keeping the lights on. So where we’ve started to use this more agile approach is I’m going to call it almost analogous. You know the google 20 percent work on something different than you are job approach. So we’ve created a meritocracy where if you have proven yourself to be particularly skilled and resilient, we are going to give you the capacity to spend some amount of your time every week on a radical new breakthrough using an agile methodology with a notion that we’ll do it fast and will fail fast. So the thing that it was a bit of a caveat is when I see people doing innovation centers as skunk works, just separate from operations, doesn’t work so well because you can’t get adoption. But when I take people on the inside of the operation, carve out some protected time, give them an agile methodology and have the move really fast, try things and fail, at least for us. That’s worked.
Bill Russell: 45:25 Yeah, that’s awesome. um, fourth thing what are, what are the qualities of a great cto, chief technology officer, uh, within healthcare do you think? What are you looking for?
John Halamka: 45:40 Yeah. so here’s an interesting challenge so I’m 56 young, young, but what was the technology that I grew up with? Well, you’re a journalist, it was called smith corona and I actually know what a carriage return is. Right? And so my problem is, is a 56 year old. I am biased by 50 years of experience of technology that I had to build and sweat and it was all hard. And so what I want a cto to be is somebody who has more neuroplasticity than me. When when somebody approaches them and says, oh, there’s this great way you can use twitter, you know, w, I might roll my eyes because it’s a thinner isn’t necessarily the native technology I speak. But a cto should say, well, I can actually, I can look at that objectively. And so it’s that I don’t come with not invented here syndrome. I don’t come with too many biases. I’m willing to be very objective evaluator of whatever comes across.
Bill Russell: 46:53 Yeah. So how many ibm selectric do you still have in your health system?
John Halamka: 46:58 Um, that I think there, I, you know, I can’t count, but certainly would be on the hundreds. Wow.
Bill Russell: 47:06 That’s amazing. Yeah, that’s amazing.
John Halamka: 47:09 Right? Forms. Try filling out a government form on a, on a printer.
Bill Russell: 47:16 Yeah. You have to find one of those old oki data dot matrix. Yeah. with the special forms. Uh, and quite frankly a selectric. It’s just easier for some people use. Um, so last, last question here. So you maintain, as people can tell, you, maintain an extremely positive attitude. I’m sure you’ve had many difficult meetings, uh, as many as the next cio. How do you keep from becoming cynical? How do you keep that positive frame of frame of mind?
John Halamka: 47:50 So what’s really important for an it leader is to have something in your life that grounds you. Right? So if I got customer emails this afternoon that said I hate you, you’re horrible. I go out and hug the lamas. Right. So remember I run a 70 acre organic farm that is a, the animal rescue for the entire boston region. And so I have horses and cows and pigs and I’m up at 4:00 AM shoveling manure and I have a bad day. I’m just out in the barn. Something that yoU can look to that is some part of a greater good. KeepS you grounded.
Bill Russell: 48:34 Yeah. And the, uh, the animals aren’t tweeting out a negative comments about you these days. I assume
John Halamka: 48:42 They are writing anonymous editorials in the New York times though, those llamas. you gotta watch em.
Bill Russell: 48:48 They’re crafty sitting over there with those neural implants connected to the internet. Uh, anyway. Hey john. Thanks. It’s always awesome to have you on the show. Um, what’s the best way for, for people to follow you?
John Halamka: 49:01 Sure. So of course, you know, I blogged though, I’m doing a little less blogging these days because the attention span in general in the world is less so twitter @jhalamka and I’m also on facebook and of course you can find out what’s going on on the farm by looking @unityfarmsanctuary.org,
Bill Russell: 49:22 unityfarmsanctuary.org, a great, fantastic. You, uh, you can follow me @thepatientscio. You could follow the show @thisweekinhit on twitter as well, a website thisweekinhealth it.com, a catch all the videos on the youtube channel, which we talked about earlier. And, uh, and we will be cutting this show down into three minutes segments for our attention span. Uh, and actually it’s not a, it’s not a joke, there’s so much going on in it, you just have to make it consumable for people. And so that’s what we’re doing here. Uh, so thanks again john. Uh, please, uh, you know, for our guests, please come back every friday for more news, information and commentary from industry influencers. That’s all for now.
Dr. John Halamka joins us to discuss Promoting Interoperability Final Rule and what are hospitals expected to do about the Social Determinants of health factors in their communities. This plus a pragmatic look at IT.