This Week Dr. Lee Milligan joins us to discuss the privacy and ethics of 23 and Me / GlaxoSmithKline deal. In addition, we take a look at Employer lead healthcare initiatives and explore scaling analytics in a health system.
This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.
Welcome to this Week in Health It where we discuss the news, information and emerging thought with leaders from across the healthcare industry. This is episode number 30. Today we take a look at 23 and Me Privacy Scandal, and we check in on uh, J P M. Berkshire, Amazon Health, uh, which really cannot get a name soon enough so we don't have to keep saying J P M Berkshire Amazon Health.
Um, this podcast is brought to you by health lyrics. Health systems are moving to the cloud to gain agility, efficiency, and new capabilities. Uh, getting to the cloud and operating. It represents new skills for the IT organization. Work with a trusted partner that has been moving health systems to the cloud since 2010.
Visit healths com to schedule your free consultation. My name is Bill Russell, recovering Healthcare, c I o, writer and advisor with the previously mentioned health lyrics. Before I get to our guest, an update on our listener drive, our sponsor has agreed to give thousand every additional hundred subscribers to our iTunes channels.
We are just 30 subscribers. Short of raising $3,000 for Hope Builders, an organization that provides disadvantaged youth life skills and job training needed to achieve enduring personal and professional success. I've hired their graduates and their stories are really inspiring. Join us by subscribing today to be a part of giving someone a second chance.
Today we're joined by one of those people whose articles I really enjoy reading, someone who has one of those bios. And, and we have several guests on this show that make me feel like such a slacker. And, and your bio is another one of those that just makes me feel like a slacker. Uh, today we're joined by, uh, Dr.
Lee Milligan, c m i o, for Asante Health System in Oregon. Good Lee, and welcome to. Morning Bill. Very happy to be here. Thanks for having me. Well, you know, your, your bio does make me laugh a little bit. It's, uh, it really is a, a, a blueprint for a type A personality. You know, you have, you have four wonderful kids and, and, uh, a wife who's, uh, also works in the healthcare industry.
You have, uh, you know, the, the usual credentials graduated from University of Utah, George Washington, university of Med School, uh, U C L A internship. Uh, you practiced medicine for 10 years, but then you, you know, you did that, that crazy stuff. Like you, you went back to school in 2010 to pursue a computer science degree.
I mean, what precipitated you doing that? I mean, why? You're, you're, you're already a physician, why go back for a computer science degree? Yeah. I've had a lot of my friends ask me that same question. Um, I think ideally it came down to, uh, I started to see this real bad accident about to happen between this intersection of medicine and computers, and some of my friends at other institutions were on EHRs and we're describing the really bad experience.
Uh, utilizing it. And so we were on paper at that time and we were thinking about going to an E H R and it just dawned on me that kind of getting my head around this space would really be advantageous not just for myself, but for. Yeah. And so in, uh, in 2011, you guys make that e h r uh, leap. You become, uh, epic credential trainer.
You also become an epic physician builder in, uh, in basic orders and analytics is, I mean, do you, was that really valuable? And, and is, is that a, a something that you really encourage your physicians to do, uh, at, at your health system? I do, we have, uh, 11 physician builders here. And, uh, I feel like it's a critical element of getting the docs part of the conversation.
Um, I think it's hard for docs to really want to be part of that whole process if they feel like they don't really even understand the vernacular of what the IT folks are talking about. And in many ways, this program is a way to introduce them. To those concepts. So I think for the system, it's a terrific, uh, uh, advantage.
So, and, and this is, this is where, you know, I just feel like a slacker. So you have, uh, multiple board directorships, the health system, uh, the physician partners, uh, propel Health, organ, a c o, uh, and uh, I guess in your spare time, you're also a C M I O, um, or is that your, that's your primary job, right?
That's my primary job, although I will say that, uh, my areas of focus are on data analytics. Yeah. Uh, data governance, uh, this physician builder program and most recently our health information systems, uh, department now report to me and we're in the process right now of unifying our ambulatory and our inpatient health information services.
Which is great. So I, that's what I'm looking forward to in the, uh, soundbite section. We're gonna really go into informatics and, and what you guys are doing in that area. So you're also board certified in clinical informatics, which you say is the hardest board exam you've ever had to take. Uh, is that really, is that really true?
Yeah, I mean, I, I think for everybody it's a bit different. Um, you know, it's been around for five years now and like a lot of these, uh, new board certifications, they start out by grandfathering folks in if they've had enough experience. And then eventually you have to do an actual two year fellowship before you qualify to take the test.
I was part of that first cohort who qualified to take the test without the fellowship, but the test itself was pretty dang hard. Studied pretty hard for it, but it was a really great learning experience. I have to say, a lot of the stuff that I applied today, I actually learned in the preparation for that board exam.
Wow. So, um, so one of the things we like to do is just ask our guests, you know, what they're excited about, what they're working on today, and, and really, I'm just gonna turn the floor over to you and talk about what, you know, whatever your, whatever has your, your interest right now. Okay. Um, there's a lot of things, uh, on the plate and a lot that hold my interest, but I would say in general, the theme of minimizing friction is top of the list and that that could be patient friction.
Um, you know, near and dear to my heart is provider friction. Um, and one of the ways that we're looking at doing that is really improving voice recognition. So, right. You know, we've used voice recognition up till now, but it really, we haven't really gone further than a kind of a superficial utilization of it.
I would love to get to the space where eventually we're using voice recognition to do all of our navigation throughout the electronic health record. We could just simply say, we. Pull up chart review, pull out all the, uh, cts I've had in the last three years and list them for me and then you can see them beautifully.
I think we're on the cusp of getting there, but not quite. Um, but that's really what gets me motivated right now is that voice recognition and n l P world that we're heading into. Um, the other piece from a patient perspective is just, you know, simple stuff like optimizing the patient portal. Simple concept, but really, if you get that thing right, it impacts everything else.
If patients can see their, um, their lab results, their chart notes, then they're not actually calling our call center. And so that volume comes down. So it has a tremendous, um, enterprise wide effect. We get that piece right. Yeah. You know, we don't, we don't talk about, uh, vendors that much here, but I mean, when you talk about voice recognition and voice navigation, are you talking about, uh, like, like just the, the basic nuance kind of stuff that nuance products or are you talking, uh, a, a suite of products that you're looking at right now?
So both m Modal and nuance are, uh, in this space. They're kind of the Microsoft and Apple kind of going at it, uh, which I think is good. I think it's healthy for, uh, for the industry to have those two kind of going at it. And yeah. Uh, we happen to use nuance, um, and we're looking at their cloud-based program now, which, you know, the, the cloud-based voice recognition has a lot of advantages moving forward, including the fact that from the provider perspective, you don't have to train it.
So normally when you have a network edition, the doc has to sit there for 15 minutes and train this thing to get it up and running, and frequently the file corrupts. So now picture this, it's two in the morning. You know, you're seeing patients in the er, you're flying through patients. The waiting room's packed, you're about to dictate, you know, your thoughts on a patient, and the thing corrupts.
Right. Yeah. And, and so we've been doing, we've been doing charting and notes for a while within Nuance. I think we were doing that, um, probably five or six years ago, maybe seven years ago. But, uh, but now when we get to navigation, I mean, it really becomes more of a Star Trek kind of thing where you're just, you know, Hey, computer, tell me what the vitals are for this patient.
Hey, computer. You know, that kind of stuff and yeah, that, yeah, and I think I, that's a good point. And I think, you know, that's where we get over that hump of us working for the computer versus the computer working for us. Yeah. I think the first couple years we were on the E H R, really, we felt like we were all data entry monkeys putting stuff into the computer and it wasn't really doing much for us.
What you just described in my mind is really the computer flipping that and now working for us. Yeah. And it's, it's, you know, it's those movies that gave us the picture and you have your, uh, Luke Skywalker hope, uh, poster there behind you telling everyone Yes. My, my nerdiness is coming out. Yeah, . Exactly.
and my Star Trek reference. So, uh, all right, so let's, let's jump into the show. So what we do is, uh, in the news and soundbites in the news, we each pick a story to discuss and soundbites. I ask you a series of questions and, and we'll go into analytics this week. Um, so I'll kick us off 23. So story. It's 5 million client genetic, uh, data with GL Drug giant GlaxoSmithKline.
And, uh, this is from live science. I'll read, read some of this. Actually, I'll read a bunch of this because it's, it's really fascinating. So, uh, 23 and Me GlaxoSmithKline, um, have, uh, signed an agreement. They're sharing, uh, the D N A data, uh, during a four year collaboration in the London, uh, giant. Uh, Glaxo and 23 and me, uh, will get access to 23 and me's genetic database to zero in on possible targets for treatments for human disease.
The goal of the collaboration is to gather insights and discover, uh, novel drug targets, driving disease progression and develop therapies. Says, uh, spokesman for GlaxoSmithKline. And, uh, it's not yet clear what conditions, but here's an example, uh, that they give. And it's, uh, the, the, uh, the gene L r r K two, which is linked to some cases of Parkinson's disease.
Only about 10,000 of 1 million Americans with Parkinson's disease have this disease because of that gene. So GlaxoSmithKline has a test, has to test about a hundred Parkinson's patients to find just one potential candidate. However, 23 has already provided two 50 Parkinson's agreed recontacted. For glaxos uh, clinical trials, which may help.
The pharmaceutical company developed a drug much faster. Uh, however, not everybody's really excited about this. Uh, uh, one quote, if a person's d n a is used in research, that person should be compensated. Uh, it says Peter Pitt, center for Medicine in Public Interest. And, um, Goes on to say, you know, if they opt in, they should be refunded the, uh, cost of the, uh, test.
Uh, in addition, even though 23 and Me gets consent of its customers to use their genetic data, it's un unlikely that most people are aware of this. And, uh, Yale, uh, law School, information Society project, um, fellow Tiffany C. Lee says The problem with a lot of these. Ts and Cs and privacy policies is that no one really reads them, which is probably true.
We've all clicked through those pretty quickly. Um, And, uh, but it's important to note that the new collaboration isn't the first time of 23. And Me's vast pool of genetic data has been mined by scientists. San Francisco startup has already published more than a hundred scientific papers, and, uh, it goes on to say 23.
And me has more than 5 million customers who've had their d n I analyzed for ancestral data. People who would like to close their account, they actually can do that. However they go on to say that, uh, any, uh, research that has been done to date, Uh, will not be withdrawn, uh, but they will throw away your samples and those kind of things and not use them moving, uh, forward.
So, uh, social media runs the gamut on this. You know, there's, there's, uh, here's some quotes. This is an outrage, another quote. The whole enterprise is of fraud. They're selling to naive people, useless tests under false pretense. Then there's who are responding to. You didn't think their whole business model depended on $99, uh, test kits.
Did you? Uh, and then there's, uh, even some support, you know, it doesn't bother me in the least, uh, the, the quote, you know, it doesn't bother me in the least, uh, if they're able to help just one person carry on. So, alright. Uh, that's, that's a ton of data. I just wanted to set that up. So, should we be surprised or even outraged that, uh, our genetic data is being sold?
I think this is the, the question of the 21st century. I think we're gonna be having our, um, bio, I'll call it biologic information, uh, captured, stored, and sold, um, in a variety of formats. Whether it's facial recognition or whether it's d n a, um, I think it's gonna be, um, something that will be from a technology perspective ahead of the law.
I do think, uh, the legal aspect is an important one, and I think the, the legal folks need to, um, uh, move a bit faster in terms of keeping up with technology. But right now it's a legal thing to do. Um, if you, you know, that that really long thing people are signing, we all do it, as you said. Um, but I think from a legal perspective, it is legal.
Um, is it ethical? I think there's a lot of debates on both sides of that. Um, I think the business aspect. Um, that people have called out is interesting. I wonder if moving forward, if they could gather even more information from the individuals and then pay them for that more information. So, for example, I.
Everything is sent over de-identified, but what if people were willing to be identified and then on top of that, they were willing to provide additional context to their clinical scenarios throughout their life. So they could, you know, I've been a smoker for 30 years, you know, or, you know, I have a family history of X or I've personally had y That would be really helpful in terms of connecting the dots between the, the genes themselves and the actual phenotype.
The outcome of what, what occurs. So, um, That's just one thing to consider. Yeah, no, that's, that's interesting. And I, and I think that's, you know, I always come down on the, um, on the side of the patient on this. So I, I look at it and somebody gave me this analogy and I forget who it is. I wish I could give them, uh, Credit, but, uh, we're, we're almost like digital serfs working the land for, for the lords of the internet.
So we, we generate all sorts of data on ourselves and then we allow them to sell it. And they're, we've done this on another show. We talked about how the medical record and parts, not the medical record, parts of the medical record are being sold and that's now become a 30, $40 billion industry. Um, but none of that value accrues back to the patient and.
You know, that's, that's where I think we're, we're a little out a kilter here. I think people are upset finding out that their data's being sold and none of the value accrues to them. None of the financial value accrues to them, but none of the, you know, intrinsic value of knowing I'm helping in this study.
So I, I would really like to see. Um, well, let's go to an area where we, where we live. So at the health system level. So the medical record, um, you know, we've talked to a lot of, uh, a lot of people about, you know, should the medical record be owned by the patient and there's a general consensus that it should, but it's, but it's not really.
Um, and I, I think we'll know it is at the point where the. Patient has access to the entire medical record, every note, every, every piece of information that's on it, they become the carrier of it, not the h i e. So they're the ones who are going from physician to physician or system to system, and they're granting access to those systems.
They're essentially going, Hey, you're gonna, you're gonna care for me for the next year. I'm gonna give you access to my medical record for the next year so that you can be a part of it. And they share it. Your system takes it in. You care for them for the year and add to that medical record. But it all goes back to them.
And then they, they control it. They sit there and go, Hey, you know what? I wanna participate in this cancer study. And they go, I'll, I'll share my information over there. I get, you know, $25 for that and I'll share it with this heart study. That's another $25. And people can then start. To feel good about contributing back to society, but also getting financially compensated.
Do you, do you think a model like that is, is possible, or how far away do you think a model like that would be? I think the devil's in the details on something like that. I, philosophically, I agree with it and I, I love the idea of the patient really owning, you know, their record. Not just so from the ownership perspective, but what, what it would signal to me is they're more engaged in exactly what's happening.
I. And one of the big challenges we have in medicine, of course, is having doctors communicate with doctors effectively. And if the patient's more engaged in that process, I think they can encourage that, that communication. Uh, but the devil is in the details. So let's say the patient, uh, owns the record and has access to the record and can even edit the record, I.
Ooh, yeah. That, yes, that becomes a challenge. Yeah. And, and if you look at the, the major vendors right now, they're providing that capacity within the patient portals. So then what do you do with that information? That I think one of the things that docs, um, you know, count on as they interact with a patient's chart is that the other person writing it, that person is accountable.
For what they wrote a hundred percent. And I can hold them accountable and I actually know that guy because we go to lunch together or whatever. And so you have this, this kind of a relationship. Uh, if the patient can edit the record, I think the, the level of confidence in what's in there will be changed.
Right. But there's, there's a way to handle that. I, I, I think the simple way to handle that is to, uh, document the source, the source of the data, right? So, uh, and, and we had to do this within our systems, especially our h i e and other things 'cause. To be honest, I mean, doctors don't trust other doctors in some cases.
So they wanted to know what doctor put that information in and then they would look at it and go, ah, I, I don't trust his data anyway. And so they can then make their, their determination based on that. But we could do that with patient data as well. 'cause we did that with demographic data. Yeah. Uh, You know, if, if they say, Hey, I have a new address, we say, you know, patient entered data new address, which triggered in our, our processes.
What it triggered was a verification phone call, uh, to that person to say, Hey, we noticed you updated this information. Is it accurate? They say, yes. Once it gets verified, then it becomes part of our, uh, you know, record, uh, system of record. Yeah, that makes that, that makes perfect sense. And the other way to do it is to have the patient be able to add, but not delete.
So if you can add, um, and you know who it's coming from, you can see what that is and, but you can't delete something that was there before, that might help as well. Yeah, it's interesting. And you know, one of the, one of the arguments that has been given to me is that, um, we're really trying to protect people from
So I you're, you're a physician, so you're gonna understand 99% of what's in your medical record. I'm only gonna understand 60% because I'm in the industry. But there's some people, uh, who are only gonna understand 30 or 20% of what's in their medical record, and they could, um, They could make poor decisions based on data that they don't understand.
Uh, how do we, I mean, what's your response to that? How do you, how do you address that? And is that just, you know, part of the, that's just part of the, you know, we, we give people control of cars. We let 'em become parents. We let 'em, you know, there's, there's a whole bunch of things that we, we, we give people control of because that's the right thing to do.
And then it's really up to them to, to make, uh, Good decisions from it. I, I think, yeah, I think there are things we can do to, to smooth that out. So for example, have you heard of the OpenNotes collaboration? Yep, absolutely. So, you know, OpenNotes is a, is a terrific philosophy and it's, um, been, you know, um, I'd say six, seven years now has been kind of rolling out across the country.
Some hospital system been more, um, Amenable to adopting it than others. Uh, we're kind of working our way towards it right now, but in general, it allows the patient to immediately see what the note said. And you know, the docs who are participating in that, they have to change their vernacular. You know, they used to, you know, I would say I got a 25 year old, you know, white male.
Uh, s o b , you know, an s o b in my world is shortness of breath, right? You know, but if I'm not medical and I read that, I'm gonna be pretty upset. So the docs need to adjust their vernacular, recognizing the patients are gonna read it for the first time. But how amazing would it be for the the patient to be able to look at their note and clearly see the plan for each issue they're dealing with, and have that as a record.
That's huge. Yeah. You know, and if I read that, you know, white male, s o b, I would say, how does this physician know me so well? Anyway, , I'll kick. Let's, let's go to the next story. I'll kick it to you. You could set it up for us. Uh, okay. So I wanted to talk a little bit about the Berkshire Hathaway, uh, JP Morgan Chase, um, Amazon approach.
And I, as you have said, stated already, I would love to have a specific short name for this entity that they're, they're working on. Uh, but essentially they decided to get together and work on healthcare as it relates to their employed individuals. And so that's about 1.1 million individuals. Um, but what I wanted to point out was after they, um, after they selected a tool Gowan to be their c e o and to kinda lead this area, that really gave me a lot more confidence in what they were doing.
Not only because. Is highly respected, legitimate, authentic, comes up with long-term practical solutions to complex issues. Uh, but because we have a point of responsibility and accountability in their process, um, before we had three guys kind of at the helm, And I think about it like when I send an email to three people and I say, can somebody take care of this for me?
Chances are nothing happens , right? But if I ask one person to do it and maybe cc the other two, uh, it can actually have a chance of happening. So, uh, I thought that was a great step forward when that happened. But most recently listening to Jamie Dimon talk about their process. Has also, uh, given me a little more confidence in the direction they're going.
He referenced, um, Bezos, uh, as it relates to Amazon and how he spent 10 years focusing on books, recognizing that it was gonna be the everything store at some point, but really getting books right. And that's, that's the model that they want to use moving forward, which I respect because this, uh, this topic's hard.
Yep. And there's a lot of complexity to it. Listening to him talk, he's already getting the vernacular down as it relates to, you know, this issue they're trying to solve. So he is talking about everything from, you know, drug prices to supply chain, to, um, irrational variants, uh, in the, the way we deliver care.
These are terms that people don't usually use. Unless they've really been, um, involved in these discussions. So that kind of got me a little more excited about the process they're going through. Yep. You know, I, I think the expectations are so high for this thing that the leaders are walking it back a little bit just to say, Hey, you know, we're gonna , we're gonna take baby steps here.
We're gonna, we're not gonna solve healthcare, if you will, uh, overnight. There's, there's, you know, we're gonna, we're gonna address this. Um, you know, this is the second week in a row. I'm gonna be giving, uh, this gentleman a little bit of credit and, um, Uh, Dale Sanders, president of Technology for Health Catalyst posted a uh, uh, Saying about 10 hours ago on, uh, on LinkedIn, and it's, uh, it's a picture of Jeff ml and it says GE was spending 3 billion a year on healthcare costs more than their annual revenue of their healthcare line of business.
And, uh, and that gives you some idea of the, of the challenge and the problem and why these employers have decided to step in. Um, here's what I, uh, There's, there's another article on, uh, apple, right? So we have, we have competing, I mean, models. So you have, you have J p m, Berkshire, Amazon, you have, uh, apple.
Intel's had to model for, uh, quite a number of years, and there's, there's, you know, other employers, I'm sure Nike has one probably in, in, up in Oregon and others. But I, I, I'd like to talk through this, this, uh, the Apple one because there hasn't been a lot written about it. And, and two days ago there's an article on it.
So Apple's first hires, uh, for their its health clinics show how it's, uh, thinking different about healthcare. Okay. So Apple's Wellness Clinic, AC Wellness, has made quick work of hiring more than 40 people to provide concierge health and wellness services at its Bay Area employee, uh, for its Bay Area employees.
According to a LinkedIn search, apple is far from the only employer starting it. We've talked about that, but Apple's approach stands out in its focus on care as not just treating disease on its website, AC wellness details, how it's looking for candidates to join the group who have, uh, experience in patient experience, and a passion for wellness and population health.
Here's a couple of examples. Uh, most of the team hires so far aren't doctors. In fact, it, the hires skewed towards wellness professionals, nutritionists, exercise specialists, and nurse practitioners. A lot of the hires have background in alternative or functional medicine, and there's even a wellness lead.
Uh, Jennifer Gibson, a former head of coaching at Vita Health. Healthx Startup Gibson, according to her profile, is passionate about things like nutrition, stress management, smoking cessation, which aren't always offered as primary, uh, at primary care practices. The company has also brought in at least a half dozen care navigators who don't have medical degrees, but do have a background in directing patients to the most appropriate care.
In some cases that might, uh, involve follow-up conversations with a specialist or lifestyle change that might alleviate the problem on its own. And that can help reduce costs. The MDs IT has hired, uh, are like former Stanford community physician. Darren, uh, Phelan, uh, emphasized on their LinkedIn profiles that their vision for AC wellness involves putting care back into healthcare.
Another former crossover doctor, uh, has a distinctive focus on sleep, which is uncommon in primary care clinics. And, um, you know, many of the new hires have, uh, you know, worked in healthcare startups and other things, and, um, So as you look at this, they're really moving away from this. Their, their mindset isn't fee for service like a traditional health system.
Their mindset is how do we keep these people healthy? And we all know that this is, you know, this is the future of medicine. We've known that if we can keep people out of the hospital, that's a much better model. Um, so I guess my question, On this as we sort of compare, maybe not comparing, contrast, these two models, let's compare and contrast these two models with what's going on in the traditional health system.
How is this, how are these, the emergence of these models covering millions of patients going to change traditional healthcare organizations, and what should our traditional healthcare . Organization response be to these, these kinds of programs, do you think? So? You know, hi. Historically we've had just an awful model.
Our off, our awful model consisted of, um, you know, patients come into the hospital, um, for, uh, appendicitis and they get a folic catheter placed, which gives them an infection. Um, we then get to diagnose the infection and get paid for that. Then we get to treat them for it, and if they get really sick, they go to the I C U and get to get paid for that as well.
All stuff we've caused. , right? And so, um, I think everybody's in agreement that that model has to go away. I think that, again, the devil's in the details on how it's, um, executed. If you think about these primary care setups that you're describing, it has to be under the tent, under the umbrella of a financial model that pencils out because there's all kinds of great things we can do, some of which have been proven.
Others of which have not been proven. There's a lot of kind of back and forth literature on care management, for example. Right. And I would say that, uh, in order for it to be truly successful, it's got a first start out under a financial model that rewards doing good. In other words, if I have finite resources and I can work on 10 things, uh, I need to pick the 10 things that I know will not only.
Uh, not only help the patient, but also we'll pencil out at the end so this whole process can move forward. So in our system, for example, as we think through how to implement population health, we're carefully identifying areas of opportunity that put us in that frame of reference where we're doing good for the patient.
Perhaps decreasing imaging, for example, low back imaging, m r i, for example. But again, you wanna do that under a model. Where it's gonna reward that goodness. And I think the danger on a lot of these, um, systems that I'm seeing is that, um, they have a lot of great ideas going in, but at the end of the day, it's gotta funnel into making it financially move forward.
And so, you know, going back to that GE example, it, it pencils out for employers 'cause they're spending $3 billion on healthcare costs. Well, it only, it only pencils out if all the things they're doing actually, um, do good. And I think there's a lot of great ideas about sleep and, and, and cooking and all kinds of great ideas of what to do.
We haven't fully proven each one of them, and that's, that's one of the challenges. Yeah, so it's, it's, uh, it's evidence-based medicine, uh, needs also to be applied to these things. It's not just entrepreneurial Silicon Valley types who are, uh, coming up with new models and saying, Hey, you know, we're gonna.
It's not the trial and error that we're used to in Silicon Valley. It needs to be right. Uh, it needs to be, uh, supported by rigorous, uh, clinical practices. Is what you're, I think what I hear you saying is that, is that right? That's exactly what I'm saying. And, and without that, it's uh, frankly, destined to fail because you're gonna end up spending your.
If you look at your expenses, they're just gonna be through the roof. These may not be physicians, but they're not cheap. And, uh, you know, to have them do a bunch of things that ultimately don't decrease the cost of the spend, um, is, you know, really challenging financially. I. And I agree with you, and both of these programs are actually under the, uh, guidance.
So you, you talked about, uh, Dr. Atul Gwane and, um, I forget the gentleman's name, but there's a physician from Stanford, uh, who's heading up the Apple, uh, model as well. So, I mean, these are, these are trusted physicians and I, and I'm sure their background is such, so I, you know, I, I'm optimistic. I, I know that you're optimistic and others are optimistic to see, you know, let's.
Let, let, let's, you know, try some new things. Let's see, you know, if we can't, uh, move the needle forward and, uh, uh, make things better. So I, I'm gonna transition to the soundbite section. We're gonna talk a a little bit about analytics. During this section, I just toss out questions, one to three minute answers, which is in some of the topics we're gonna be talking about.
It might be hard, it's more of a guideline than a rule. If you feel like, uh, you're not done, just keep going. I'm not gonna cut you off. I'm pulling this information primarily from a health system, uh, cio.com article that you wrote, the human element of health it, uh, which is a great article, and we're gonna, we're gonna really focus on data going around this.
So you guys, at, at, uh, at your health system, you started at data-driven, uh, strategy, but it hit a roadblock. Can you give us, can you sort of set this up for us? Give us a little background on the data-driven strategy and the roadblock that you hit. Yeah, we, uh, you know, when we went live with our, um, enterprise E H R, we followed all the recommendations that we were told.
Um, we built a small reporting team, uh, with a supervisor and we pushed forward and launched and, uh, you know, we expected magic to happen. We really expected to be able to push some buttons and get information out in a way that was really actionable and operational. Um, and what we found, um, about a year after we went live, the honeymoon period was over the folks who had been expecting and waiting for great information to make better decisions, recognized that nothing was changing and they weren't really getting what they needed.
And when they asked for it, the time it took to actually get what they asked for was ridiculous. Right. And so it, you know, started a conversation within the organization at the highest levels. Finally, our chief strategy officer, our chief financial Officer, and our, our C I O Marquettes, um, got together and they talked through what this might look like, and eventually asked me to look into it.
I pulled together, um, a, a very small team, uh, myself and our, um, bi developer supervisor at the time. And we brainstormed quite a bit. And I'll tell you, when I first started to unpack this issue, I was a little bit, uh, cataplexy about it because it's so complex. There's so many layers to it, and everybody had a different opinion of, you know, why it wasn't working.
And so ultimately we decided to go down three main roads to evaluate this. The first was to look internally. I. Really understand the need and the ask. So we met with every C-suite and VP at our institution for a half an hour, and I did that by design a half an hour because I wanted it to be a short period of time where they knew they had 30 minutes to tell me what's on their plate and what they need and they need to spit it out.
So we took 29 pages of notes from that, compile that into themes, and that was one piece of our evaluation. The second was we looked external at places we thought were doing this pretty good. My brother works at Intermountain Healthcare. Got a few connections there, spent some time with them to understand their model.
Spent some time with some academic centers who have a lot more resources than a typical community, uh, medical system. Um, spent time with four or five total. Um, compiled a bunch of notes from that. And then lastly, we looked at industry best practice, and we didn't limit it to our industry. We looked at Lyft and a few other places and spent time actually talking to folks in other, other companies, compiled all that information ultimately into what ended up being a ridiculously long report.
Although it had a pretty brief executive summary, I. Um, and I ended up presenting that to our executive team at our quarterly meeting, basically for approval of the recommendations that were coming forth. It was a two, sorry, go ahead. So can you give us an idea of some of the recommendations that were made?
I mean, so you, you, you do a, a, a complete analysis of analytics within your organization. You talk to the executives, you gather that information. Um, Can you give us an idea of, uh, of, you know, what the executive summary to that might be? Mm-hmm. , uh, to the group? Yeah. So part of it had to do with we were just understaffed and I, I, I always, you know, hate the idea of adding staff if we don't have to.
It's a, it's a big topic, but the truth of the matter is we didn't have the right number of staff to make this happen, but more importantly, we were missing key. Staff functions that weren't there. And part of the problem was a disconnect, frankly, between what the requester wanted to communicate and what the BI developer heard and then built.
And so that the delta between that communication, um, is really what led to a lot of inefficiencies in our system. So what we did was we took a look at that and we said, how can we best improve on that? And one of the positions that we added was this clinical data analyst. Which is really somebody who can really understand well what's being asked.
This particular, uh, hire, who's awesome, uh, had a background as a an I C U RN and then was an Epic analyst for several years, really understood things well and then learned reporting, so we understood the clinical side. He had some operational background as well, and then now he understands the reporting side.
So those, those conversations now are much cleaner, much clearer, much more, um, Defined out of the gate so we understand the ask a lot better so you're not pump out 30 iterations of a report. The se, the second big position that was kind of different was our principal trainer position, which was, you know, this is a brand new position, not usually something you find in analytics, but what we recognized is that folks weren't able to really fully comprehend the information we were delivering to them.
So the goal with this position was twofold. Number one is to really spend time with the folks who are requesting information and getting reports and dashboards, et cetera, so they really understand what's in front of them so they can take action on them. But the second piece, which is probably even more important, is we also recognize we couldn't continue to scale and just add positions.
We had to add self-service. As one of our, one of our, uh, strategy frameworks moving forward. And so this person would be in charge of teaching folks how to leverage self-service to meet part of their reporting needs. So that was, that was a big piece of it. The other big piece was data governance, and we can talk more about that in a bit.
But that, that whole data governance piece was huge. And when I built the re the initial, um, uh, presentation to the executive team, I actually didn't dive terribly deep on that topic. What I did was I focused on the why, and then what I wanted to do was spend dedicated time on data governance in a separate forum.
And fortunately, it, it worked out that way. Yeah. So I, I do wanna talk about governance. I also wanna talk about, uh, scale, uh, and, and, and a little bit more about the role. So let's, let's go to governance. So, Um, you know, you, you emphasized the human element of, as the title of your article indicated, the, the first first group that you really focused in on was the clinicians.
Um, how do you, how do I ideas for, talk to us about the process. So somebody has an idea for a, a new, uh, set of data they wanna take a look at and some new information that would really help a process. How does it, how does it go from. Uh, concept to, uh, to at development, to, uh, operationalizing it and maintaining it.
Give us an idea of the life cycle of that, that whole process. Well, I should start by saying that we're, um, you know, fractionally through our, our, uh, execution of putting this in place. So, um, we're not fully there yet, but um, as it stands right now, we spend a lot of time focusing on accountability and trying to make sure that throughout the organization we have the structure in place.
So that when we have a request that all the pieces are lined up, ready to go, and that starts out by, we have our, our, almost our entire C-suite. Is our steering committee and they meet on a quarterly basis, uh, to set a strategy around data governance. Below them, we have our data governors, and these are usually, um, VP level and director level folks who have a specific domain that correspond to their area.
And then below them we have the individual data stewards who do the actual work of understanding workflow and, and solving the data governance problems. And then below that, supporting all that, we have our office of data governance. We've built in the process where we can so far. So for example, um, we have a P M O now and, uh, in our P M O project request process, data governance is one of the key line items there.
So we have to get data governance to sign on and to sign off of projects. We've also incorporated data governance into, uh, ServiceNow. Uh, I'm not sure if you guys use ServiceNow when you're at St. Joseph's. But ServiceNow is basically a way to track the work that we do. Um, and having that built in on those levels allows us to be able to get folks to utilize those tools.
And then as a flow of that, it comes into the workflow for data governance. So for example, let's say, um, let's say you're using a report and you identify there's a data quality issue. You look at it and you're like, oh, these numbers are way out. This, you know, this can't be right. Or you're looking at something even better.
And you know, this thing says length of stay. How are they defining length of stay? What does that look like? They can actually log a ticket in ServiceNow that day. It pops into our queue within the data governance, uh, queue, and then it gets worked by the appropriate data steward that corresponds to that domain.
If that makes sense. Yeah. Did you have a, did you have a big cleanup project? I know that when we, we did our E H R uh, implementation. I remember the day when somebody came to me and said, look, we have, we have 5,000 reports in, in, uh, in Orange County that we need to, uh, you know, reconcile and, and whatever.
And I'm like, 5,000 reports. I'm like, well, let's, let's start with the basics. Like, when's the last time they've been accessed? And a good couple thousand of them haven't been accessed in. Yeah. Um, did you have that same kind of experience in cleanup? We did. We, we did that, uh, that type of cleanup, both for, uh, traditional, uh, SQL based reports as well as for our dashboards and our reporting workbench reports as well.
How do you, so how do you scale this? So you have a, a clinical data analyst and you have a, uh, uh, a new training, uh, principal, trainer for analytics. Mm-hmm. , uh, that's, that sounds like two people to me. Um, so it would seem to me that they would either be extremely busy, um, or that they figured out a way to, to scale the work that they're doing, uh, across the organization.
Uh, is, is there, is there a trick or is there something about scaling that you guys have figured out? We also added a few bi developers, traditional BI developers, um, to that mix, one of which we, uh, have dedicated to registries, which, uh, might sound a little odd, uh, but with the, with the work that's happening now around our a c L model and our, um, our, basically our ambulatory contracts that we have, having somebody dedicated to registries is key because most of this stuff feeds off of registries.
So getting those right people in the right positions was important, but the single biggest improvement in our ability to deliver has been our work around lean and agile. So about two years ago, we started going down that road. Um, my manager, Michael Olsson has, uh, has really been at the tip of the spear on this, and we've completely revamped our approach based on agile and lean methodology with three week sprints.
Wow. That's, uh, well, I, I love, I love it when we close out a show by introducing a topic that we could do a whole show on. So, um, we'll, we'll, we'll definitely have to have you back and talk more about that. 'cause that, that sounds like a, a, you know, a lot of people talk about. Lean and a lot of people talk about agile and, um, I, I think they're at least agile for, from where I sit is one of the least understood terms within IT organizations and it gets, it gets kind of convoluted.
So, um, Lee, I wanna thank you for coming on the show. What's the, what's the best way for people to follow you? Uh, so they can follow me on Twitter. It's uh, Lee md it, I'm not sure that's the best name, but that's the one I got going right now. And then, uh, of course I'm on LinkedIn, uh, and, uh, they can always email me with any questions they have, um, at Lee firstname.lastname@example.org.
Wow. Sharing your email address. That's bold. I, uh, I appreciate that. Um, I. Also, you can follow me on Twitter at the patient cio, uh, my Writing Health on the health lyrics website. Uh, you can follow the show at this week in h i t and check out our, uh, new website. Uh, actually it's not new. I, I, I have to change this.
It's, you could check out our website this week in health it.com. Uh, catch all the videos on the YouTube channel. I think we're up. We'll produce another five or six from this show and put 'em out there for, uh, social media. So please come back every Friday for more news information and commentary from industry influencers.
That's all for now.