October 3, 2023: Join us as Clearsense’s Data Governance Principal Advisor Terri Mikol illuminates the evolution of analytics by drawing parallels to the telephone switchboard operator, revealing pathways for democratizing data use across healthcare organizations. Discussing unique illustrations such as the employee at UPMC defining ‘patient name’, Terri will explore the potential of every employee to contribute to data quality and the implications for organizational learning and efficiency. From exploring ‘divide and conquer’ methodologies to analyzing the lessons from cell phone adaptation, we’ll delve into how these strategies can revolutionize data handling and promote widespread competency in analytics. Terri will address the persistent demand and supply imbalance in the field, provide insights into incorporating data roles across functions, and discuss the alignment of analytics with organizational needs. Reflecting on shifts in organizational culture and the statement "data is a part of everybody's job", the conversation will explore the balancing act between accessibility and technical demands, the breakdown of knowledge silos, and the decentralization of data responsibilities.
October 5th at 1 PM ET/10 AM PT, discussing challenges in healthcare interoperability. We'll tackle key issues like fragmented technology systems, data privacy, and cost-effectiveness. Engage with top-tier experts to understand the current landscape of healthcare IT, learn data-driven strategies for patient-centered care, and discover best practices for ensuring system security and stakeholder trust. Register Here
This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.
Today on This Week Health.
Better data is everyone's job, and really, it's the only way you're going to scale. The questions we're being asked now are so much more complex and require all kinds of data knowledge. The days of one team being experts on all your data are over, and it's never coming back,
alright, here we are for a conference campaign, and I'm looking forward to this conversation. We had a webinar last month on AI. And we have a webinar this month on interoperability outcomes, a discussion of what's possible. That's on October 5th, by the way, 1 p. m. Eastern time, 10 a.
m. Pacific time. You can sign up on our website. We've got great guests. We've got Aneesh Chopra. We've got Mickey Tripathi and Mariann Yeager, and we're going to talk interoperability. But as we're doing that, it's become apparent how critical data is. Data is to feed. Just about everything within the organization and data is significant, obviously, for improving outcomes, improving the experience and whatnot.
And so, I thought today we would reach out to our friends at Clearsense. And we have Terri Miko l, who is a principal advisor for Clearsense and spent 18 years at UPMC. on their enterprise analytics on the provider side. So extensive experience in this, and we're just going to talk analytics.
Terri, welcome. welcome back to the show.
Thank you for having me. I'm looking forward to the conversation.
Well, I'm looking forward to it as well. What's holding us back on the data and analytics side?
So let me start by saying that I am a software engineer by trade.
So over my Decades of experience with the right hardware, software, and data, we can make pretty much anything happen. We've created thousands and thousands of excellent analytic solutions and are really driving a lot of new insights. However, there, there remains a challenge that was a challenge 35 years ago when I first started in this space and is still very much a challenge, and that is that the demand for data and analytics continues to far exceed the supply of people that we have who can actually work in this space.
So the challenge is this asset that we call data has become so strategic and we still have a small portion of people in the healthcare industry that can actually work with it. This is what I believe. We need to solve next because the technology will continue to get better. The data will continue to get better, but until we enable more people to work effectively with it, I just don't think we're going to get over the next barrier and really begin to scale.
Yeah. It's interesting as a CIO every year, and literally it was every year. The team would come in to me and they would say. Hey, we need 10 more people on our data and analytics team. And I'm like, well, what's going on. And it was this whole demand aspect of it. But when we looked at it, we had countless reports that were never being used anymore and all that other stuff.
So that's the low hanging fruit. That's the, you just look at that and you go, all right. We're just going to stop creating things that no one uses. But we're hearing all this stuff about AI. We're hearing Epic is coming out with generative AI front end to slicer dicer.
Like, oh, all of a sudden this is going to solve everything. Is it though? I mean, we still have to know what the definitions are, what the metadata is, all the people still have to understand what they're looking at in order for that to be effective. Yeah,
we focus a lot on the fact that we still can't provide answers to the common questions everybody has.
What does it mean? Where did it come from? Is it accurate? We're still struggling there. So what we've done at Clearsense is we've learned a lot from our history in this space and I want to share a couple of stories with you to really explain why this approach that we're taking to solving this supply and demand imbalance is really taking off and making a big difference out there.
So let's start with a long time ago when the telephone was first created and... Switchboard operators became very critical to the world because everyone started making phone calls and the data scientists back then assessed the volume increase and determined that eventually everyone in the world would have to become a telephone switchboard operator because the call volumes were, expanding so much.
Well, what has happened? Here we are, all these years later and each and every one of us is our own telephone operator, right? So that shows how you scale. You scale when everybody gets involved. The second thing that history has taught us is that people can learn new things that are outside of their normal focus.
And I'll use the cell phone as the example. I mean, could we have imagined 20 years ago that the average person would be able to get on a cell phone and do all of these things? There was no training class, really. There was no... Heavy, intensive, major culture shift that we all drove. It was just something that was needed, something that was important.
We put it in people's hands, made it as user friendly as we could, and look at what's happened. So, so those two stories really been the foundation of how we approach analytics and making data something that every single person in an organization can work with. Because we believe if we don't do that, we're just still never going to meet the demand.
So, that's the gist of what I wanted to talk about today, and we can dive into a lot of details on how we've actually done this, because you hear all the, oh, we're too busy, we don't have time, we don't have money, it's too much of a lift and I believe that, again, history has shown us ways to completely transform your entire workforce without it being So disruptive and so dramatic that people are overwhelmed by it.
So over the next 10 minutes, you're going to explain to us how we're going to go from the switchboard operator that, takes the call and does whatever that thing is, anyone who's watching on video can see me now, taking the cord and moving it from one to the other. But essentially that whole paradigm, which is what we have in analytics today, how it starts today, is somebody has a need at the, wherever, a nurse or whatever at the bedside, they talk to clinical informatics then goes.
to the data team and says, Hey, where is this data? Is this data available? And then we essentially, we rebuild it every time. We rebuild those connections every time. So , in the next 10 minutes, you're going to explain to me how we're going to go from that to Zoom, essentially.
Yes, and we're seeing this working already. So it's a combination of, as you always hear, people, process, and some technology. I'm going to start with the people side. I want you to think about human resources and what human resources looked like decades ago when it was just starting up. Employees weren't yet viewed as a critical asset to an organization, but once we changed our minds and realized that employees are important, you start to see new processes and new roles and new technology.
And the most important change that we saw was that HR started showing up. in every job description. And how did they do that? Because everyone already had full time jobs, and everyone was already busy. Experience has shown us that when an asset becomes that strategic, it starts showing up in every job description, and every person plays a small role in The management of that asset to make sure that it is of high quality and that it is being used efficiently and appropriately.
I would argue that we've experienced the same exact things in finance. And facilities and supply chain with all of our equipment when things become important, you start seeing it in everyone's job description. So we are using a divide and conquer methodology on the people side and getting engaging your entire workforce in a small data role.
And they come in all shapes and sizes, you've got business folks who are experts on data collection and what data means out there. in the primary functions that a business performs. You've got IT people that are experts on your data sources and then you have analytics folks that are experts on all your reporting.
So there's a lot of ways that you can engage people and take just a piece of everyone and begin to scale. And I'm going to give you one perfect example. At UPMC, we gave someone who worked in the registration area in admissions for a long time. We gave her the term, the business term, patient name. Go define patient name in our business glossary for the entire organization.
And she already had a full time job, but in her own time, she made a call down to the emergency room and learned about how they handle patients that come in with no identity. Do we use Jane Doe, John Doe, is there some numbering scheme, because we need to document that somewhere so that the secondary users of data know that when they see Jane Doe and John Doe, perhaps that's not their real name.
Then she spent some time up in the maternity ward and had the same conversation about new babies that aren't named yet. Do we use baby boy? Baby girl? And then she talked to it about all the testing we do in production systems with fake people. What do we call them? Is it Mickey Mouse? Is it Michael Jackson?
Is it test 101 or what's the main vision?? So long story short, she gathered all of this knowledge in her own time, put it in the business glossary. And now everyone in the organization is benefiting from her knowledge. So instead of us all starting over every time with patient name, we put one person, we put it in her job description.
You're going to own patient name and you're going to keep documentation on how it's used across the organization. She grew, she gained data skills. We all benefited from what she learned.
So I heard two things there. One is data is a part of everybody's job. Therefore, we're going to empower them to be, whatever aspect.
Data stewards, data miners, data, archaeologists, essentially is what she was. She was, looking at what we have done over the past years. That's one aspect of it. I think that's really fascinating. But the second you talked about was a business glossary. Like, there was a place to put this information so that others could have access to it.
We fail there an awful lot, don't we?
We do. So that's the second piece now. We're going to talk about some of the tech, right? So, in my experience, all of the tech we have in this space that's designed to teach us about our data and make it easier to find things, is essentially... Google for your internal organizational data.
They're all very very technical and they require training and extensive consistent use in order to get really good with the tools. That's just not going to fly because we need people to be able to spend an hour a week on their data role, or. They're not going to be in there all the time, so it has to be simple, and it has to be the user interfaces that they're used to elsewhere, like on their cell phones, or in gaming, or when they're shopping online.
We have to use things that they already understand and know how to navigate so that there isn't any training. In addition to that, I'll use the word transparency the breaking down of knowledge silos. We in healthcare still very much run on the phone a friend world in the analytic space. You get a new data question and if you've never encountered that before, you've gotta engage other people and find out who knows what and get to the right place because we don't have this, Complete transparency where is all the data?
What all does it mean? Where are all our reports and dashboards? What are the known data quality issues that we need to be aware of and work around? All of that knowledge that exists out there, we need to put it in technology that is easy to reference, easy to use. Everyone needs to be able to contribute their knowledge.
Everyone needs to be able to get it back out. The user interfaces are very critical there, as you can imagine.
that's really interesting. So you said people, process, and technology. You're touching on technology a little bit, but we jumped over process. What is, what does the process look
So I, I love that because whenever you look at any kind of organizational models for managing your data assets, they're highly complex and. What we've done is we use the same organization that we've had experience with in human resources and finance and supply chain. In that, You have a C suite VP level.
They're up there making decisions about what we're going to do, how fast we're going to do it. They're aligning resources, both people and money. They're setting and approving policies. So they're making the big decisions. Then we have this director level. And in the data world, we call them information governors or data owners, something like that.
And that is where the vast majority of the data related decision rights and accountabilities lie. That's where the expertise is, and so we divide up all of your data pretty much in line with the director level in your organization, and we put these individuals in charge and give them formal decision rights over data.
And then we have a analyst manager level, that's your stewardship community, and we engage hundreds. in that space. We've got data stewards out in the business that are experts on data collection and primary use. We've got analytic stewards who are experts on all our reporting and dashboards and data science stuff.
And then we've got our data source application stewards that are experts on the data that's inside of your system. So we've taken the traditional data steward role and we've split it up. So that we can engage hundreds of people and elevate hundreds of people. We don't want to go out and hire full time data stewards because then they just become middlemen and you're going to want to hire new FTEs anyway.
It's better to just retool, the people that you already have and give them a skill that they so desperately need and will really benefit from. How many executives have you met? I've stood up in front of a room full of people and presented a dashboard and been terrified that someone was going to ask them a question about the data because they think it's misrepresented or wrong and they wouldn't be able to answer the question.
Yes, I've been in those rooms, you've been in those rooms where we're sitting there going, I know that data's wrong. Like, that's just wrong. Like, and you're just sort of scratching your head. The thing I like about the scale of it is that before that data even gets to the meeting, it has been sort of vetted.
You have more people sort of going over that data and going over the definitions and contributing. to the body of knowledge and moving forward. I find it interesting that we're mostly talking about a cultural change and an approach rather than a technology per se. We're essentially saying we've got to change how we approach data just the whole data life cycle within an organization and the culture of how we approach data.
Data used to be this, this asset behind this glass wall that we'd have to go talk to somebody. They'd go behind the glass wall, get the data for us and then bring it back out. And now what we're saying is, nope, in order to scale we're going to empower the entire organization. Well, not the entire, but a significant portion of the organization as part of their job is to contribute to the value.
of the way we use data and the way we define data and explore data. It, am I capturing this well? You
are and we've done this before. This is not new to us. We've done it in HR. We've done it in finance. We've done it in supply chain. We've done the exact same thing for our other strategic assets.
So when you approach it that way, From the top down and the bottom up, people weren't thinking, Oh, this is some new big thing. This is going to be hard. We don't know what we're doing. It really is just asset management and we've done it before.
We have. So HR is an interesting example because now people have implemented a system.
And it used to be, we had a fairly sizable HR department. We'd call them up and say, Hey, I've got this benefits question. I've got this question. I got this question. You have this question. And a significant portion of that now is done, like we just query a system and it gives us the information that we need.
I mean, there's obviously there's still... Transparency, but
there's transparency there, right?
Right, right. And there's not a you also have done away with the culture of... Job security. I'm going to hold on to the data so that I can hold on to my job. As an HR person, you recognize, you know what, there's enough work here for HR to do.
there's tons of other things to do. We don't have to be the data stewards and the data stewards is a nice word the data gatekeepers on all this stuff. Let's have everybody have access to it because you know what, they're going to have that question. at seven o'clock at night when they're sitting down with their family.
you're right, and interesting that you bring up the whole job security because you're talking to a girl who was the data expert. That's how I, that's my career, right? But here's what I learned. Statistically, people who are in charge of analytics in healthcare don't last very long.
And it's because they're not meeting the demand. The only way I succeeded in lasting as long as I did was because I enabled other people. I focused on enabling other people to go and explore and become heroes on their own. It was not about one centralized team solving the world's problems. It was about giving everybody the ability to do it.
All right, so
we're going to break this down. You are now talking to yourself from, 10 years ago and other people who are in that role, as well as people who work in that role, potentially some nurses and whatnot who are saying, okay, I'd like to champion this kind of approach in our system. Let's break it down, make it , as, simple as possible.
What's like the next three things we're going to do?
So where we start is right at the top, because we want to get the C suite and your entire VP level to understand that this is just asset management. So we walk leadership through. We're talking about inventorying things, assigning people responsibilities, managing their quality, becoming transparent about where it all is, these basic things.
So we start at the top and make sure that we have support from the top. Yes, data is an asset. We're going to move forward as an organization and use what we've learned. before and apply that to data. So we have a very informal session that we take leadership through and we describe what they've done in finance and HR and supply chain and show them how to transform that.
So that's number one, very important. Number two thing that you do is inventory. The biggest win that you will have in an organization is providing complete transparency to where is all our data. Most organizations have application inventories, but they have more than one. They have many of them, and they're not updated, and they're not in sync.
So simple as assigning a new role. You're an application steward, and these are your five applications. You're responsible for publishing metadata about those applications. That's your job now, and it's going to take you an hour a week. And you're not going to do it once a year. You're going to do it as part of production change control from now on.
You're always going to keep this metadata updated. Such an easy change, right? But instantly inventorying where all your data is. Then we inventory all the people. Where are all the people in the organization who work in the analytics space? Part time, full time, whatever. There are hundreds. There are hundreds out there doing this work, and we don't even know who they are.
So we find them, we profile them, we publish their bios for the organization to see here's what we do for the organization. Here's the data we use. Here's the solutions we provide. So it's all about inventorying the assets first. And you're not
talking, you're not talking about re org ing. You're just talking about...
Inventory! Yeah, just who are you? Why would we want to make any changes whatsoever until we know what's going on? And then we start inventorying the data quality issues as well. Why not provide transparency into those things before we start talking about, oh, we need to change our data quality strategy.
Why don't we go out and actually Find out what we have in place already. Who's doing what? Maybe track all our issues centrally so we can start to figure out, can we do some things better, right? It's just, there's some basic asset management things you can do to get started. So that's where I would start.
It's interesting. The last project that I was working on at St. Joe's before we announced the merger with Providence, which takes your data problems. Exponential, because we were 16 hospital system and then became a 52 hospital system. And if you don't have it, if you don't have it baked at 16 hospitals, you're not going to have it baked at 52.
And you're going to have to step back and make those things work. But it was interesting. We went out and inventoried. All the people that had just data and there was like five different titles that people hid things with, like data analyst support analyst, just a bunch. It usually had the word analyst in it in some way, shape or form.
And then we
used to have informatics. It used to have informatics. Remember, that was popular for a while.
And you'd go, what we ended up doing is going into the job description and going, Oh, my gosh, this is. 100 percent data, this is 50 percent data, this is 15 percent data. And then we just went out and started talking to people, but it was important to establish that foundation of trust.
It's like, look, we're not looking to bring all these people into IT. We, we don't want to manage all these people. In fact, we want just the opposite. We don't want more telephone operators. We want more departments that are empowered with data, that understand data, and you may end up getting more resources as a result of this.
More people working with your data and making it more available. And so that foundation of trust is so important as we go into the organization. And by the way, we found... Oh my gosh, hundreds and hundreds of people that were working with data and some of which we had no, as the as the technology owners of the data silos and that kind of stuff, we had no relationship with them.
And so by establishing a relationship, all of a sudden they were exposed to, Oh my gosh, I didn't know we had that data. We didn't have these mechanisms. Oh, I didn't know you, you wanted my feedback into the data definitions. It's like,
absolutely. Breaking down the silos.
Breaking down the silos. And they now have a career path because they're working with other people who are in that space as well on some things, right? We're not saying, Change everything you do, keep doing what you're doing, but we want you to work together on some things now as a community of analysts, right?
Just a ton of opportunity there to let people share what they know. I'm not talking crowdsourcing. I'm talking formalizing the people who will share knowledge and publish it for others to use, because we want to make sure it's. Of high quality and well maintained and whatnot. But that transparency is just so critical in order to elevating everyone's ability.
then how long can people expect for a transformation like this to take?
So we like to do it one data domain at a time. And by data domain, I mean a group of data elements that we're going to assign to somebody. So like encounters, medications. Images, surgical services. Those are some examples of domains.
So we like to start off doing them one domain at a time and just hitting the big things, the things that people always ask questions about because that's a time saver for the stewardship community. Once they publish that knowledge, they start getting fewer phone calls, right? People start calling them less because they're now using their internal Google.
And imagine, once we capture all of that, And we start putting ChatGPT like stuff on top of their own internal institutional knowledge. Now, we really will have Google for your data inside of
I'm looking forward to that time, when we have that natural language front end, and it's delivering the quality data on the back end.
But you can't short circuit this process. Like you can't
get there with just a handful of people either. You see that, right? Everyone contributing.
Well, people are so excited about the natural language front end. And they're like, you can still have a natural language front end.
Ask it a question and get crap back, it's like, it's, you still have to do the work on the other side it's not like the generative AI model is going to go into the data and go, Oh, I'll make sense of it all. Here you go. No, it's going to hallucinate. It's going to give you whatever you have to make sense of it on the back end and have the metadata, have the definitions, have it.
All that work done. And I love the fact that what you're saying is, it's not a team of 10, 15 people. It's an organization of a couple thousand people in some cases, and a couple hundred thousand in other cases, who are all working in a direction, a cultural direction and a stated direction towards creating value from the data that we've collected.
Better data is everyone's job, and really, it's the only way you're going to scale. It's what we've seen over and over, and believe me, I miss the days when trending readmissions made me a data hero. I miss those days. The questions we're being asked now are so much more complex and require all kinds of data knowledge.
The days of one team being experts on all your data are over, and it's never coming back,
never coming back. That's amazing. Well, Terri, I want to thank you for your time. It's always a pleasure to talk to you. I appreciate you sharing your... Experience and wisdom with the community. Thank you very much.
Bye bye. See ya. By the way, thank you Clearsense for making Terri available and Clearsense for being a partner for several years now. We really appreciate them. If you have the opportunity we had a great webinar last month on AI.
That's going to be aired on our podcast channel. But if you hit the opportunity, we have the Interoperability Outcomes, a discussion of what's possible, October 5th, 1 p. m. Eastern Time, 10 a. m. Pacific Time. Aneesh Chopra, Mickey Tripathi, Mariann Yeager, and I will be moderating that, and we're going to talk about What are we trying to acheive? Why are we doing all these things around interoperability? What's the government really pushing towards and what can we expect in the next couple of years? So that's what we're going to be doing. And I want to thank everybody for joining us. Thanks for listening. That's all for now.