The amount of data generated continues to grow exponentially, with the volume of healthcare-related data estimated to reach more than 54 zettabytes worldwide by the year 2025. As The Big Squeeze continues, health systems are faced with a growing need to deliver excellent care more efficiently. Modernizing business intelligence (BI) capabilities and the advanced analytic tools that come with it is the pathway to efficiency.
In this episode of In Network’s podcast feature Making Rounds, Head of Thought Leadership Dr. Jerome Pagani speaks with Digital Health Practice Leader Kevin Erdal as well as Michelle Peranelli, who leads the data and analytics function at Nordic. They discuss what’s needed for successful BI modernization, the challenges of healthcare data analysis compared to other industries, a typical BI modernization journey, and some interesting health-specific use cases.
In Network's Making Rounds podcast is available on all major podcasting platforms, including Apple Podcasts, Amazon Music, Google, iHeart, Pandora, Spotify, Stitcher, and more. Search for 'In Network' and subscribe for updates on future episodes. Like what you hear? Make sure to leave a 5-star rating and write a review to help others find the podcast.
[01:22] Modernizing healthcare infrastructure with business intelligence (BI) modernization
[02:36] First steps toward BI modernization, including requirements
[05:51] Healthcare enterprise analytics
[06:44] How the healthcare industry uses data today vs. other industries
[08:14] The unique challenges of healthcare data analysis
[09:12] The cost of data wrangling in healthcare
[10:52] Using AI and automation to process data
[12:05] How health systems should think about investing in data infrastructures that support analytics
[12:52] The history of data use and analytics in healthcare
[16:14] Real-time data reporting
[17:52] A typical BI modernization journey
[21:26] Typical use cases for health systems
[25:51] User experience in healthcare
[28:49] Data as an asset
Dr. Jerome Pagani: Hey, Kevin, Michelle, thanks so much for joining us today.
Kevin Erdal: Happy to be here.
Michelle Peranelli: Thanks for having us.
Dr. Jerome Pagani: Kevin, we have a perspective around modernizing healthcare infrastructure and getting to the right architecture to support business intelligence modernization. Can you tell us a little bit about that and what's on our client's minds?
Kevin Erdal: Yeah, absolutely. So really, at the end of the day, what we want to do from a business intelligence standpoint is get data in the hands of people to help them make informed decisions. Right? So, we want to make sure that we're removing the barrier of technology to prevent some of that from happening. So, in healthcare, there's tons and tons of data, many different sources. We have to make sure that whatever architecture we're designing is also going to be able to scale over time. And that becomes critically important when we start coming across net new use cases. The other piece of the puzzle is that we have to have repeatability. So, as we start to consolidate data and we make data available for a specific use case or maybe a specific user group, we have to be prepared for those next two, three, four, five use cases that are coming down the road. And we want to make sure that we can reuse and repeat some of what we've already started to accommodate the first two or three use cases in that particular example. So really at the end of the day, it is as simple as making sure data is available to the masses and that we could use some of the business intelligence technologies to put it in the hands of the right people at the right time.
Dr. Jerome Pagani: So what are the first steps that you typically see on that journey towards BI modernization?
Kevin Erdal: Yeah, usually we see a first step of trying to figure out a particular department or maybe a particular user that is trying to get at some very specific information. And then we start getting into the nitty gritty details of what information do they have today versus what are they trying to get towards and what are the barriers preventing them from getting that information? Oftentimes we hear things like “I’m using three or four or five different independent reports,” oftentimes going to be an operational report. They don't have a good way to combine the information from those reports into a singular solution. So, in that particular example, we might start breaking down to say, well, where is this data originating from? Where is information being entered into a system, and how do we start consolidating that data to then build a singular point of truth, if you will, from a business intelligence standpoint so that user, instead of trying to use four or five different operational reports, can go to one solution and ask four or five or six different questions of the same solution, but get the answers they're looking for from that singular solution that they're comfortable with and they have been trained on and they're going to on a regular basis to get information to help drive the business or help change patient care, whatever the scenario may be that they're responsible for within a health system?
Dr. Jerome Pagani: From an infrastructure perspective, what's required and, you know, what are the alternatives as folks are sort of considering that modernization to support BI?
Kevin Erdal: Yeah, so for infrastructure, we need to be able to accommodate huge volumes of data, right? So, we're seeing a lot of our customers in today's world look to public cloud platforms to accommodate this need and start to create more of these dynamic data platforms. Or in some cases people are talking a lot about data mesh or data fabric-type of strategies to make sure that data is available to support the BI modernization effort going out within the institution. So really, any platform, it doesn't always have to be cloud, but we're seeing a vast majority of our clients look to the cloud platforms that can really scale over time and that you can spin up or spin down some of the infrastructure required to support the volumes of data, the various sources of data that a lot of our institutions are depending on a day to day basis.
Dr. Jerome Pagani: So, what else do we need besides that infrastructure and that solid use case to get started?
Kevin Erdal: Yeah, really, what you want to have is a long-term mindset and a broader strategy, right? So, oftentimes, there's going to be a singular use case that we can build an MVP or a minimal viable product around, which is great, but we also want to start to understand what is the 2 to 3 year vision. Is there a merger or an acquisition on the horizon? Are organizations starting to work with other entities outside of their organization to ingest data to support things like clinical research? Are they trying to de-identify data? Our customers in today's world are getting deep into the de-identification of their data to support things like artificial intelligence. So as we start to think of these different use cases over time, we want to make sure that whatever infrastructure we're building here and now to support those institutional objectives around business intelligence, for some of the operational needs, for example, we also want to keep in mind where we might want to go in the next 2 to 3 years so that the infrastructure and the overall strategy is supportive of the comprehensive institutional needs.
Dr. Jerome Pagani: So, when we're talking about analytics within a health enterprise, what sorts of things are we talking about?
Michelle Peranelli: Yeah, that's kind of the fun part of analytics is it really crosses everything that's going on in an organization. You know, we have all of our clinical analysis. Operationally, we use it to drive how we're managing the business day to day. And financially, it's important to understand, what's the impact of all of the items that are going on within the organization? So really, when we talk about analytics, it's, you know, let's understand what's happening today, how we can use that information to understand why things are happening, what can we do to potentially improve performance, and then really use it to drive decision making for how we can offer care in a better way.
Dr. Jerome Pagani: So, how are health systems today using data, and how does that compare to how other industries are using it?
Michelle Peranelli: It's an interesting dynamic because healthcare hasn't had digital and really electronic information for as long as other industries have. It's a lot of information across a lot of different types of data. So even just being able to have line item data versus a lot of images. Healthcare has a very interesting mix of information to take in and really address very specific issues. When we think about things like the financial sector, they've had information electronically for a long time. They've had a lot of ability to really look at what's the best way for me to use this to drive every decision that's coming across the table. So, they've been able to take advantage of, you know, not just visualizations like all of the stock tickers and logging into your account, being able to see lots of fancy graphs anytime that you want. They also are getting really into AI and using it to manage entire portfolios, trade stocks without anyone having to get in there and do it manually. And that's really where hospitals and healthcare organizations are starting to move towards. Of, we have all of this data. What are the things that we can start to do to really innovate and allow our folks to see that data within their workflow and use it to make decisions?
Dr. Jerome Pagani: Yeah. And you brought up one of the challenges that healthcare faces is that the types of data that we have to deal with are a far greater variety than we see in other industries. Does that present some unique challenges for healthcare, and how does that affect our ability to combine and interrogate those data?
Michelle Peranelli: Yeah, it's one of the biggest challenges that healthcare faces is really being able to have clean, normalized data. It is very difficult when you're working across not only systems but different clinicians, different service lines. Not everyone speaks the same language. Not everything is consistently a checkbox that then can be cleanly reported on. So there's a lot of work that goes on behind the scenes to really clean up that data, wrangle it in a way that we can make it so that it's easy to report on and actually build analytics from.
Dr. Jerome Pagani: So, Michelle, you mentioned that data wrangling is one of those sort of significant components when it comes to being ready to ask the kinds of questions that the enterprise needs to ask. How significant a lift is that for healthcare?
Kevin Erdal: Yeah. So, what we still see on a regular basis, unfortunately, is that 80% of the time is spent on data wrangling. And that's been a consistent stat now for the last couple of years, whether we're talking BI modernization or artificial intelligence or really even enterprise data warehouse, the vast majority of the time is really spent understanding what is the current state of a particular data set and then how do we curate that data to really support the desired output. And that again, that's been the story now, unfortunately, for the last several years. We just feel it a lot more right now because we're getting more sophisticated, which is the exciting piece within healthcare, right? We're asking more, and we're demanding more of our data, and frankly, we're expecting more. So that part's really exciting, but it kind of goes all the way back to the infrastructure component and creating repeatability. When we do this wrangling one time, we want to make sure that we can support multiple use cases over time because at the end of the day, you know, we aren't in finance or we're not in retail. CPT codes and ICD-9 codes and ICD-10 codes and LOINC codes and SNOMED codes don't map quite as easily as something that would be related to the cost of a particular SKU, for example. So anyway, that's where the repeatability really comes into play, and that wrangling piece is still happening in today's world. There are a lot of different services and solutions out there to help expedite, but at the end of the day, it's still a very heavy lift that we need to be able to plan for and then support long-term.
Dr. Jerome Pagani: So, I hear that part of the answer there is to get that infrastructure that allows you to have a repeatable process so it's not a net new lift every time. Will things like AI and automation end up being useful for speeding up that cleaning process any time soon?
Kevin Erdal: When done right? Yes, absolutely. It will be very useful. We need to be careful, though, on what we're automating. So there's nothing worse than automating a process that is integrating data the wrong way, for example, and I know that's a generalized statement, but there are some nice net new tools and services coming from big tech specifically right now to help with metadata management in general, or just data management long term specific to healthcare, which is not something we had seen in the last even 2 to 3 years, right? More people are putting an emphasis on what it really means to integrate healthcare and or clinical data, but also maintain and sustain over time. Again, I kind of go back to where if we do a project in 2021 and we do a lot of this data wrangling, we don't want to have to do that again in 2023 to support a similar or possibly a net new initiative that just happened to be using some of the same data.
Dr. Jerome Pagani: So, how should health systems be thinking about how to invest in data infrastructure that support analytics?
Kevin Erdal: Yeah, so we talked a lot about scale today, but we also want to talk about some of those metadata management technologies and services that are out there as well. So it's one thing to have a lot of data in one place, that's very exciting, that can be very helpful, right? But if we don't have a mechanism to monitor data over time, we're really not going to be making any progress, which is the unfortunate reality. So that's where that 2 to 3 year picture becomes critically important to understand what we need to do right now to accommodate a specific use case, some of which Michelle just hit on a minute ago. But also, how are we going to ensure that our data remains healthy, if you will, for the next 2 to 3 to 10 years in front of us?
Dr. Jerome Pagani: So, Michelle, what did the state of the art for data analytics used to look like?
Michelle Peranelli: Yeah, especially within the healthcare space, a lot of what we were using data and analytics for was really for describing what's been going on. You know, I want to understand what happened yesterday. I want to understand what happened over the last year. And really being able to start visualizing that was a big step in the right direction for organizations. So being able to have dashboards that are integrated into clinical workflows, being able to have work lists that people can actively work off of, was really the healthcare industry starting to get into “how do I actually use the data on more of a day to day basis?” The thing that we see, though, is a lot of what is being worked on within healthcare organizations is really based on “What do I absolutely need to do?” Which makes sense because we want them to be focused on providing care. And the patient experience is really one of the biggest components. So, a lot of times, innovating in this space was based on “what's the new regulation that I have to report on? How do I report on that well? How do I clean my data so that I can submit it in the way that's required?” And what we really want to start moving towards is “how do we start innovating without having to have that regulatory requirement that’s driving this?” and start to move towards, “let's use the technology that's available for really improving care as well as understanding what's happening today.”
Kevin Erdal: Yeah, Michelle, you hit on this middle piece, right? I remember the days where we were super excited if we could just automate an extract to send to that external agency. That was super exciting because then somebody didn't have to manually run it, and then we'd be happy to look at those KPIs of what happened last week, last month, last quarter, last year. That's certainly not the case anymore, right? We need to know what's going on here and now. And then also, what do we think is going to happen over the next two weeks or next month? Because that's becoming critically important. Funny how times have changed.
Dr. Jerome Pagani: So obviously, health systems have been really focused on meeting the regulatory requirements. So, for how long have they been using their own data to look at either clinical or operational functions?
Michelle Peranelli: It's a great question. I think that it has varied based on how much the organization really understands about the data that they have available and trust that data. I think, you know, for a very long time within healthcare, people are asking great questions of, you know, if only I knew how often I was changing over my O.R. beds and how quickly I could do that, and what was the time that it took between those. But a lot of times they didn't trust the information. You know, we didn't document it right, these averages aren’t accurate. And I think that it's really been in the last 5 to 10 years as the EHRs and in systems have been able to handle that type of documentation and be able to report off of that documentation in a better way that organizations have been able to start really looking into those innovative questions and how do I actually change my actions based on information that I'm able to get out of the system and trust?
Dr. Jerome Pagani: And while those sort of external reporting out functions have a regular cadence that are sort of dictated, what's the goal for the internal stuff? Are we really trying to get to the point where it is almost real-time?
Michelle Peranelli: Yeah, it is definitely one of those goals. It would be so much better. I always think of, you know, when you swipe your credit card, and you get an immediate alert from the bank of, “Hey, was this you? Is this an accurate charge?” Those are things that organizations are really trying to move much faster towards of, you know, I'm ordering a test. What are those alerts that should be coming up, then how do we use data to actually give that provider information on how likely is it that this is something that you need to consider adding an additional test or a different type of test based on this patient? So it's really, you know, how do we start to use that information to real time alert the users and have them think about what's the right way for me to treat this patient or, you know, move forward with my day to day?
Kevin Erdal: Yeah. Michelle, you hit on that famous word that I love so much action, right? How do we put things in action via, whether it happens to be an alert or we can help people utilize the information that's coming to them to create the next step in a particular process. Not everything, of course, has to be real time, to your point, but the things that can and should be procedurally that can support a real-time action, that's where the analytics really comes critically important, and we can get that in front of people as quickly as humanly possible. It just creates a better patient experience.
Dr. Jerome Pagani: So, what is a typical BI modernization journey look like, and does it differ if you're starting with a couple of use cases within a department versus going for enterprise-wide transformation?
Michelle Peranelli: I think it differs in just the scale of what you're looking at, but really you want to start with understanding what is our long-term strategy as far as what are the questions we're trying to answer, who's involved, what tools are they using, and then really building out, how do we approach this in the most efficient way given those questions? So, whether it's with a department or two or the entire organization, a lot of times we start with what are you trying to accomplish over the next two, three, four years? To really start to understand, what are those key initiatives that are going to be undertaken? What are those key questions that we need specific data to help to support? And then really dive into what tools do we have today and where do we have gaps now that we know what we're trying to get to? So really, a lot of it is getting to spend a lot of time with the operational and clinical users of this information to see how they're using different analytics tools today, how they’re interacting with reporting and data, and then looking at what all has been built so far by the analytics teams, to really start to align, how do we move towards a place that we're focused on building information and using our resources in a way that people are using versus having to do one off build each time a new question comes up? So this is really the approach that we use to say, let's take what we have today, use it in the most efficient way moving forward, and make sure that our ongoing initiatives are supporting what the organization really wants to accomplish in the next few years.
Kevin Erdal: Yeah, very well said, Michelle. I mean, the reality is virtually nobody in healthcare in today's world is starting from nothing, right? There's always been something started, and that's fantastic. So, let's make the most of that. We're not going to recreate the wheel in a scenario just for the sake of doing something different. And the other piece that you really hit on that I think it's been hitting home with a lot of folks recently, is the technology and what tools are out there. And we're starting to see more and more redundancy. Every 2 to 3 years, there's a net new tool that comes out, you know, into the market that people get really excited about. And that's fantastic, right? And it might be filling a gap, like Michelle said. But we've seen a lot of scenarios where some of the tools are actually starting to really bleed together and be duplicative within a singular health system. So how can you make sure that we're helping the IT spend, if you will, picture along this journey throughout the entire healthcare organization? Sometimes a particular department might like one BI tool versus another, but at the end of the day, is it really necessary to have two tools doing the same exact thing? So as we start to think about it, like Michelle said, we're taking a half a step back here to not only look at the operational components of where the data is coming from, but let's look at that technology as well to see where we might be able to save some licensure cost long term and all the while starting to align with the organizational initiatives and making sure that the tools that are being utilized in today's world can accommodate what those initiatives are going to require in months and years to come.
Dr. Jerome Pagani: So, I think you both are touching on what BI modernization can do for a health system that allows them to do things that they couldn't do before. What are some of those typical use cases that you're seeing in our clients that kind of come up all the time?
Michelle Peranelli: There's so many areas of the healthcare world that are impacted by analytics, but I think that one of the things that we're really seeing require that type of real-time reporting, real-time access to information is the care coordination space. So being able to really coordinate not only within your organization but also with broader community organizations, with the payer, with the patients themselves, with pharmacies. Those are the things that, without really clean data that can be normalized and understood outside of your organization, there is no good way to really connect the dots there and make it seamless for the patient to be able to say, all right, I got my order from this physician. Now I need to go to this radiology center. I need to call my insurance company to get that authorized. There's so many different pieces of the puzzle, and having, you know, a more seamless transition of that data between organizations and understanding of it is going to make that process so much better.
Kevin Erdal: Yeah, Michelle, that reminded me of really two things, right? And one, I'll get straight to the point from the healthcare system perspective, what we support on a regular basis, but the other is a real-life example, right? When I buy something from Amazon.com or Homedepot.com, for example, I don't care what it takes to fulfill my order, right? I just want my product. And when I think of that as a patient, I don't care what it takes to integrate or to facilitate communication or to transfer data from a payer to a health system to my bank account as I'm paying some bills to my portal app when I'm trying to schedule an appointment, right? So as a patient, I just expect this to happen, and that's what I think we need to primarily focus on as we continue to forge ahead in a bit of a new area, right, for healthcare in terms of how we're utilizing the data. We've hit on a little bit of the real-time nature in the action-oriented piece of the puzzle kind of within the health system, but also how are we informing some of the patients that are coming in the door and we're seeing some of the geofencing and things of that nature that are already starting to help the ease of use from an access perspective. So, when I go into a health system in today's world, I just expect my provider to already have my information when they need it, which may be real-time or maybe it's retrospective, from the other health system that I maybe visited three months ago or six months ago. So, when we talk about claims integration or when we talk about referrals, when we talk about previous visits, when we talk about family history, things of this nature just have to be inherently available to our care providers. You know, to go along with that, I also want to make sure that there is capacity, right? So, we're seeing some of our organizations start to look at making real-time decisions related to a specific clinic based off of patient volume. So, there's a kind of a unique urgent care example that's come up in the last six months or so that we've been helping out with where patient volume spiked, right? And we saw some of this during the pandemic, or we've seen some of this during some of the COVID testing as examples as well. So how do we take that data, that information, and then deploy net new resources? It might be nurses, it might be front desk staff, it might be an opportunity for us to maybe reconsider a kiosk at a given location. That's all data driven, right? We need to be able to understand what's happening over time, but also the here and now. And if we need to solve a problem within the health system, we can use that information and say, okay, let's deploy some of the folks that we know can help the patients that are at that particular physical facility in this particular example. So just one specific use case that we've seen come up. We've seen a little bit more of the command center conversation come up within a lot of our clients in today's world as well. That's a really big conversation that does start to get beyond some of the traditional analytics. But the point being, how do we create that action to really support patient care? And that means not just, you know, the physicians and the nurses and the front desk staff, but really at the end of the day, providing the needs for that particular patient as they're walking in the door.
Dr. Jerome Pagani: I think there's some great examples of how health systems are beginning to use data in ways that impact patient care, make care delivery more efficient, and actually improve the overall function of the enterprise. So those are fantastic. With the Amazon example, I think you raise a really interesting point, which is that these are businesses that are designing for a particular experience, and they're using data at every step to make sure that that experience is smooth and they're getting to the outcome that they want. And that's what we ultimately want to do with healthcare, is get people to be engaged with their health on a regular basis and stay healthier for a longer period of time. So, it provides that great prevention lens. But then you're also talking about making sure that once somebody is sick, they return to their maximum level of health as soon as possible.
Kevin Erdal: Yeah, that's why we're in the game, right? To make sure that we're actually supporting those of us that need the help and the guidance from the healthcare providers. That's really where we're going to start to make a difference and impact on the society as a whole.
Dr. Jerome Pagani: And you know, we mentioned some of the nice retail experiences that folks are having, but I think that's an important point because some of those same folks are looking to get into healthcare as well, and they've been able to create that seamless experience elsewhere. And so, our clients really need to be thinking about how they are going to modernize their enterprise to do the same.
Kevin Erdal: Yeah, we hear a lot, right, that healthcare is a couple of years behind, in some cases, some of the other industries, in retail, finance, what have you. Let's use that to our advantage, right? Let's figure out what others have done. Let's figure out how others have had success in a similar fashion and learn from it and apply it. We can do it a little bit quicker.
Michelle Peranelli: And one of the interesting things that this brings up as far as there's a lot of great solutions out there that are being used by other industries is now all of those solutions are coming up to how do we actually apply this in healthcare? There's a lot of third parties out there that are bringing very targeted solutions to our healthcare providers and saying, hey, I have this amazing tool that's going to help you, you know, get that referral this much faster from your physicians. But all of these are introducing new technologies, new systems that have to be integrated to the existing systems, you know, potentially new data points that now people are going to want to be able to report on in addition to their core systems. So, a big piece of this is really understanding, what are those additional tools that really are going to be worthwhile? How do I measure that? Those tools are going to give me that return that we were expecting to see, and then how do I actually integrate it into my existing systems in a way that makes it seamless for our providers?
Dr. Jerome Pagani: Yeah, that's a fantastic vision for what the future of health is going to look like. So, this conversation really satisfies the scientist nerd in me. When I was in grad school, I had a mentor who when anytime he was asked about what the lab did, would just say, “Data, data, data.” And so, I'm really excited for how BI modernization is going to change the way healthcare entities really interact with their own data sets and the way that's going to transform patient care.
Kevin Erdal: Yeah, data can be a new form of currency, right, within healthcare. It can help us in a lot of different ways. So, it's a very exciting time, and it's very exciting to see people jump on board and start to learn more about how we can utilize data differently and how we can look at data as a true asset. It's a fantastic time to be in this field right now, and I can't wait to see what we're going to be here in the next 3 to 5 years.
Dr. Jerome Pagani: Yeah, I love that perspective. Our clients are sitting on a pile of riches just with all of the vast amount of clinical and operational data that they're sitting on.
Kevin Erdal: That's exactly right. We just need to unlock it.
Dr. Jerome Pagani: Kevin and Michelle, this is absolutely fascinating. Thanks so much for joining me today.
Kevin Erdal: Happy to be here, thank you.
Michelle Peranelli: Thanks for having us.