peaker 1: You are listening to Your Practice Made Perfect, support, protection, and advice for practicing medical professionals, brought to you by SVMIC.
Renee Tidwell: Hey everyone, welcome to this episode of Your Practice Made Perfect. I'm your host, Renee Tidwell. In this episode, we will be discussing medical malpractice data and analytics. Essentially, how can data be used to ultimately help increase patient safety and why is it important for a provider to understand those analytics. Joining me today to help us all understand this topic a bit better, we have the vice president of Candello, Mike Paskavitz. Welcome to the show, Mike.
Mike Paskavitz: Thank you, Renee. I'm really happy to be here.
Renee Tidwell: I'm really looking forward to digging into this topic with you. Before we get started, we'd love to get to know you a little bit better. Will you take just a minute to share a bit about yourself for our listeners?
Mike Paskavitz: Sure. So I am five years actually at CRICO leading the Candello division of CRICO. My career in healthcare goes back to the olden days of the late '80s, early '90s. I started my career at Boston Medical Center and Lahey Clinic out here in Boston. My involvement in safety actually started again in the early '90s. In 1992, I started a national publication called Briefings on Hospital Safety, and it was because I had tapped into this network of thought leaders who said even at that time before any of the literature was out there that safety was a problem in healthcare, these folks were saying that, "This is going to be a big issue in healthcare and if you want to learn about safety, you really need to go outside of the industry to learn about it." So we created this publication and it was a real extreme success, which was a real indication that this was going to be a very significant coming movement.
From there, I spent five years in England working with the National Health Service on two initiatives. One was a corporate governance initiative and the other one was helping set up the National Patient Safety Agency for the NHS in England. And those are two phenomenal experiences. Since I returned to the US, so to speak, I was involved in a couple of startup companies that were using technology and healthcare in really interesting ways. One was a web and mobile online community platform for doctors called QuantiaMD, and the next one was a really advanced learning system called knowledge factor, or their amplifiers, their product.
I became very fascinated. And I'll tell you that the real driver behind both of those interests for me was data. Knowing that technology had an ability to capture data just by interacting with it was a really powerful concept for me at the time. And actually it's what brought me ultimately to CRICO because they had, when five years ago and still do, the largest database of its kind anywhere. So there was just such potential, I think, to do really meaningful work in healthcare. And it all dates back to the early '90s when I became really interested and fascinated with safety.
Renee Tidwell: I can tell you're passionate about this and I'm really excited to learn a little bit more about how the data can improve our safety for our patients. So you mentioned Candello briefly tell me a little bit about what is Candello, what is it that you do?
Mike Paskavitz: So Candello is really the data division of CRICO. CRICO is the captive medical malpractice insurer for the Harvard Medical Institution. So that's Mass General Brigham or MGB as they're now known, Lahey Clinic Medical Center, and Beth Israel, Dana-Farber, Children's, the large institutions around the Boston area. And in the same way that any other malpractice company has a claims division and underwriting and patient safety, so does CRICO. I think what's different is that Candello is the data division and it's a data division because the company cares deeply about data and the value that can be provided to the industry through it.
Renee Tidwell: So how did Candello get started?
Mike Paskavitz: So the Harvard Medical Institutions founded CRICO in 1976, and it was really in response to the malpractice crisis at the time and the Boston institutions feeling that we really were over a barrel with the commercial sector and we think we can do it a better way. And so they created what I think was the first captive insurer in the United States and started CRICO in that way. So CRICO is a not-for-profit member-owned insurance company. And from the very beginning as a company, they were very interested and fascinated by data. Now, back in 1976, data probably took the form of cocktail napkins and sticky notes because we were really sort of in the beginning curve of the information age.
Renee Tidwell: Still information though, even if it's on a napkin, right?
Mike Paskavitz: Absolutely. Absolutely. So CRICO actually was really committed to learning from these cases and these claims that they were managing and defending. And so they committed to the idea of data and developed a taxonomy, a data language. And that taxonomy endured for them. And for 20 years, CRICO coded its cases and had a database and did analytics using just their own data. I think even after 20 years, the data that they had really wasn't a substantial enough database even after 20 years, simply because it was CRICO's own data, to be able to know when a trend was a trend. I think the idea of starting a national data collaborator was inspired by this, by the CRICO board and the CRICO leadership at the time.
So really, CRICO and Candello, and Candello at the time was known as CRICO Strategies, again, a division of CRICO, started this national database and went out to other insurance companies around the country and put out the idea of forming a national data collaborative and adopting the same data language and coding those cases, putting into a database and then either allowing them to access it to do their own analytics or have CRICO, or CRICO Strategies at the time, now Candello, do the analytics for them.
As important as the data, I really can't emphasize this enough, was the concept of community. This idea that CRICO and the leadership thought at the time, and still do, is that data without people and without conversation is just data. Nothing really happens from data unless people learn from it, get inspired by it, get direction from it and do something with it. And so the community aspect has always been part of what CRICO did for its own members in convening the surgical chiefs and in convening the OB chiefs and having a lot of education as part of what was provided to the CRICO members.
So that concept of community was really a reflection of what CRICO believed in his own DNA. And so Candello, or again, CRICO Strategies at the time, Candello is the name as of a few years ago. We rebranded, but for many years it was CRICO Strategies, has always taken this concept of community. So the combination of data and really good data and very carefully curated data was complimented by this community of people that all have the same interests and the same pursuits. And I'm trying to learn from each other and as much as learning from the data itself.
Renee Tidwell: You mentioned a couple times members. So can you talk to me about that for a few minutes? Who are your members? And then it sounds like you've got a lot of great services, innovative at the time and even now when you first started and then again today. But talk to me for a few minutes just about who are your members and what services do you provide specifically.
Mike Paskavitz: We have three types of members like SVMIC, like MedPro, The Doctors Company, Constellation. These insurers are typically larger, but we have other smaller commercial insurers. And then we have captive insurers and AMC or academic medical center captives like CRICO is, whether it's University of California, University of Michigan, MedStar Health in Washington. And then the third is the risk retention groups. Those are groups like Cassatt in Pennsylvania area. So it's really those three kinds of companies that join and contribute data to the database. Those are our members and they're data contributors. We have a team of professionals who provide that support to all of our members.
One of the things that we do when we welcome a new member is have a conversation about where they are and where they want to get to. That's typically in the areas of data analytics, patient safety, what we call engagement, meaning that people are paying attention to the data and that's its own kind of pursuit. And then also your traditional claims, underwriting and finance as well. So a new member comes on board, we want to see where are you now, where you want to get to. And the services we provide are directly related to that.
I will say that even in my time in my five years here, data analytics has advanced so much. Whether it's the technology, the analytical tools, the data and the amount of data out there in general for the industry has accelerated really quickly. So where I think CRICO Strategies prior to 2019, a lot of the work that was done with members was done with us doing the analytics for them and a lot of the coding was done for them. We've found, and I think it's not surprising, that most organizations have a commitment to data analytics themselves. And where they used to have possibly no capacity or very little capacity, they're now seeing the importance of data and needing to have that internal capacity. So where we used to be very much a service provider, concierge, we have received a strong message from our clients and our members that self-service is one of the big themes that we've been honoring over the last few years. So there's just a general desire across the industry, I think, that we need to be better informed by data than we historically have been.
Renee Tidwell: I'm interested to learn a little bit more about the data collection process. Can you talk to me about how do you collect and compile the data for use?
Mike Paskavitz: It all starts with the taxonomy. And the taxonomy is, again, our data language. Everyone in the Candello community basically puts their hand up and says, "Yes, we'll adopt this language. We believe it's going to deliver valuable data. And more than that, comparable data." So if we're benchmarking, if you don't all collect the same data, then a good percentage of what you're trying to analyze is not comparable. So it's really an important point that the shared taxonomy is really the spine of the database.
Renee Tidwell: It gets everyone on the same page, right?
Mike Paskavitz: Everyone's on the same page. And so when a member joins, they've already seen the taxonomy, and that's hundreds of codes. We have a very substantial infrastructure to support the taxonomy and ensure that it continues to grow and mature. And by grow and mature, I mean, as there are emerging risks like COVID, virtual care, that those kinds of codes are included in the taxonomy so that organization can start to contribute that data and then analyze it on the backend.
We've also introduced a whole range of new kinds of codes that allow you to do new kinds of analytics. In development now we have what we call litigation codes, and that will enable organizations to do litigation analytics, predictive litigation analytics to help them determine the likelihood of defending a case and how successful that defense might be or even just the decision about whether to try a case or not. So we're always getting pushed by our CRICO members and our Candello members to make sure that the language and the taxonomy is best in class and current. So that process is really important.
What happens once a member has signed up to the community is that they determine how they want to code and who wants to code. By how they want to code, we mean do they have their own internal team? Typically, all of our coders have historically been nurses and continue to be with significant experience. There's judgment involved in the coding process. The taxonomy appears in the form of a coding manual, and that coding manual guides you to collect the data that is being sought. But there are points along the way where you have subjective judgment by the coders. The more experience they have in doing the coding, the more proficient they get in making those decisions.
What it also means is that we're able to audit the data. So all of the coding has constant and rigorous auditing to make sure that there's integrated reliability, that the data is thorough, consistent, reliable, all of that. So it's really a matter of deciding who's going to do the coding, you meaning the client or the member or us. We have a coding team and it's a well-resourced, very experienced coding team. It can take anywhere from an hour and a half to two hours to code to deeply code a case. And deep coding is really where the significant value comes. So it's a really critically important job in the whole process. We treat all of our coders. Whether they're coders who work for a client or a member or employed by us, we treat them as the same. So they meet regularly. It is as if they work for us in the sense of the support they get.
Renee Tidwell: So from this process that you've just described to us, what kind of data are you hoping to collect?
Mike Paskavitz: This is a really important point, and I've referred to the term deep coding. Deep coding really means that you have a deep reservoir of source information that is coded using the taxonomy codes. And so that deep source information includes information from the claims file, expert witness testimonials, depositions, root cause analysis if they're available, and even the medical record. So the clinical coding, which is really the clinical information that describes and that allows the coders to understand why the event happened that led to the claim. That's actually the real differentiator in our data and other data that's out there. It's really about the why. Not what happened, where it happened or who was involved, but actually why it happened. That clinical data is incredibly valuable when you're trying to affect change on the ground or try to understand trends where interventions can be meaningful. If you don't know why something happened, it's really hard to correct it.
Renee Tidwell: Absolutely. It sounds like this is super important and valuable. We ensure physicians. Can you maybe explain to us why this type of data specifically is important to a practicing physician?
Mike Paskavitz: Sure. So I mean malpractice, there probably isn't anything in a physician's career that is more impactful than being part of a malpractice claim. We have a long, long history of defending physicians and working through them emotionally, intellectually. So the impact that a malpractice claim can have is really dramatic. So the learning involved from a malpractice claim, and I'll say this and it sounds a little bit dramatic, is that other than maybe a natural disaster like a hurricane or a workplace violence incident, there probably isn't much in healthcare that is more studied from more perspectives than a malpractice claim because you're getting lawyers, you're getting unbiased experts. The level of scrutiny that goes into understanding what contributed to a malpractice claim and then learning from that, I would argue it's very, very hard to find. Even root cause analysis done on serious reportable events, while they can be very rigorous, there's a lens on that that isn't part of the conversation, and that's the insurance lens and the financial lens.
So the richness of the data is really important. And what that means is in the world of physicians where they care a great deal about evidence, I mean as arguably their DNA, is to make decisions based on evidence and experience. So evidence means a lot. And the basis of really sound evidence is there is a rich amount of source information and a rigorous process to draw from that. And that's what's really different. So what it really means is that physicians can be ensured that this was learned from real things that really happened that had real consequences both for patients and for the providers. There are real dollars involved so you can understand the financial impact of what's happened, and that is increasingly meaningful.
So I think that what we've heard from our members is that the data is really meaningful to physicians because it is based on credible evidence or credible data. It is national, right? So you're never a profit in your own land. When you try to convert somebody and it's based on your own either data or opinion, a national broader perspective is often really important to turn the tide on someone.
So I would sum it up this way. The data is inherently rich and very carefully curated. And it's also very focused on the clinical aspect of the event. So not just the claim itself and the defense of that claim, but also what can be learned clinically. And there's just so many examples of things that have changed on the ground by providers based on the malpractice data that they saw.
Renee Tidwell: So Mike, transitioning from that, how has the use of data and patient safety evolved over time?
Mike Paskavitz: Well, I think the data has evolved as the demand for it has evolved, but also the taxonomy keeping up with that demand. And I will say that almost all of our members, if not all of them, and then the CRICO members as well, are raising the bar for data analytics over time.
As I mentioned earlier, data is just data without people to work through it and learn from it and do something with it. And a couple years ago, not that long ago, data and analytics was sufficient, just having the two of those. But now the leadership of our members are saying, "We need to start seeing needles move. We need to start seeing change on the ground inspired or using this data. We need to start to see an effect on our business." So it's really representative of the evolution of data analytics. So I think what that has done has challenged us in two ways. One is to really make sure that we are out in front and on pace with the way that the clinical environment's changing. So the data's meaningful when it's generated.
Physicians often may dismiss data that's outdated or not expressed in a meaningful way. So I think there's a pressure that comes with that demand. If the goal of getting involved in data analytics is to change things, then you have to engineer the front end and the backend. And I think in the last couple of years we've been working very significantly with our clients, and it kind of came from this, I mentioned, self-service analytics. One of the things that accompanied that was we survey our members regularly. We meet with them all the time. And one of the things that they started to share is "We need more people in our organization to be exposed to and understand the benefit and the value of the data." In many organizations, it was really their analysts that knew the data or maybe the risk manager knew the data. There's an adage in patient safety that says, "If the only person that knows about patient safety is the safety manager or safety director, then that's probably the only person in the organization who's safe." So safety is an evangelical thing, and so is risk management.
So we started to get the signal very loudly that the data needed to start to be disseminated across organizations. Boards needed to be better exposed to it in a way that was meaningful to them. Clinical leaders needed to look at it and say, "Okay, this is credible. This is a really valid point." Insureds themselves. So the insured physicians, we have members who, when one of our members will use a tool in front of a practice they're insuring, they see that there's national data and that matters. So it's really been twofold. One is improving the data on the front end and then also improving how the data can be disseminated and received on the backend.
Renee Tidwell: So are you seeing a lot of your members using the data for education purposes? Or how do you see them using the data?
Mike Paskavitz: Absolutely. One of our products is a product called Candello Discover, and it's a search engine. It's a tool that is really a query tool that allows you to reach into the data and say, "I just had an event or a claim or a potentially compensable event like this. What can you tell me, Candello, about other events like this around the country? What's been learned from them? What was the average indemnity payment? What was the close with pay rate? What were the most common contributing factors? What were the top major injuries?"
So we made that tool so easy and so accessible that anyone could really use it as long as you knew what kind of event you wanted to understand better. And what that's done is it's become an educational tool. So all of the CRICO patient safety folks use it with the CRICO members. We have members who are using it to teach residents. Residents want to know what's the malpractice profile. And even medical students, so they use it when they're rounding. And so we've tried to put these tools in the hands of our members so that they could engage their stakeholders using data. And that didn't used to be true just a few years ago.
So we have a set of dashboards that really are designed for executives to look at and understand what the trends are, ask questions. It was developed for them. They're the recipient of the data, so the data has to be meaningful to them. So that's been a big part of when I talk about the front end being the data collection and the backend being the expression of that data. That's really a lot of the focus that we've had in the last few years.
Renee Tidwell: Do you have any real world examples of how the data has been used to affect change? Any specifics you can share with us?
Mike Paskavitz: Sure. I mean, there's lots of them. We have these things what we call value stories. When I said that the leadership in our members has been raising the bar and looking for results and needles to move, one aspect of that is the classic term return on investment. What is the dollar for dollar return? Anyone who's been in the risk management or patient safety business know that that's a very hard equation to produce. And I think we're moving towards that, as is everybody. But I think what they're really asking for and what they told us they were asking for is we need to start to see that there's value to all this. And so we've been collecting these value stories that our members provide to us. These aren't puff pieces that we go out and write. They're really what's happened on the ground for our clients.
One, for example... This is a really interesting example. One of our members that has a pretty low volume of claims, so their ability to learn from their own data is limited by that volume. However, because they have access to a national database, they can go out and look at if they have a particular claim and say, "Is this a national trend?" It really empowers them to make changes or recommendations or decisions in their organization. So the national data is really meaningful. In one particular case, one of these clients didn't even have a claim, but just did a query on Candello Discover around the problem of anticoagulation and saw that there was a pretty significant trend around anticoagulation and realize that this doesn't seem to be related to geography or organization type. It's a national healthcare trend and issue and risk. So based on national data that they had not experienced themselves, they were able to convince their clinical leadership to invest in a system-wide anticoagulation program.
So that was a perfect example of looking for risk and opportunity in data before you're affected by the problem. And I think that's ultimately what risk management is about, is getting out in front of risks. But in the malpractice space and the patient safety space, we just have a long history of being responsive to events after they've happened. And data analytics is really designed to help us get in front of that.
And so a number of other examples, we have another client that has been using the data and is particularly in OB to identify learning gaps to improve the intervention of education. I mentioned that there have been some advances in our taxonomy that have really helped the data be useful in new and different ways. One of those is a set of what we call linking and waiting codes. And those are just a set of codes that allow a coder to not just code to the defendant, but our cases, we code to everyone involved in the event. These codes enable you to essentially determine the extent to which those who were part of the experience, part of the event that led to the claim, were actually involved in the claim. So in a surgical case, was it really the surgeon, 100% the surgeon? Or was there an issue in a surgical tech or with one of the assisting surgeons?
Renee Tidwell: So it really helps you learn what all the moving pieces are.
Mike Paskavitz: Absolutely. I will tell you that has been a huge gap out there, is team-based learning. And trying to understand where knowledge gaps are and where things happen on a team, it's that classic Swiss cheese model because team-based care is both the answer, but it also becomes a challenge when trying to understand the dynamics of a team, especially when something goes wrong. So these kinds of analytics are very meaningful to our clients and they're starting to be able to do what we would call precision interventions. So they're not just taking an education course and carpet bombing everybody with the same education when for some it may not be relevant at all. That kind of precision intervention is really the byproduct of really trying to understand in a very, very deep way, what went wrong, why, and then the extent to which individuals contributed to that event.
And it even reveals team weaknesses. So it's not even necessarily attributable to an individual on the team. It could be communication is a classic problem in healthcare. So the analytics are getting more and more precise, and our members are using that data. We've had other clients who were looking at OB cases and had an assumption about what the problem were with these OB cases and found a completely different one. One looked and found, to their surprise, how the root causes of, or the contributing factors for shoulder dystopia were nothing on what they thought they were. So there can be real discoveries in looking at the data when it's deeply coded and granular and includes that clinical insurance and financial information.
Renee Tidwell: Well, and what better way to learn than from someone else's issues that they faced and try to prevent it from happening on your end.
Mike Paskavitz: Yeah, absolutely. And again, one of the things that CRICO is proud of and is appreciated for is that it's a very competitive marketplace in Boston. One of the things that CRICO members agree on is that they don't compete on safety. So you get CEOs from competing institutions that say, "We don't compete on safety. We just don't." And I think there's a similar agreement between the Candello members, many of whom are competitors, but there's insights to be gained from each other. And the extent to which they can, it's very transparent.
Renee Tidwell: Shifting gears very briefly before we wrap up, can you talk to me just for a few minutes about what the future looks like for data collection and analytics?
Mike Paskavitz: Yeah. I think the word of the day is AI. Everyone is very interested in AI and its use in healthcare specifically in our world, in malpractice and how it can be used. We've seen just our own internal dabbling. We've seen it's got great application for ChatGPT specifically for generating content, which can be helpful in some ways. The big question most people ask is, what does AI have to offer in terms of coding? Can you code in an automated way using AI or an AI product and can it do it proficiently and faster and less expensively?
At several times, CRICO has investigated the use of AI in natural language processing. And I think what we found along the way on many occasions is AI doesn't know what it doesn't know, and we aren't particularly willing to take the risk of what it doesn't know until it does know it. And so I think we've had a very careful eye on the use of AI and we're continuing to explore it. We think that it could be a very productive assistive device in the coding process. It probably has application and auditing. And then the use of AI itself, an AI product or model sitting on top of a database has lots of potential too. But we're taking this very deliberately. We have a lot of information that's important and sensitive, and we're just being very thoughtful about how we approach it.
So I would expect that there isn't going to be a watershed event where robot replaces man, which is I think what a lot of people are fearful of. I think we're going to see that in a lot of different parts of our economy, AI is used in different ways and it'll be a real benefit to a lot of things. We see it as a real benefit to us. We've seen the ability for people in our communications team to use ChatGPT or something similar to generate content which they then review.
So if you were to apply the idea of an AI approach to coding, I don't think you can ever really eliminate the human aspect of that because there's incredible judgment and context that often results in real insights in the data that even the most advanced AI model can't really pick up. So there may be a different change in the way that the dynamics of people in technology, but I don't see this as when the data entry clerk lost... That profession sort of went sideways in the 1980s when people had their own PCs.
And I think forcing functions are really healthy things and concern and fear can be a great motivator. And so I think that when you see something that can be as potentially game-changing as AI, it's going to bring out the best in people as they figure out how to best use it. And so I think it's probably a healthy accelerant to the data analytics space in this industry.
I do see also, it's our goal at Candello and CRICO, that malpractice data has a far greater exposure across the industry because again, I have just a huge amount of confidence in its core value just from the richness of it, the depth of it. And I think it's an unpolished jewel in the data race around the industry.
Renee Tidwell: I think that's all been super valuable. Before we wrap up though, would you like to share any last minute tips or pieces of advice?
Mike Paskavitz: I will say that one of the things that we've noticed, it wasn't so long ago that a lot of our commercial members felt as though they were too far removed from the point of care, the point of service to really make an impact. What we've found over the last couple of years to the one is that insights, patient safety, learning that comes from a credible source is extremely well received by physicians. And I do think that there's a great opportunity to build a partnership with the physician where it's trusted, it's credible, and it's actionable. And again, that actionability is really important, I think, when you're talking about physicians and providers. When data paralyzes you, you almost get frustrated with the data. But when data tells you where you ought to be pointing yourself and it's done so in a really, again, credible.... And I use that word often, but it's important. When it's done that way, it's very impactful. So I think malpractice providers have a great deal of opportunity to build partnerships with their insureds in a way.
Renee Tidwell: Absolutely. Mike, thank you so much for being here today. Listeners, we always appreciate you listening. We will add some links in the show notes for some resources if you're interested. And with that, we'll say goodbye.
Speaker 1: Thank you for listening to this episode of Your Practice Made Perfect. Listen to more episodes, subscribe to the podcast, and find show notes at svmic.com/podcast.
The contents of this podcast are intended for informational purposes only and do not constitute legal advice. Policy holders are urged to consult with their personal attorney for legal advice as specific legal requirements may vary from state to state and change over time.