Informatics: Diagnosing the Healing Power of Informatics
The following excerpt is based on the book Tomorrow’s Jobs Today, available at fine booksellers from John Hunt Publishers.
Dr. Katrina Miller Parrish is a physician, researcher, author, and Chief Quality and Information Executive for L.A. Care Health Plan. In her distinguished career, she has held leadership roles in prominent health organizations, received noted fellowships, and lectured at institutions, including USC’s Keck School of Medicine. She received her bachelor’s in biology from Reed College and MD from Eastern Virginia Medical School.
Q: Dr. Miller Parrish, a hotly contested provision in the 21st Century Cures Act aims to increase IT interoperability and information sharing amongst healthcare groups. Yet those rules have met pushback over privacy and security concerns. How do health organizations and executives manage regulatory pressures, and how do they impact a workforce’s overall tactical capacity?
A: A health organization like ours is beholden to regulations. LA Care Health Plan is a public health entity, and we receive our authority directly from the State of California and the County of Los Angeles. That means while a board approves everything we do, it can also be reviewed by the LA City Council or the state. Every day is spent making sure we’re adhering to all of the rules. Those could even come from the federal government or a line of business serving the Medicaid population or perhaps Medi-Cal in California.
We’re regularly audited by the Department of Healthcare Services (DHS), and so we have to adhere to regulations, or we wouldn’t exist. During an audit, if they raise an issue, we’ll get a finding and sometimes have to develop a corrective action plan for it. We can fix it, but we have to commit to a resolution that takes time. It affects our capacity, and in contrast to for-profit organizations and how fast they can move, we have a few more hoops to jump through.
It can take years to get initiatives to implementation because of all the necessary steps we have to take. I think part of working with a public health authority or a government entity is just knowing that’s the case. You understand that you have certain requirements. Part of what my world is all about is figuring out how to go through all of those processes as quickly and efficiently as possible. Asking what can we do in parallel? What has to be contingent upon something previously done? Then trying to make it happen as quickly as possible.
Q: Does that imply you’re in favor of legislation like the 21st Century Cures Act?
A: I’m in favor of interoperability as much as reasonably possible because, for us, we want to get as much data in our door as we can, especially for population health management. The more we get, the more we understand about our population, our members, and about our providers and network too. We’re interested in having access to as much data as feasible. So basically, anything that comes along that safely decreases information blocking for a public health benefit, we can get behind.
Q: In the Affordable Care Act (ACA), the actual standards defining the Electronic Medical Record (EMR) were debated and delayed. How do you adapt to waiting for regulatory specifics? Do you drive forward with your technology initiatives and hope they’re agile enough to adjust to the final regulations?
A: In our network, we work with eight different EMR’s. The top five systems represent a majority of the population, but we have to try to figure out how we potentially work with the rest, which is why we lean on the health information exchanges. We work with three of those. It’s still not enough, and we’re very early in making that work. But what our “Health Information Ecosystem” strategy is partly about right now is working with whomever we can, getting as much good data in as we can for our population health purposes, and processing it efficiently and on time.
That takes a little initiative until you discover where your barriers are, and it hasn’t been so much the regulatory barriers. It’s technical barriers that challenge us. When we’re talking about regulatory barriers, they have to do with the type of data. For example, when we’re talking about mental health data, substance abuse data, there are some carve-outs that the state Medicaid programs will insist on treating and protecting differently.
The exchanges need to speak to each other, and they’re getting there, but the ACA legislation didn’t adequately address them, and that’s been a challenge. If considerations had been put into the ACA in a meaningful way, we might be much further down the line, but they didn’t make it about true interoperability, and they didn’t give enough incentive to the vendors to do anything about it.
Q: Health informatics leverages AI and big data strategies to analyze health populations to improve overall outcomes, but the quality of the data sets used carries the potential to influence and produce unintended results. We already see inherent bias in other industries, but what’s it mean for the medical world?
A: One of the things that we try to do, for example, in the case of claims, encounter, and population data, is to develop risk stratification or other adjustments. Because what we’re trying to do is say, “This person or this group has a higher risk or higher severity of illness.” Or, “Here is some potential for a higher cost. How can we address them differently than in a lower risk population?” So, yes, we do assign quality control resources and monitor those analytics to make sure we’re understanding the data correctly.
All of our groups come to us frequently and say, “You’re not calculating the risk adjustment or the risk stratification correctly. We think our population is higher risk than what you’re representing.” But there’s not a lot we can do with that anecdotal response right now. All we can do is grab all the data we can find.
Of course, a lot of data isn’t perfect. It may be a spectrum of corrupt to bad data entry. If we’re talking about the kinds of codes that we’ll get representing a claim or encounter, the quality of that data will vary. A claim is when you’re asking for reimbursement for services, so it’s a fee for service scenario, like an invoice. An encounter is when services are under a capitated payment where we’re already paying that entity a monthly amount to take care of a certain number of people. We get data in the door about the services they received, but they’ve already been paid for it. Therefore, the incentive to send us good encounter data is far less than the claim data.
Q: As the quality assurance professional in your organization, even though you’re primarily responsible for population health care, are you also possibly catching some fraud?
A: Yes, in some cases. We have a special investigations unit, SIU, and I’m the chair of the Credentialing and Peer Review Committee. That’s where we look at those kinds of issues, and we work very closely with our SIU unit. For example, when they identify providers who are just writing tons of prescriptions for one particular medication and we find out that these patients whose names are on the prescriptions never received them (and don’t even have a diagnosis matching them), it raises a flag. We can identify them through algorithms, through data from payment integrity or our pharmacy data.
Q: Do you develop your systems and tools for this kind of data mining and analytics, or is there software already available?
A: It’s a mix. There are software tools out there where you can do that first pass at running the data and finding trends, but I think that we still are learning how to set those algorithms up. It’s another place where bias could come into play. We could identify the wrong people, and so we have to review the first pass of the data carefully. If it doesn’t make sense, we try to confirm it. If we see a trend, we have to ask, “Okay, is there a good reason for that trend?” Let’s say we have a provider identified for tons of prescriptions of one type, like risky or expensive, but then we go and find out that that’s a neurologist who is dealing with kids with intractable epilepsy. Okay, well, then that makes good rational sense.
Q: Do you have to prioritize who you put under the microscope, especially with the SIU, because you can’t go after every single problem or person?
A: Yes, and we have ten thousand providers! In this committee, we’re focused on the providers’ side of things a little bit more than the member side of the equation, but the numbers are so huge you have to use whatever is available. This is what we do with population management, as well. We’ve got to figure out what that spectrum is, then decide where to put resources.
Q: In response to technology’s effect on litigation, most states require attorneys to demonstrate technical competence. With the growth of Informatics, do you expect medical professionals to be held to similar education requirements?
A: I hope so. I’ve got a family medicine background, but I also have a clinical informatics background, and there are maintenance certification requirements for both of them. Again, we’re all working with EMRs, and if you have folks who could do so much better with that knowledge, even with just basic EMR wisdom, like knowing about pre-checked order sets or templates for notes or ways to find different orders, it would be advantageous.
Q: Would that type of education be as relevant for an oncologist as opposed to a podiatrist?
A: It depends. Everybody needs to have some level of experience to enter and pull information out of an EMR. There are some specific tricks to be able to do that well, and it’s not just about how to make sure the information gets in appropriately. Successive folks are going to be seeing those patients and have to understand what happened to the patient. If you have a medical assistant or office staff, and they’re trying to find a report on a patient, how easily can they get to that data, and in how many ways?
Q: We now have changes to board structures where Chief Information Security Officers are no longer reporting to a Chief Information Officer. They’re going directly to the CEO because cybersecurity is so paramount. Is that now the case with medical organizations? Does the Chief Medical Information Officer sit in the boardroom?
A: Generally speaking, I think the C-Suite remains similar to what it has looked like for a while. In our case, we have a Chief Medical Officer, and then you have some others reporting up and transitioning all of the time. We have a Chief of Enterprise Integration, and I’m Chief of Quality and Information. I think the reporting structure itself can be a little bit different sometimes depending on the organization and the people. If you have a COO who is better at overseeing some areas of medical operations, then it’s appropriate for some of those administrative medical staff to report up through that person. It depends. The technology efforts that we have here are just ever-present and constantly changing, and of necessity must be as efficient and flexible as possible at all times. The reporting structure almost doesn’t have to matter so much as long as you have the affected stakeholders making sure the right things are being done for projects and initiatives to succeed.
Q: So, what exactly made you choose the field of Informatics?
A: I went into Informatics because I liked the combination of clinical with data and with business. I think it’s part of being a family physician. I like being able to have my hands in a lot of things and understanding a lot about data, which is the informatics side. I like having that variety. When I was doing family medicine, I felt like I could do way more than taking care of one person at a time. I loved my patients, and we had great relationships. I learned a lot, they learned a lot, but I really love population health because I can make an impact on millions of people with huge programs that can not only affect a community in Los Angeles but conceivably far beyond that.
Q: As a trained physician, you have a perspective on the stresses of data-driven life on the body and mind. Society is just beginning to understand the side effects of excessive dependence on our devices. How do we address infringements of technology on work-life balance?
A: I have a department of about 90 people, and I try to make sure they understand their priority is themselves, their second priority is their family, their third priority is work. I try to reiterate that all of the time for my department, and when it comes to individual people with their own issues, I try to make sure they’re focused on the right things. They’ve got to have their priorities in order and believe if they come to me and have to take some days of PTO that I’m going to understand and put that before the demands of an audit.
Audits will happen; work will happen; L.A. Care will continue to exist. The most immediate thing is that these people take care of themselves and their families. Maybe that comes from me being a family physician or being a family person, but it’s in there. It’s ingrained in me that I need to make sure everybody knows that.
That said, I have to try to model the behavior. That part is not as easy. I do get complaints from colleagues that tell me I’m not always practicing what I preach. So, every day I too have to work on that. What time do I have to be at work? What time can I leave work? Do I need to be here seven to seven? Do I need to be the one taking care of editing and reviewing all of the documentation that comes in and out of my department? I want to do a little better with that, but I think it’s a familiar challenge.
Q: We now have this concept of social media as a prism through which people begin to see themselves, where every person must have their brand. This is an incredible pressure that I don’t think anybody ever expected. How is that affecting people’s work life?
A: Well, millennials are a whole generation that has grown up with these devices attached to their hands, and even more so the later generation. One example of my own sort of experience with that was when I was on call. I had a beeper at the time, and I started to have a physiologic reaction to every time that sound would go off. You knew you were going to deliver a baby, or you had to go to the ICU or the ED, or something was going to be an intense situation to deal with. And what I progressed to is now I rarely, if ever, have my phone actually on a ring. I always have it on vibrate at this point. I don’t pay attention to it sometimes. I put it away, and, in the evening time, I might not pay attention to it for the whole night.
You have to do what works for you. Some people are workaholics, and they want to work every day of the week. That’s fine, but I do want to try to espouse and motivate for a better type of balance.
Q: You’ve achieved success in medicine, and now you’re immersed in informatics and business optimization. How has your medical training informed your approach to solving the business piece of it?
A: I think first of all, in medicine, we learn to assess a situation, take in the data, try to figure out what a differential diagnosis would be. You’re never trying to go right to a solution. You want to see what all of the possible solutions are in all relevant scenarios. I kind of think of differential diagnosis in a way, like a root cause analysis, where you’re trying to look at all of the possibilities before you get to your final answer. And then, when you get to your top three, top two, or even the only one it could be, based on the data you have available, then you move into your solution. You could look at your solution as a project, as an initiative that follows a particular process.
We use the System Development Lifecycle process and others. And it’s interesting how much my medical training set me up for being able to assess data in a way that falls in line with almost any kind of stepwise assessment.
Q: Your work has taken you from the L.A. marathon to as far as Tanzania. What’s the best lesson you learned going out into those communities and abroad that you’ve been able to bring back into your work and professional life?
A: It’s a simple answer, but it’s to listen. And if we’re talking about the Los Angeles marathon, you’ve got five seconds to listen to that person and see what their exact issue is and try to figure out what to do about it because they’re running and they’re going to come in for a few seconds, and you’ve got to get them out on the road again.
In Tanzania, it is so foreign to western allopathic medicine physicians as to what could be going on with a patient, that you’ve got to listen to what their story is to understand that, number one you may be looking at something like malaria or something unusual in the United States, but it could also be something totally different. For example, we had issues in one particular area where there was a myth that if you had AIDS and you had sex with a virgin that you would be cured. And so, we had to deal with that situation and those who truly believed. We had to listen and think about what we could do with the population there to try to redirect them to the right people who could change that perception. But we had to listen to those folks to understand where to focus our efforts.
That being said, with medical care, the minute the patient walks in the door, you’ve got to let them tell their story. You got to give them the time because you can’t assume by looking at a person or looking at the data that you know what’s going on. Not only that, you don’t know what their priority is. If you’re not working with them on their priorities, then they’re not going to trust you in terms of how you’re working with them.
The same thing is true in business. If I don’t listen to what my direct boss, my CMO, is telling me about his preference or his opinion or priority, I’m going to go the wrong direction completely. I’ve got to listen to what he’s saying, to listen to what the CEO is saying, and put it all together to make sure that I strike the right balance. The same thing can be said for my department. If I’m not listening to them and or understanding what their real issues are, then we could have problems in terms of employee engagement.