From 3M Health Information Systems
Podcast Episode Transcript: Finding a common language: Documentation and quality improvement
Dr. Gordon Moore: Welcome to 3M’s Inside Angle Podcast. This is your host, Dr. Gordon Moore. With me today I have Dr. Geoffrey Rose, who is the chief of adult cardiology at the Sanger Heart and Vascular Institute, and Heather Joyner, who is the director of health information management, both from Atrium Health. Welcome, Dr. Rose, and welcome, Ms. Joyner.
Dr. Geoffrey Rose: Thank you. Delighted to be here.
Heather Joyner: Good afternoon.
Dr. Moore: The context of the conversation is based on this work that we do in healthcare around coding in clinical documentation improvement. The story from Atrium that I understand from colleagues and have seen for myself is pretty powerful around that and raises the questions “Why do this work? Why is it important?” We’re going to get into how you’re doing it and what kinds of results you’re getting. I’m going to start with Heather, if you don’t mind. Why do you do this work, and what’s the work that you do with your staff in your department?
Ms. Joyner: That’s a great question. First and foremost, we do this for the patient. In the day and age of health care across the continuum, multiple providers, the diagnoses and the information that follows the patient needs to be accurate. Part of what we do is support that process. I’m an old coder—been around for a while. Back in the days, we would just code what the doctor wrote. We didn’t really ask any questions and there wasn’t any back and forth.
As we’ve moved from the volume world to the value world, we’re measured by external agencies. We’re measured internally. Patient outcomes are more important. Even though there’s still a gap between the coding world and the clinical world as far as terminology, it’s ever so important to try and close that gap by working together to get the documentation to be as accurate as possible so the coders then can turn around and assign the most appropriate codes. Whether it’s for reimbursement, measurement or simply so the diagnoses that will follow the patient are right.
Dr. Moore: You have a whole crew of coders who are looking at hospital charts and doing that work, but there’s also this thing called clinical documentation improvement. Tell me a little about that.
Ms. Joyner: Yes. We have an entire staff of inpatient and outpatient coders, and we also have a very seasoned team of CDI specialists, all nurses with whom we have a very strong partnership and work with on a daily basis. We do that first and foremost through our concurrent coding process, where we’ve teamed up coders and CDI nurses on particular units for particular service lines—cardiac, neuro-oncology surgery, for example.
The documentation improvement—Atrium Health has called documentation a team sport, and it is absolutely a team sport. The coders and the CDI nurses work together. They work with the nurse navigators. They work with the physicians, preferably while the patient’s in house to work towards getting that documentation, and it might be identifying one particular area that we can focus on. For example, acute myocardial infarctions (AMIs)—I know Dr. Rose can talk about that, or COPD, or other diagnoses where we focus on making sure we get it right, but it’s a process where it might be a face-to-face conversation.
We have tools that give the coders and the CDI the ability to work on the same platform so they don’t have to be necessarily standing next to each other. They can either be remote or not together and still communicate their findings and their questions—which facilitates that process by which we are trying to get the documentation to be as accurate as possible. It’s really robust. The nurses get to share their skill level as nurses. The coders get to show their skill as coders. Everyone learns from the process. I know the physicians have learned more than they’ve ever wanted to know about the coding process, but it’s very robust and everyone has their role and it is really a team sport.
Dr. Moore: I get the idea that we need to make sure that we represent stuff to pay the bills, and that’s fine, but at some point I think I hear from clinician colleagues they can have a hallway conversation: “I can think of somebody else. The patient has congestive heart failure and they know what I mean and we’re good to go.” Dr. Rose, you’re chief of adult cardiology. Is that not the case? Why are you engaged in this enterprise, and why are physicians participating in this work?
Dr. Rose: To follow up on Heather’s point, we want to get it right. If we’re just taking care of patients in small circles, we could rely upon conversations of, “You know what I mean” and “I know what you mean,” but that’s not how it works. We have to be able to speak in a common language about how sick or how well our patients are, and that has to be codified in ways that are similar between how—we’re in Charlotte, North Carolina in the North Carolina region down here, but the patients that we’re taking care of have to be able to be compared in some way, shape, or form to those in the Pacific Northwest or in the Northeast. That’s the reality of healthcare, and there’s a variety of reasons for this.
One is—we already talked about the financial aspects in the value-based world, but the other probably more important aspect, at least to me, is the issue of process improvement and care improvement. If we can’t speak about it in terms that are standardized and measured and so forth, we don’t really allow ourselves the capability to actually get better at taking care of patients. We’ve learned this. We know this through clinical registries. We know this through a variety of vehicles that are out there for care improvement, but this is an extension of that in my view and why I think, as a service line leader, it’s something that really merits our attention and our energies.
Dr. Moore: That sounds like a big shift. If we start a conversation about the revenue cycle coding morals, and now you’re shifting it into how you understand standard language around patients and the opportunity for quality improvement. I wonder if that’s part of the message—if you talk to other clinicians about this, and if that resonates and if you have any pushback.
Dr. Rose: That is what we lead with, because to be candid, if we’re talking about revenue cycle improvement to a number of clinicians, the room is going to empty pretty quickly. For most, that’s not what has drawn them into the field. When we talk about it with the real aspect of the ability to improve care, then we get some buy-in. To give you a quick example of this, one of the reasons we were drawn into this—a couple of years ago we were quite happy with our performance and the care that we were providing in patients with myocardial infarction. We knew we were doing well through registries where chart data were extracted through—pretty big risk processes, national registries and so forth. We felt very comfortable with our performance, only to be surprised when we got our data back from Medicare, which is information that all comes through claims-based data, that we didn’t look so good.
There was this discrepancy out there in the public space in terms of our performance from registries versus how the payer, if you will, viewed our performance. It all came back to a lot of issues, a lot of gaps that we had with respect to how we were coding and documenting and so forth. Why should we care about that?
There are some financial implications to the organization, but also for us to be confident in the care that we’re providing and know that it measures up to our colleagues in other areas of the country, we have to have a common way of how we’re talking about this that’s implicit. We found that we weren’t doing the things that we needed to do in terms of the discipline of how we were characterizing our patients. It gave us a number of gaps to fix. We’ve gone through and fixed those, and I suppose we’ll talk about that later in the podcast, but it was easy to find buy-in on my clinical colleagues when I was able to share with them how Medicare and thus the public viewed our care with respect to a condition that was very close to us as cardiology practitioners.
Dr. Moore: That’s interesting. You saw that gap and recognized it. As you were just saying, I know the truth in the sense I have the registry. I can see the clinical outcomes and to heck with that external thing, but then we’re being held out to the public and they’re waving a flag over our heads. It would be nice to have that congruent—what we know about the patients. It was extra work for me to have to bring that congruence to bear. That’s the work that you guys are engaged in, in trying to make that seamless and easy. What have you guys done at Atrium to make that easy? Did you have to set up structures, committees? How do you do that work?
Ms. Joyner: We started a couple years ago, simply trying to understand the gap. From a coding perspective, we put a hard stop in the coding system—that a patient with a certain diagnosis would not file a bill until it got a second level review. That was the beginning of our process to have as much of a real-time conversation with our physicians around a specific diagnosis at that time.
One of my goals of this process was to build the trust. I wanted the physicians to trust my team’s ability to do their job and code what was in the record. Through that process, we started with baby steps. We started looking at some charts, identifying clearly where we had opportunities, and then we built the process out, not just looking at one diagnosis or one diagnosis on expired patients. We built it out where we have a pretty robust second level review, process while the patient’s in house, and on the back end if we still need to do it on the back end.
Whether it’s CDI or concurrent coders or nurse navigators, we built queues where patients were funneled so nurse navigators could look at real time data. We have CDIs communicated to the coders, and the coders communicating to a physician champion. This whole team sport approach grew organically, but really well, and it’s now three, four years later. We have a pretty well-run machine. We are able to identify gaps that still exist, get the physician champion involved, and have that peer to peer conversation.
You were talking about the hallway conversation. This is a very formal hallway conversation that Dr. Rose or another champion can have with his or her peers. I’ll work in the documentation so then optimally it’s coded appropriately. We’ve seen some results from that. Dr. Rose, if you want to share some of the results of the work we’ve done on that process.
Dr. Rose: Thanks, Heather. Let’s look specifically at the gap that got us starting down the pathway on this work with AMI Care. We’ve gone from lowest decile performance, at least in terms of claims-based data, to top quartile performance across our organization. That has been through commitment to this process where through the initiatives that Heather’s outlined, we’ve been able to identify cases in real time, involve our clinical champions, work with our hospitalist team—I think that’s an important point to bring into place here because care is based on teams. It is in the old days where perhaps one provider would look after a patient throughout that patient’s entire hospital course. Now, tt’s teams of providers taking care of populations of patients, so we have to have that interdisciplinary communication.
In putting this in place, we’ve been able to get some real traction in terms of our results. It’s really served as a springboard for this work to extend even further in our organization. We have other service lines that are beginning to embrace the same kind of approach—the “getting it right.” I think this has implications with respect to quality assessment and quality improvement. If everybody has the same baseline, we’re all talking about the same populations of patients.
A new area for us as an organization but one we know is very important is the value space and beginning to have some sense of “What does it cost to deliver that care?” If we can’t speak to it in terms of measurable units that have some degree of reproducibility, we’re going to be totally lost when we start getting into that other important domain of care. They only way an organization can take a serious, logical approach towards improving value of care is by first defining the elements correctly and putting processes in place to ensure that you’re getting it right at the time or point of care.
Dr. Moore: Back to you, Ms. Joyner, on the question, or maybe Dr. Rose, you can let me know: you talked about AMI as an early example. How did you guys select the diagnoses to go after or the patient types to go after?
Dr. Rose: We went after AMI in particular, and the first area that we wanted to look at—if we wanted to use the tried and true low-hanging fruit categorization, that was really it for us. It was an easy way to try out the process by having Heather’s team help us find out, of AMI patients, who did not survive their hospitalizations, who was being put into the broad bucket of an MI fatality? Clearly given the nature of that particular clinical condition there will be patients who do not survive it, but we found that there were a greater number, and a lot of it had to do with the non-STEMI or the type two myocardial infarction and so forth. There was a lot of opportunity for improvement in terms of how we as clinicians were describing the clinical care.
We put in place a hard stop, as Heather mentioned, that none of those charts would go out the door until they were reviewed by a clinician—one of our physician champions. We were able to find out rapidly, yes indeed that was a myocardial infarction event, so that would go through. If it wasn’t, one of the physician champions would reach out to the physician colleague who had coded or had not really documented well and would seek clarification. It took some work, but as we went forward we didn’t have to do it as frequently, because I’m proud to say there was a great deal of institutional learning around this topic.
Ms. Joyner: I agree completely. We were able to use that as kind of a launching pad, and other diagnoses we’ve picked—for example, heart failure was around the 30-day mortality rate—or other core measures, or anything else where we had a population that we knew we could get our hands around to have this coordinated effort in the same vein. We also tapped into what I’ll call the controversial diagnoses. I think if you ask anybody across the country, the diagnoses that people are struggling with are acute respiratory failure, sepsis, encephalopathy, shock, AKI, for example.
One of the other things we put in place was a documentation committee, and this committee has been around for five years. Dr. Rose happens to be our chair. One of the things we’re able to do through that documentation action committee is when we identify these diagnoses, a subgroup of that committee, half of whom are physicians—a subgroup gets together and talks about what is Atrium Health’s definition of acute respiratory failure? What do we think that eighty percent of our patients look like, present with, for example?
This past year we actually published our adult and pediatric acute respiratory failure definition internally, because outside of 30-day mortality rates—those sorts of measures—we have payers that are denying payment based on clinical validation where they say the chart does not support the diagnosis that we have submitted for the bill. Our physicians are engaged, as you can see that they are. Through this documentation action committee, we were able to come up with that definition. One of my colleagues who manages the denial process has started using that definition as the basis for appealing those denials and we are seeing a decrease in denials. We’re winning more appeals as well in the acute respiratory failure space.
Obviously we can’t do everything all at one time, but we know we have some high target areas, whether it’s a measure or whether it’s denial, that we can apply pressure, and sometimes it’s through our DE committee and sometimes it’s through other passageways. That’s how we’ve kind of scaled it.
Dr. Rose: To quickly follow up on that point, we can’t control what the various payers are going to have as their internal definitions for a particular condition. On the flipside, we have over 2,500 physicians in our physicians organization. We can’t have 2,500 different definitions based upon clinical perception around common conditions. We have to have some kind of standard language. By putting that discipline in place and through a clinically led process, arriving at those standard definitions actually gives us a lot of traction when we are having to have a conversation with Payer A versus Payer B versus Payer C, because we’ve arrived at a standardization that is based on clinical judgment and is reliably put into place. It’s a good place to start from. It doesn’t totally get us away from that area, but it sure helps.
Dr. Moore: Nice. As you guys were describing this, it occurred to me that you probably have a pretty clear idea of how you describe the success of this. I could imagine we can count the number of queries as a measure of success, but I heard you say that as the clinicians learned about the questions they would change the way they were documenting stuff and maybe a decline in queries would be a measure of success. How do you guys define it?
Ms. Joyner: Good question. Clinical validation and denial from payers is relatively new. As we build a program out, we actually see the number of clarifications or queries. We call them clarifications that we send—increasing to get on top of—at the same time we’re sending education out, which is natural, because all of a sudden the education is out there and we’re having more conversations because now everyone’s aware of this opportunity. Initially the number clarifications goes up, so as the clarifications go up, we were winning our appeals more. As we go on, yes, the number of clarifications should drop and the number of denials should drop that we don’t have to appeal in the first place, so that is definitely one of our metrics.
Dr. Rose: One of the things that has evolved as we’ve progressed in this work—in the beginning it was very manual, and there were a lot of emails and messages and phone calls and so forth. Particularly around the AMI work we have tremendously fewer cases to chase down our communications to put in place, and I think that that represents learning.
We’ve also had the opportunity through our EMR to put in some electronic querying systems that query a point of care back to the provider and are very seamless in terms of integrating it into the workflow. We’re all working inside our electronic medical records all the time and delivering care. That’s the reality of how we’re documenting and so forth, so if we can come up with real time solutions that are not intrusive that can ask some very simple clinical questions, that goes a long way towards satisfying the concern about the particular nature of a condition. It began as a very manual process, and now as we’ve gone a little further, we’ve been able to utilize electronic tools that work well with our electronic medical record to amplify or facilitate this work.
Ms. Joyner: Because we’ve become more efficient in that process, we are able to get the data out, put it in using SAS and Tableau. We use systems to do data visualization, and when we’re able to share the hard data with physicians and our physician partners—we actually had a group come up with their own metrics—goals they were going to hold themselves accountable to as far as response rate, and in addition to the response rate was a meaningful answer.
We’ve made some serious progress, and part of that is having an efficient system that we were able to get data out of and then share with our physicians, because without data it feels like we’re wasting the physicians’ time. That’s not something we want to do, so we’ve made some headway there as well, even as far as the groups coming to a point where they’re self-managing with their own metrics.
Dr. Moore: Starting by hand gave you a lot of learning. It helped you figure out where to go and how to work the process, and then figuring out an automation scheme for information flow and technology helps expand the scope of this—that you get more, and that has been successful. The question in my mind as I hear about ramping up the number of queries—that it may be frightening to clinicians on the front line to think about all these questions coming at them when they’re trying to take care of patients, but I’m not hearing that there’s been an issue with that.
Dr. Rose: Surprisingly, no. It’s not because people all of a sudden start answering questions, but if you can put tools in the hands of clinicians in real time, that makes clinical sense. You have the chart open for Mrs. Smith and a query pops up on the side. “Is this cardiomyopathy, congestive cardiomyopathy, ischemic cardiomyopathy, or what have you?” It takes a second to click the box. If you don’t want to click it, you don’t have to, you can just move on. I think that people recognize, “Yeah. I guess I need to know this. Let me just answer it and we’re done.”
If I have to open an email message in the EMR and then go find a medical record and then go look through it and make a call, I don’t have time to do that. I’ll delay it until somebody asks me—until I get some kind of emergency message that I have to get it done. That’s the reality of being a busy clinician. If you could do things at the point of care with the ease of use that are efficient, you will get effectiveness, and I think that’s what we’ve seen.
Dr. Moore: Where from here for you guys? Ms. Joyner?
Ms. Joyner: It’s scalability. What do we go after next, whether it’s a diagnosis or a service line. I mentioned before we have concurrent coding. That is something we want to expand. There’s a lot of value in coders and CDI working together, also financially from a coding perspective, because the chart’s encoded while the patient’s in house—the amount of days between discharge and the final bill are reduced. That’s a very healthy metric. Nobody can deny that’s a good thing. It is finding the diagnoses, finding the service line, finding the physician champions, building our definitions, and continuing the team sports that we have developed at Atrium Health.
Dr. Rose: The continued message within our physicians’ organization is that while every patient is different and every patient has his or her unique needs and so forth, there has to be a capability for us at an organizational level to begin to categorize our patients to some degree, because from the task of improving quality, improving value and so forth—we can’t do that unless we have the common way of categorizing a classification. Providers have training in this in terms of the scientific method, and this aspect of clinical science is in its early days. People need to recognize that.
They also need to recognize that it’s critically important. Healthcare in the United States costs too much and it’s not reliably effective, and there are myriad reasons for that, but one of the elements is we have to be better stewards of our resources. In order to do that, we have to have a common language for how we’re applying care. We have found that that message seems to resonate when it’s said in that manner. If this is a conversation about revenue cycle and so forth, that’s a dividend. That’s a byproduct. It’s important for the health of any organization to be efficient in its operations, but that can’t be what one leads with. Quite honestly, the other part is perhaps more important in the long run. That’s improving the quality and the value of our care.
Dr. Moore: Dr. Rose, Ms. Joyner, I want to thank you for your time today and a wonderful conversation.
Dr. Rose: My pleasure. Thanks for the invitation.
Ms. Joyner: Thank you.