Russell Howerton, MD, Chief Medical Officer, Wake Forest Baptist Health, discusses "System Based Approaches to Improving Quality of Care for Cancer Patients."
RUSSELL HOWERTON: So when you speak to someone who has an administrative leadership role about performance improvement, or improving something, the quality of care, a standard set of tools comes into mind, and I think many of you will be familiar with these. Performance improvement is a fairly well-developed science with certain tools. There are models for how one would improve. We have teams of people in the health care facility who know how to do that and to help others do that. You see a few listed here. They come in many flavors and names. And they all have a fairly common premise-- finding something to set a name around, to develop measures, and to select changes that improve. But there is just one small detail before you get started. So we were-- Perry asked me to speak about improving the quality of cancer care. So what is the definition of quality in cancer care? Now, I recognize some old friends here that I can call on. It's nice to have a small audience. Paige, can you help me? Help me know, what is quality in cancer care? AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: OK. Michelle? AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: So you can see why I was a little stuck when Dr. [INAUDIBLE] asked me, what are we trying to improve? What is quality of care? AUDIENCE: [INAUDIBLE] good care. RUSSELL HOWERTON: Good care. And how would I measure that? AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: So that would be reasonable. You know, quality in health care is a very elusive thing to measure. The Institute of Medicine has a definition for quality. It is care that is safe, timely, effective, efficient, equitable, and patient-centered-- the so-called STEEP criteria. Well, that's a step. But even then, what would be a measure that would speak to each of those domains? So we're all-- I'm assuming almost everyone in the room is a professional engaged in the care of cancer patients and overwhelmingly motivated to do good and bring good to the care of our patients. We have professional knowledge and expertise, and we carry sets of professional metrics with us-- radial margin in rectal cancer, number of lymph nodes in an axillary specimen, rate of deep vein thrombosis in post-operative patients, readmission rate. I can name all kinds of individual metrics. As a matter of fact, I did a Google search getting ready for this talk, to think of, what metrics could we use? So as a surgeon, I naturally went to the National Cancer Database and-- partly sponsored by the American College of Surgeons and the Committee on Cancer's quality of care measures. I extracted two here, to make the point that there are dozens and dozens of measures on the Committee on Cancer's site on quality of care. Are there medical oncologists in the room? I know at least one. Good. So the American Society of Clinical Oncology has the Quality Oncology Practice Initiative. It is a tool set developed by oncologists, meant to facilitate the measurement of quality in ambulatory practices. I think if-- maybe y'all can correct me, but there's 142 measures in the measure sets that are available to be used. And they have downloadable tools by which you can develop processes in your practice to measure them. As people interested in oncologic disease, the Institute of Medicine in 2012, a very learned body of-- speaks authoritatively in the health care industry, convened a group of experts for this very question, delivering high-quality cancer care. What is high-quality cancer care? It makes for some pretty interesting reading. Little sobering to read the first premise, for those of us in cancer care. So I won't read that out loud, but I think all of you can see it. That's the starting premise of the group that produced this report. Would anyone here be ready to refute that statement right now? I probably wouldn't. Their concept is at a very high-level here. So they set some noble aspirational goals for a system that would care for cancer patients. And I don't think anyone could argue with any of these. But they're at a scale considerably greater than the practice lives of most of us. Translation of evidence into clinical practice, quality measurement, and performance improvement-- no one's going to argue with that. None of us. We all want that. Most of you probably can't impact getting an adequately staffed, trained, and coordinated workforce. Right? That's-- at a micro level, most of you are trying to do that in your given practice environment. But those are not really manageable for us individually. This is the summative output of that document, which describes a system-- and it's a great learning system-- that starts with patients in the center, and wraps them around excellent interactions with a workforce that was described earlier using technology and measurement for performance improvement tools, and then bring change. This is mom, apple pie, and Chevrolet. We all agree with this, right? How would it come down to us as individual clinicians, or even a practice, or even a system? You know, we're a medium-sized system at Wake Forest Baptist Health. How would we translate that into our system? So I want to ask someone here. Can anyone here make a statement to me about the quality of care of the patients in their practice? Anyone want to make a statement, an assertion about it? Paige, can you tell us? I know you're doing-- you're delivering good care, I know, based on the way you're training. Can you tell me something about the quality of care of the patients in your practice? AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: I have absolute confidence you are. In reference to an external set of knowledge, formed a set of professional skills and opinions, and you're affecting those to the best of your ability. Would that be-- yes? AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: Oh, ladies first. AUDIENCE: [INAUDIBLE] so I tend to see them way downstream. And I don't know that it's [INAUDIBLE] but everyone tends to work as if they're in a silo. And much of our challenge is trying to coordinate communication, and should it be going upstream, and that kind of truly patient-centered [INAUDIBLE] RUSSELL HOWERTON: I've got a great slide for you coming in a minute. AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: I've got a great slide for you. Hold that thought. Yes? AUDIENCE: Well, I come from more or less a small practice setting, solo practice, and rural setting. And the metrics or dynamics are unique, compared to an academic, tertiary-care setting, or large single-specialty group. We just don't have the personnel to touch on each area that is becoming sort of the benchmark of quality care. So we have to select out what's most-- what's the highest priority. And as a physician in the field, trying to keep up with cutting-edge practice, I need a resource, like the [INAUDIBLE] guidelines, some sort of consensus, a resource, in order to know that I know my patients are receiving the same oncologic care as they would be if they went to a larger community. I can't participate in clinical trials. Clinical trials are another way to judge the quality of the practice. So we do a number of things. We don't have, let's say, patient-centered care, except in the fact that from the oncology standpoint, the treatment part, the diagnostic aspects, that's our first priority. And that the patients are satisfied. But we know from the results of the satisfaction survey that patients look at things like the time you spent with them, did you explain things in a way that was understandable, how long did it take-- how long did you wait for your appointment, and such. So it's very taxing on a busy practice, not only a solo practice, but to take on complete care [INAUDIBLE] patient-centered home approach, is a real challenge for community oncology. RUSSELL HOWERTON: Well, you might be surprised to know that even at scale, in a big health system, the actual resources to do this kind of work are a challenge. So these are control charts. These are time series plots of data. Probably most of you know what they are. I don't doubt that. And they lend themselves to analysis of the performance of a process. In general, if you can't make a statement about the patients in your care based upon data derived from the actual care of your patients, you can't really make a statement about the quality of the process you're delivering. You can reference point initiatives. Now, let me be the first to say, I'm at a major academic medical center. I got wonderful partners I work with. I cannot produce a graph like this about the patients I'm caring for. So it's not a judgment on anyone, that we're not-- it's very hard in our volume-based system today to build into ongoing care measurement of metrics that matter to patients, and then to find the time to gather in a group and understand and analyze them, and drive performance improvement. So if you can do that today, I want to commend you. Do any of my-- can you do that about any of your patients? I know I can't. But that is how you would understand the voice of your process, what it performs like, and be able to predict it to the next patient who might be in your system. At a system leadership level of a large system, it's kind of embarrassing to admit that that's hard to do at scale. How might we turn the lofty goals of that IOM report into real action on the ground? So you have to decide what you want to do and what can orient people around it-- diverse people with diverse professional and public opinions. Almost all of us can understand the concept of value, which is outcomes over costs. But that's outcomes that matter to patients. Those of you who've read Michael Porter, who's a writer about change in the health care system, will know that he groups outcomes into tiers. And most of the outcomes we measure are tier one kind of outcomes-- very hard outcomes, process measurements, short-term outcome measurements, all the way to tier four, which are deep, long, patient perceptions of life, quality of life. They're hard to measure. It takes teams to actually own that work. We either have to resource the people providing the care to have the time to do it, or resource layered-on resources. So we've got 13,000 people working at Wake Forest Baptist Health, and we struggle to find that, to produce those for the ongoing care of our patients. Not a study, not something we do retrospectively, but in the measurement of the process. A way to get started is simply to choose measurement as a strategy. Now, most of you will know there are a plethora of quality measurements. And there are pros and cons to almost every measurement. At some point we have to get over that and just decide we're going to pick a starter set. We're going to measure. We're going to share together the data, of which cannot be Lake Wobegon, right? We generally, as professionals, are a little bit uncomfortable in looking at data about performance that reveals variation other than everybody between 95% and 99%. But all of you who asserted care that you deliver based on reference points of good quality, almost certainly, if measured, are at variance from what you intended. Without an active process to listen to the voice of the process, to understand variation, and to drive it closer, it is almost a given, without knowing anything about you or your practice. This is hard. And left to our own devices, in a volume-based system, we are not in general doing it. How many of you-- so the outside world has recognized this, and they are changing. The payer community is changing. Is anyone here a participant in an accountable care organization? Yeah, it wouldn't surprise me. It's largely primary care-driven. Does anyone know what I mean when I say accountable care organization, a value-based payment model? Something that changes from payment per widget, or volume, to a payment based on value, which means you have to have an outcome or a quality measure over a cost. So interesting. I was getting ready for this talk. And yesterday in the mail, from one of the things that comes to you when you're in an administrative position, I got a display. So this is Medicare, our biggest payer. These are current Medicare payment innovations, different than fee for service Medicare. These are payment innovations where the payment model is at risk around equality in an outcomes study. So obviously you can't read it, but I think you can count. It's about 12 or 15. I hope most of you have heard the secretary state that they have an avowed goal to move 90% of their payment model by 2019 into value-based payment models. So what we aren't doing, currently, internally, they are getting ready to do externally with the claims data that is generated whenever we bill. So one of those value-based payment models is an accountable care organization, where a group of entities agree to attempt to treat patients over a period of time, and a benchmark of expected cost is derived. The fee for service claims are paid out during that time. And if the incurred costs have savings over what were the expected costs, the entity can share with CMS some of the savings. It turns out that-- matter of fact, things just like Dr. Clark mentioned, interventions to improve mobility that might reduce readmissions, sending someone to someone's home, to [INAUDIBLE] may have financial benefit if you can harvest those savings. And if you can predictably do that, you can build a system of adding those resources in, even though there's no RVU for that. There's no technical charge. There's no fee. You can hire those people to do that. Well, let's say you save millions of dollars in that model. There is a set of quality metrics that unlock those savings. You can't access that savings as a group participating in an accountable care organization, unless you hit the quality metrics. Now, we sat here just a few minutes ago and we were struggling to decide what those quality metrics are. And I can assure you, they don't derive from the Committee on Cancer, and they don't arrive from the American Society of Clinical Oncology. They come as close as they can, largely with claims-based data, because they don't, at scale, have detailed clinical data. So how does that relate to cancer care? So you might think, I don't-- that's a primary care thing. Those are accountable care. One of the demonstration projects, one of these-- this one, you won't see it right here-- is the oncology care model. I've been in medical oncology. You've probably heard about it. You may have heard about it. So that's a bundled payment around cancer care, where instead of the traditional medical oncology model of fee for service, physician visits, chemotherapy administration, whatever it takes, an entity can agree to get slightly reduced fee for service payment, and they get a per member, per month fee for caring for those patients, and an opportunity to share in the savings if the episode of care comes in lower than the expected episode of care. Now, that-- it hasn't gone live yet. That was-- applications were this spring, and there's going to be a pioneer group of practices that are going to start that this winter into next summer. Much remains to be determined, but the goal-- the avowed goal of that large payer-- is to incent people to think through how to commit resources and time to sit and gather, to have people to help with that data, to drive an improved performance. It will have exactly the same phenomena, though. There'll be a set of claims-based quality metrics that you have to hit to unlock the savings. So you might say, well, why does that matter to me? I didn't sign up for that. I'm still doing what I'm doing today. A different payment model was around a surgical procedure, joints. So Medicare spends a lot of money on hips and knees. And they have been incenting, in pilot groups, bundled payment for that. And for some years now, organizations around the country, in a demonstration project with Medicare, as well as multiple demonstration projects with private payers, have been delivering hips and knees on a bundled basis. They agree to one price for it, same sort of thing, quality metrics, cost, you get to share in the savings. It's had pretty wide uptake on a voluntary basis. No orthopedic oncologic surgeons in here, are there? So what drives success on that? Well, it turns out the number one driver of generating savings is largely around your interaction with your post-acute care provider. How many of you send patients, in any form or fashion, to skilled nursing facilities or rehab as part of your care model? Do you know the readmission rate for the skilled nursing facility that you send your patient to? No. You will when you're in a bundled payment, because it will directly impact the success of value to that. So if there-- does anyone here run a skilled nursing facility? Good. They come in several models of medical direction, one of which is a once-a-month visit. And if anything happens between that, call the ED. Take the patient to the ED, and they almost always get admitted. Others have a much more hands-on touch mechanism of medical oversight and medical direction. You might imagine there's a very different readmission rate from each of those. In bundled care for hips, knees, and joints, understanding that, figuring out the right resource in the post-acute care setting, is the primary driver of success. Most of the hard outcomes-- length of stay in the hospital, cost of the implant, DVT rate in the hospital-- those have all been driven to near national standard means. And there's not much variability in that. And it's built into the pricing model. How you interact with that post-acute care setting is what drives those models. Now interestingly, you actually need an increased initial dose. If you underdose the post-acute care treatment, the wrap-around care, you promote readmissions and failures. So it's often paradoxical what works to help. So back to oncology. How does this relate to quality of oncology care? So that's been going on for some years. You may or may not, if you aren't in this world, been aware that just a few months ago, unsatisfied with the pace of adoption, Medicare adopted a new strategy around hips and knees. So they believe it works enough in pilots that they named geographies, zip codes around the country, mandatory. So it happens that the Charlotte region is one-- anybody here from Charlotte? [INAUDIBLE]? We're not. Wake Forest Baptist, we're not in that geography. But Charlotte is. So someday soon in the future, you're in that model, whether you like it or not. Now, it takes a ton of work to get ready for that. You have to allocate time, as a team, caring for the patients, to sit and gather and work readmission data, look at run charts around that. We don't have that in our care model today. We would have been in big trouble if this was the geography where that was chosen. So this oncology care model is way upstream. It's where hips and knees and joints were some years ago. There are some years to go before it proves itself out as evidence. But the largest payers are going to change. And we, together, are going to have to pace our change in system to just about the same pace they do of finding resources and time to work data around our patients and be able to speak to those processes. I want to show you something that resonated with me as-- so, is there a way to make this come up with my screen view? Is there a way to just make my screen view come live on the screen? SPEAKER 1: I actually don't know how to do that. RUSSELL HOWERTON: OK, well, it possibly won't happen. So I'll go back here to my slide. Does that go back and work? So in the New England Journal-- oh, you made me go away. Just a few months ago, was something called instant replay, a quarterback's view of primary care coordination. Dr. Clark will know this. I will know this. Patient went to his primary care doctor, had kidney stones, CT scan, seven-centimeter mass characteristic of glandular carcinoma in the liver. Not many other major medical problems. Comes to a primary care doctor and sees that. There ensued 80 days between that visit and the date of surgery by a hepatobiliary surgeon to take that out. The live video of this, there's, like, 20 patient encounters with all these different specialists, 100 emails, 50 phone calls, interventional radiologist, biopathologist to get this patient to that outcome. So in a value world, that is what we as individual practitioners, grouped in clinically integrated networks, or health system leaders, have got to address. There is implied patient harm there. I wonder how much anxiety this patient had between each of those events, awaiting to find out if it was scheduled, not scheduled, accomplished, the results. You probably hear about that at your end of the care of these patients. And that is our shared challenge. I wish I could tell you that I stood here in 30 minutes and could give you an answer to that problem. But my purpose is really to just give you pause to think about what this would be like. AUDIENCE: [INAUDIBLE] RUSSELL HOWERTON: So in every health system discussion, when this comes up, a nurse navigator model comes into play. But to your point earlier about the challenges of your resources in your smaller independent practice, it is not scaleable at 13,000, to nurse navigate every problem. What we have to do is A, engage patients, B, open access, and C, segregate or parse the population to understand how to dose the additive resources, and largely alter delivering professional skills and asking patients to work around the structures we've found to deliver those skills, and instead orient our skills around a single patient's needs throughout the system. But it's a degree of flexibility and consciousness of patient needs that is not currently common in our industry. Now, you will be familiar with a different graph that I didn't put on here, of the spread in cancer care of curative treatment and palliative and symptom-based treatment. And a not-uncommon current strategy has almost all curative treatment, and then sort of a fall-off of that, and palliative treatment at the end. Much cancer care-- I'll just say, pancreas cancer patients could all benefit from a palliative care visit on the day of their first visit with a hepatobiliary surgeon. And it should be graded up. Figuring out a model to resource that and get that into place is our challenge. So if you ask me what system-level improvements in cancer care are, they rest on a foundation of wonderful, skilled teams working all of the individual metrics that would be in the National Cancer Database. But it is the challenge of building a system of self-examination and measurement for ongoing performance improvement based on data about the actual patients we are caring for. There are examples around the nation where this is happening, but it is by far not the common in our industry. And because it is not the common in our industry, the payers are innovating for us. And we are responding. Now, I could produce for you an exactly similar graph of models that private payers are doing. And many of you probably have spouses that work for major corporations in America, who essentially fund private payers. And I can tell you that they are similarly imposing these models on us. So I hope if I'm around to come back in 10 years, that we talk about this. Most of us will be in practice environments where we know data about our own individual patients that derived from the process of care that we were using those standard performance improvement techniques on that we saw at the beginning of the slide.
View Doctor Profile
