Gene Beed, MD
President and CEO
Horses, Zebras, and Unicorns
Gary M. Owens, MD
Gary Owens Associates
Glen Mills, PA
|STEERING COMMITTEE MEMBERS|
Al B. Benson, III, MD,
Professor of Medicine
Associate Director for Clinical Investigations
Robert H. Lurie Comprehensive Cancer Center, Northwestern Univ.
Ira M. Klein, MD,
Chief of Staff to the Chief Medical Officer
Aetna Oncology Strategy
Samuel M. Silver, MD,
Professor of Internal Medicine Hematology/Oncology
Associate Director, Faculty
Group Practice, University
of Michigan Medical School, MI
Roy A. Beveridge, MD
Chief Medical Officer
McKesson Specialty Health/The US Oncology Network
The Woodlands, TX
Jennifer Malin, MD,
Manager and Medical Director of Oncology
Los Angeles, CA
John D. Sprandio, MD,
Consultants in Medical
Oncology and Hematology
Craig K. Deligdish, MD
Oncology Resource Networks
Pharmacy Services SelectHealth
Salt Lake City, UT
F. Randy Vogenberg, RPh,
Principal, Institute for Integrated Healthcare
John Fox, MD, MHA
Associate Vice President for Medical Affairs Priority Health
Grand Rapids, MI
Lee N. Newcomer, MD
Senior Vice President of Oncology Services UnitedHealthcare
Approximately 200 oncologists, payers, employers, managed care executives, pharmacy benefit managers, and other healthcare stakeholders convened in Houston, TX, on March 28-31, 2012, for the Second Annual Conference of the Association for Value-Based Cancer Care (AVBCC).
The mission of the conference was to align the various perspectives around the growing need of defining value in cancer care and developing strategies to enhance patient outcomes. The AVBCC conference presented a forum for the various viewpoints from all the stakeholders across the cancer care continuum, featuring more than 20 sessions and symposia led by nearly 30 oncology leaders. The discussions focused on current trends and challenges in optimizing value in oncology by reducing or controlling cost while improving care quality and patient outcomes, introducing emerging approaches to management and tools that providers and payers are using to enhance cancer care collaboratively.
The AVBCC Second Annual Conference was opened by a Steering Committee discussion of 11 panel members who attempted to define value in cancer care and articulated action steps that can help to implement value into cancer care delivery. The following summary represents highlights from the Steering Committee discussion, which was moderated by Gene Beed, MD, and Gary M. Owens, MD.
Gary M. Owens, MD: The 2 target concepts we need to consider during this Steering Committee discussion are “value” and “action.” Our goal is to define what is value in cancer care, and how is that value different for a payer, a clinician, and, of course, the patient. But we also must include other stakeholders, such as employers and the government, who pay for the care. In addition, we want to develop concepts and ideas that are actionable, ones that we can ultimately transfer into true action steps.
One thing that complicates the approach to cancer is that it is not a single disease but a collection of very different diseases. Heterogeneity is the hallmark of cancer. At the same time, there are similarities in the issues that concern patients with cancer, related to costs and outcomes. Currently, the most robust area in the drug pipeline is oncology. There are almost twice as many specialty oncology agents in the pipeline as there are other specialty agents for all of the other diseases combined. So we are on this cusp of innovation. We are seeing innovation in cancer diagnostics, which parallels the types of therapies being developed, and we are now able to do testing to tailor treatment for certain genetic mutations and expression of genes.
We need to mold that innovation into what it means for patient care, what it means to those who pay for it, and what it means for those who provide that care. That is how we are going to open the discussion for the Second Annual AVBCC Conference.
Gene Beed, MD: We are going to discuss the changing epidemiology of cancer toward patient-centered medicine. This means personalized medicine and the economics of cancer care. For example, what does this mean in terms of innovation, or when cancer becomes a chronic disease? Our overarching mission today is to focus on the issues that are most important to promoting value-based cancer care—quality, costs, policy, regulation, and patient access.
What is value in cancer care? I can say that it is quality divided by cost, but those are not quantifiable terms that we can all agree on. What is quality, and quality in the eyes of whom? If there are barriers related to cost, then the question arises who will pay for the care. Are we just all going to pay more for healthcare, or spend less on other conditions? Are we going to indebt the next generation, stifle innovation, ration care? What are our options?
Perceptions of value differ throughout the healthcare industry. For manufacturers, efficacy, total patient outcomes, and unmet medical needs determine value. This is in marked contrast to how health insurance companies, employers, and regulators define value, which concerns longevity, quality of life (QOL), and cost. Cost is a key part of the value equation, regardless of what part of the healthcare industry is involved. Why is defining value in cancer care more challenging than defining value in other areas of medicine?
John D. Sprandio, MD, FACP: Value in cancer care is challenging for 2 reasons. One reason is the clinical situations in which patients present, depending on their state and their variable frame of reference in terms of their goals and their understanding of the situation. And the second reason is the cost of drugs. Cancer care is uniquely complex. This is not to say that in other areas of medicine cost is not complex, but cancer presents a uniquely complex situation.
Samuel M. Silver, MD, PhD, FACP: I agree wholeheartedly that there are 2 issues here. One issue is indeed the complexity of oncology. When I talk about qualityof- care issues to nononcology audiences, they say, “So what is the problem? We talked to our cardiovascular colleagues, and we have it down.” And, they partly have it down. If you have to treat a myocardial infarction (MI), that involves a single organ. It does not make a difference how the patient got the MI, only that that patient has MI. There are not 4 stages of the disease. It is not estrogen receptor–positive or progesterone receptor–positive; it is not HER2-negative or HER2-positive. It is not all of these various silos. The science of cancer is complex, with multiple diseases and with multiple presentations of a single disease.
Second, because we have so many silos and our cardiovascular colleagues do not, they can have hundreds or even thousands of patients in randomized trials looking at a single point. In cancer, we do not have these numbers of patients. Therefore, the science becomes more difficult for comparative effectiveness analyses.
It is very difficult to determine the degree of improved efficacy and total patient outcomes in oncology. It also depends on where and how we are looking. What does cost mean for the patient? Value is not about the total cost of management, it is about the cost that we bear personally. Some of us are only seeing a part of the costs, and some of us see more of them.
This also varies from patient to patient, depending on their insurance. The value equation changes depending on the outlook of just that one variable.
Improved patient QOL—whether we define that in terms of years of life or quality per se—is the metric that has to be the most important. And the next is total patient outcomes.
Craig K. Deligdish, MD: For the purpose of this discussion, I would disagree with my colleagues for a couple of reasons. We have been treating cancer for many years. I think we must ask, “What has changed during this time?” Why would the definitions of value, quality, or outcomes be different today from what they were 10, 20, or even 30 years ago?
The difference today is in the cost of treatment. The cost of treatment today is drastically different from what it was 10, 20, or 30 years ago. By contrast, I do believe that we have made a dramatic impact on the survival of patients with cancer. At the end of the day, one of the important questions that a physician and a patient need to answer is, “Do I have a curable form of cancer? Can the patient be cured?” And a second important question is indeed about QOL, which is even more important when the patient does not have a curable disease. That question must be answered in the context of defining “quality” and “value.”
Next we need to consider all of the treatment and diagnostic tools that are available to us today, as compared with those that were available 10 or 20 years ago. What have they contributed in terms of value, quality, and cost? We also need to talk about how we define efficacy. We need to factor all of this information into what ultimately is the definition of value. Is a diagnostic tool or a treatment efficacious if it extends survival by 1 or 2 months? Clearly, taking a pill that has few side effects and cures an otherwise terminal disease (eg, Gleevec) has high value, whereas a drug that costs $30,000 a month and allows us to live 1 month longer has less value.
In terms of value, the definition is to some degree based on whether value is being defined by the patient or by the payer. As we all know, it frequently is neither the provider nor the patient who is paying for treatment. Is the government paying, the employer, or the insurance company, or is the patient paying for the treatment? And this is tantamount to defining the issue of value. We must also consider whether it is going to result in an improvement in QOL, make the patient feel better, or cure the patient’s disease.
Al B. Benson, III, MD, FACP: I would agree that in terms of value, the central focus must be on the patient’s perspective. For patients whose disease could be cured, the goal is cure and living a long, productive life. I think that is how most patients would define value. Financially, this is very difficult from a patient’s perspective, because in many cases the patient is divorced from the outlay of funds that cover the true cost. Although we are talking about tremendous problems with uninsured patients, copays, and so forth, over time most patients with cancer end up in a situation that they would not experience in any other aspect of their daily life in terms of potential catastrophic costs and life-threatening illness.
Normally, if you purchase any product, you have to cover that cost. Within the overall complexity of oncology, value is somehow distorted for many patients, because they are not paying the full cost of the service. Another fundamental problem with value is that over time we are seeing significant technological advances in all aspects of care, whether it is in diagnosis, in imaging, or in treatments, that may not be optimally defined in terms of utilization yet can drive up costs and at the same time have the potential to extend life.
There is so much we do not understand about cancer biology, and, as the technology advances, we have been unable to study the technology at its maximum to truly define value. For example, for cancer surveillance, we now talk about survivorship and how we monitor patients over time. The literature is very sparse in terms of defining what is appropriate surveillance. In my area of expertise, which includes colon cancer, there have been some meta-analyses done, but that is the exception. Overall, we need evidence to decide how to maximally use imaging, how to define risk, which patients are at risk, and how to monitor patients over time. And if we diagnose a tumor recurrence, for example, what can we do to help that patient further?
The same is true for drugs—the way we study most drugs now remains empiric. We do not have the selection tools that we need to demonstrate true value. As we develop markers, we are certainly making some progress. We have seen this in diseases such as leukemia, breast cancer, and to some extent, colon cancer.
When we talk about value and cost, survival benefit demonstrated in clinical trials shows proof of principle, that this given intervention works, but we cannot define how it works maximally in each individual patient. Because of the way we design these large trials—in most of them we are giving treatment to a large proportion of patients who will absolutely not benefit from that treatment—we cannot select out the patients who, in fact, truly benefit from the therapy, thereby reaching the ultimate goal for the individual patient.
Roy A. Beveridge, MD: The topic of adjuvant therapy versus palliative therapy is directly relevant here. The value for these types of therapy is different. I do not know of many biopharmaceutical companies that believe that improved longevity of a patient is not very important, because everything related to drug design right now is related to making patients live longer. In terms of personalized medicine, identifying particular patient populations is among the major goals.
By contrast, the US Food and Drug Administration (FDA) has indicated a lack of interest in QOL improvements. That may be something that the FDA may aspire to in the future, but this has not been relevant so far to any major drug development approvals.
Dr Beed: From the payer’s perspective, how do you respond to what people have said so far about value, and how is it perhaps different from that of clinicians?
Dr Silver: Fundamentally, value relates to people agreeing that something has equivalent value; there is an exchange that goes on. But in cancer care, that equation does not work, because the consumer is not the payer. I may not agree that this has the same value as something else, but I am constrained, because I cannot set the prices. I cannot decide on the prices of treatment, and I am restrained by state laws that force me to pay for something, even though I do not think it has value. Therefore, we have a fundamental challenge in defining what value is in cancer care.
We have a challenge defining value in all of medicine. For example, there are new transaortic valve replacements that cost $35,000 for a valve—$60,000 total compared with $8000 for a typical valve replacement. These new valves provide an additional 18 months of life for a patient who has less than 1 year of life expectancy. We have the ability to negotiate those prices. But in cancer care, we do not have the ability to negotiate that, because there are typically monopolies, at least early on, with the release of a new drug and a new drug category.
There is hope that in the era of biomarkers there would be enhanced value, because at least we would not be spending money on drugs that were not going to benefit a patient. But with tumor biology that involves escape mechanisms, it is not clear that it is going to provide a long-term benefit. Even with the recent study that was published in the New England Journal of Medicine showing that there is a lot of tumor heterogeneity, within any given tumor there may be multiple tumor types. We just happened to pick one, because that is what we biopsied. So even the role of biomarkers is now coming into question, especially in relation to the current economy.
The challenge is that we do not compare the value in producing 1 year of life or a QOL year in the same way that we compare it to people with heart disease. We may pay $80,000 a year to a patient with multiple myeloma, which is now a chronic disease, and that person may live for another 2, 3, or 4 years, but we do not have discussions about this—what is 1 year of life worth? At the same time, we are making tradeoffs as a society, because in the state of Michigan we are closing schools and are not paying for teachers. This is a challenge for us as oncologists, because at the end of the day we have to make choices on where we put our money, and we are not having a dialogue about where we should spend those dollars.
Dr Beed: You raise a very good point. One way that oncology is different from the rest of medicine is that just a generation ago we did not even use the word “cancer” with our patients. This is very different from most diagnoses. We did not talk about cancer, and we did not mention the diagnosis. Now we want to talk about the value of 1 year of life, QOL, and what the patient’s out-of-pocket copay is.
Matthew Mitchell, PharmD, MBA: You asked what makes cancer different from other categories of management. From my perspective, some of the biggest issues are the regulatory and government mandates in oncology. Nothing comes close to cancer care, and when we tie that with the amount of money that is tied to oncolytics, that is where the difference is between oncology and other clinical areas.
Dr Beed: So cost has been the main change.
Dr Mitchell: Cost is what gets on the radar, but much of this is semantics, that is, the way the question is asked. We have to ask what is efficacy, or what is the treatment effectiveness in real-world outcomes. I am surprised that we are not discussing adverse events or safety issues in relation to value. From a management perspective, these issues have to enter into the quality of care and the value proposition. We may need to consider allowing coverage of products within a certain regimen even if the exact regimen was not studied in a specific cancer type. One example may be allowing the substitution of cisplatin/carboplatin therapies into a regimen in which both products have demonstrated efficacy in separate regimens, but one may produce fewer adverse events, leading to better patient experience and lower cost with the reduced adverse events.
Ira M. Klein, MD, MBA, FACP: My role at Aetna for the past several years has been to be the clinical leader for our oncology strategy, and in that position we have wrestled with the issue of value. I would say that the standard definition of value is quality over cost. I think we can define value in many different ways, but that is the basic premise.
Dr Beed: There is some disconnect between, “The value of an extra month of life in a clinical study does not show what this drug can really do in the real world in my hands,” versus “It’s not a free market. There is no ability to negotiate.” What is the value for the cost that is being paid with someone else’s money?
Lee N. Newcomer, MD: By the time our 5-year-old children graduate from college, their healthcare premium will be equal to the average US salary. Based on current cost trends, we are 1 generation away from spending our entire salary to cover healthcare, if we do not change something. We can have many discussions about how we ought to tweak this, but we have to make some changes in the system, and they have to be made in the next 2 or 3 years.
We need to make some decisions about whether, as a society, we can spend any time treating people with drugs that get a 1- or 2-month response, until we find out how those drugs work in a combination that gets us a much more meaningful response, such as 4 to 6 months or 6 to 8 months. And we have to decide as a society how much those months are worth.
We know that England has done that. It came up with the number of $40,000 per life-year. I do not know if that is the right number for the United States, and at UnitedHealthcare I will not be making that decision, but as a society we have to do that, because we simply cannot afford to continue the way we are spending today. It is not sustainable.
Dr Beed: In the Affordable Care Act, the discussion about comparative effectiveness explicitly says that a calculation of cost per quality-adjusted life-year shall not be used.
Dr Newcomer: Exactly, and the way that Medicare balances its budget is by simply paying the physician and the hospital less and less each year, because it is not bringing enough money in. Medicare simply lowers the reimbursement and tells providers that it is something they have to accept. Very few people can walk away from Medicare. But, we cannot take that all the way down to zero.
I have just attended a hospital board meeting to discuss reimbursement. Each year we have to find $30 million worth of new profits just to cover what Medicare is not going to pay us, because their fee schedule went down. We have to do that year after year, and now we are laying off people to get it done. That the government will not address this tells us how difficult the discussion is, but we cannot use their approach of down-pricing to zero.
Dr Beed: If the US government will not address it, who will? Where is the area of agreement between clinicians who are caring for patients with cancer and the payers who are trying to manage this pot of money?
Jennifer Malin, MD, PhD: My alma mater, the University of California Los Angeles (UCLA) has the dubious honor of being the most expensive academic center in the country, according to the Dartmouth Group. At the same time, California is in a budget crisis. The 3 primary state budget expenses are healthcare, education, and prisons. I do not see any efforts to try to rein in costs at UCLA, even though the educational mission of the university is suffering—our faculty and the entire staff have mandatory furloughs, essentially a 2-week salary cut per year. The cost of tuition has increased by 30% over the past year. Our public education system is in shambles.
Therefore, what we are talking about in a theoretical way is actually very real. It is happening now: 1 in 3 of us is going to get cancer, so what do we see as worth putting our money away for in terms of treatments? Realistically, being in the position of being able to not have cancer, but thinking about ourselves as potential patients, what is a meaningful improvement in QOL and in length of life that is worth mortgaging our children’s future for?
F. Randy Vogenberg, RPh, PhD: There are so many different perspectives in this issue. Looking at it from the perspective of the commercial payer, the self-funded plan sponsor, whether it is an employer, a union, or a municipality, they are all feeling the squeeze as well. The research that I have been involved with recently is demonstrating very clearly that employers in particular are very much engaged in this debate, more than many people realize. They are trying to grapple with the value question not only in economic terms but also with how the whole healthcare system that they are paying for is performing to deliver outcomes.
As one of the biggest areas that is so visible to everybody, cancer is a very emotional disease. Everybody has been touched one way or another by cancer, including me. And there are many dollars involved. Employers are concerned about cost, but their primary interest has been performance. There is a big gap between what we are paying for today and the results or the performance that we expect to see. It is not just about cost. Employers are looking for the performance that is related to the cost of therapy, whether it is a drug or a diagnostic test, extending life, or curing disease.
We are now also dealing with severe drug shortages. Employers are very interested in more of a 360-degree view of what is happening with the dollars that they are spending. That is unlike what we see typically with the Centers for Medicare & Medicaid Services (CMS) because of some of the congressional budget rules that CMS has to play by.
There are probably employers who are seeing this as an urgent issue. They are looking at a 2- to 3-year window, as well as a time for very rapid innovation and change, to try and shift not only the cost curve but also the performance curve among all of the vendors that they are paying for.
Dr Beed: We were just talking about the meaning of value and about how the demographics have changed. Personalized medicine, biomarkers, and resource allocation are where the rubber hits the road in many cases. Should we spend more money on cancer care going forward? Are we willing to take money from other diseases and spend less on them and more on cancer?
Dr Vogenberg: That is all part of the discussion. There’s a finite amount of financial resources.
Dr Sprandio: Cancer therapies are valuable if they are potentially curative. In the metastatic setting, there is value, and it is reasonable as a society to support 1 or 2 lines of therapy. Regarding biologics, especially biomarkers, if we can convince patients that we are using the best drug from the beginning, and the second-best drug for the second round of chemotherapy, we would have a much better chance of telling patients that there is no sense in doing the second, third, fourth, or fifth lines of therapy. But today we cannot say that. We can just say that based on a population and a probability, we are going to choose the best drug, but we do not know what is best for the specific patient’s tumor.
No one would question that we should be investing more in early detection when there is good evidence that makes a difference in outcomes. In metastatic disease, we have to ask questions about how long to treat.
Dr Klein: Because this is an epidemiologic discussion, we should point out that we do not have the granular data that we know how to get to the answers we are seeking. Clearly there is a lead time bias. One of the reasons that we do not know whether we are doing better than we were doing 20 years ago (except in some select diseases, such as chronic myelogenous leukemia) is that Medicare and most of the private payers do not capture the elements that we need to figure out whether we are comparing like to like. So we might have a diagnosis, but we do not have a stage. We do not have tumor histology. We do not have biomarkers. And we do not have performance status. Until we get that information, we cannot make an informed epidemiologic decision. Once that information is available, I think we will be able to make much better decisions regarding the value of treating a metastatic disease, and who is getting the best chemotherapy.
And if we are comparing like to like, then we just eliminated the lead time bias, because we can get rid of the disease through our ability to granularly bucket different patient categories.
Dr Beed: So when it comes to actionable things that need to happen in terms of value in cancer, I have heard about watchful worrying and the early detection lead time bias. You point out that we need to have more robust informatics that look at cancer, and that may integrate clinical trial data, claims data, and other data to get some ideas of what happens in the population. Dr Newcomer, what else would you add to this in terms of making value-based cancer care a reality, considering the increasing numbers of expensive therapies, better early detection, and better survivability that may be real or may be related to lead time bias?
Dr Newcomer: I want to underscore the need for starting to collect some registries, so we can truly find out whether your statements are true. One of the things I am struck with in my discussions with practicing oncologists is that they do not get the same response that was shown in the registration trials, because the patients are different. Patients entered into a registration trial have a performance status of zero. The patients we see in the office often cannot do half the things as patients who are enrolled in clinical trials.
Only about 2% to 3% of all patients with cancer enter into clinical trials. If we are going to accelerate the pace of making these kinds of value-based decisions with cancer, we need registries. I think they are very possible. At UnitedHealthcare, we have 35,000 patients now with breast, colon, and lung cancers, and we have the data that were requested. We have information on stage, genetics, and current status—everything but performance status. We are starting to try to mine that data now to see if we can find what delivers value, and what does not.
Dr Beed: How do we apply these to individual patients? How do we make sure we are using the treatments that are really of value? Is this just the art of medicine, or is there something more to it than that?
Dr Benson: This gets back to the point about how we expand our knowledge base in terms of understanding how to use these drugs for appropriate people. It is going to take many shifts in what we do. Clinical trials are designed to determine efficacy, to see whether the drug works. They are critical in that regard, but what we are increasingly arguing now is that we have to change the way we design clinical trials. Even if we do not have the correct biological profile, since we often do not know that at the beginning of the trial, we need to prepare an information database, including tumor banking, so that in the future we could go back to that database and tissue to explore links encompassing biological profiles, treatment, and outcome.
A good example of that is the KRAS marker in colon cancer. When some of the colon cancer trials were designed and the researchers were looking at markers, KRAS was not even one of the markers in the original design. But because the researchers had collected tissue, they had excellent clinical data and banked tissue from which exploration of markers could be pursued. They had a methodology that was consistent with the need to test the marker. We were therefore able to go back to the data and combine a number of different trials to come up with a marker that transformed the treatment strategy.
These databases are going to be increasingly important, provided that we are putting the appropriate information in the databases. One reason that the database concept is especially attractive is that we are moving away from an empiric approach to one looking for patient selection based on more precise criteria. We now know that most malignant diseases, because of tumor heterogeneity, are a collection of different biological profiles. And as we start subsetting, we are going to need larger populations to look at these small population subsets. For example, in patients with lung cancer, there is a marker that defines a 4% population that will benefit from a given intervention; in patients with gastric cancer, there is a marker that identifies 20% of the gastric cancer population that will potentially benefit from a targeted therapy. If we are going to do comparative effectiveness analyses in oncology with these databases, we have to make sure, in an ongoing way, that we are populating these databases with ongoing information as it evolves, so that we can look at these subpopulations with meaningful data over time.
It is also going to require cultural shifts in trying to convince more people to participate in clinical trials. These trials are going to be even more complex, because as the National Cancer Institute has recently stated at the American Society of Clinical Oncology (ASCO) GI Symposium, “No tissue, no marker, no study.” The point is that if we are going to truly demonstrate marked improvement in efficacy, we have to define the population that is most likely to respond to a given treatment. This changes the landscape in how we conduct oncology clinical research. It changes the landscape in real need for population-based databases. It asks less, “Does it work in the real world?” and asks more, “Will this enable us to capture the subpopulations in sufficient numbers of patients that we could actually show that there is efficacy and appropriate intervention for the subpopulation?”
Dr Sprandio: What we are getting at is that we are going to change the epidemiologic description of disease from a histopathologic diagnosis to a genomic or epigenomic diagnosis. That will give us much more potential, I think, to observe changes in outcomes.
Additional clinical studies that Dr Benson mentioned are critical to driving guidelines, and it is the oncologist’s job to follow guidelines. The question is how do we devise innovative practices to impact the changing clinical and economic burdens of cancer? What can we control in terms of the economics? We can control unnecessary resource utilization. At the core of what we have done in this model that is still evolving is that we have minimized unnecessary resource utilization on a consistent basis over the past 6 years.
Consider the National Committee for Quality Assurance guidelines; each of the guidelines and requirements include a quality of service and a value component. The guidelines help to focus on patients as a central target, streamlining the standardizing treatment, which is what we do as oncologists.
Dr Beed: In your oncology delivery system, how is my experience as a patient going to be different from another center or from the old way of care delivery?
Dr Sprandio: Patients notice improvement in service. They also recognize better coordination and communication, which are major concerns. The physicians are different silos. That is noticeable in the patient’s eyes. Access, engagement, and responsibility are important, and making patients understand that they have a responsibility to communicate and report information to their providers. The whole thing about patient-centered medical homes (PCMHs) is tracking chronically ill patients over time to be in a better position to avert the acute exacerbations of their symptoms or issues, and then not only provide a decrease in resource utilization but also to improve their QOL.
The PCMH was geared and developed for primary care, initially pediatrics, and then adopted by primary care. It is a model that also fits very well for oncology care. At my practice, we are now saving money delivering care this way. I run a 9-man practice, and we have shown that we were able to reduce our annual emergency department utilization per patient receiving chemotherapy by 68% since 2005. We have reduced our annual hospital rate for patients receiving chemotherapy by 51%, and we have reduced the length of hospital stay for admitted patients by 21%.
Dr Beed: Does that drive every insurer to your doorstep?
Dr Sprandio: Yes, it is starting to. We created many efficiencies; for example, we went down from 8.3 fulltime equivalents (FTEs) to 5.5 FTEs. There was a lot of internal infrastructure that we had to build, and that was costly. The programming was done by trial and error. There was instruction and hand holding. It cost a lot of money to develop this model, but now our practice has stabilized—we are still independent, and we are hopeful.
Dr Beed: Have the capital expenditures been laid out, and are you going to start seeing a return on investment?
Dr Sprandio: If there is no return on investment, it means that there is no payer response to this. Reform - ing payment methodology is harder than reforming care delivery.
Dr Beed: If somebody does something that decreases emergency department use and other resource use, it saves money in terms of patients with cancer. They spend some money, they save some money, and they have an ongoing process that’s saving money in caring for cancer patients. Does that mean that we should then, as a system or as a payer, be spending less on patients with cancer going forward and realizing those savings? Or does it mean that we should reward people for saving money and spend more on cancer, and then the ultimate burden on the system is negligible?
Dr Malin: One of the elephants in the room that we have not talked about is how oncology care is paid for, and the fact that most of the revenue for oncology practices comes from the margin on drugs. If we are going to make the decisions about which treatments to use on which patients as cost-neutral to the practice, we have to reward those innovative practice models and start to reimburse providers for offering care that is patient-centered and well-coordinated rather than reimbursing them for selecting the most expensive, highest technology– based novel therapy, if that therapy is not going to have a dramatic improvement in patient outcomes.
Dr Beed: What do you think about the way we pay oncologists, and the incentives that are a part of it?
Dr Mitchell: I am unable to think of another specialty that brings in revenue based on billing for drug therapy to the degree that is done in oncology. Part of the elephant in the room is the potential for treatment decisions, and more specifically medication selection, based on potential clinic revenue. One possible solution is to reevaluate the way cancer centers are reimbursed. One option would be allowing a management fee, similar to a medical home but more comprehensive, because the sophisticated centers do have interdisciplinary teams. Knowing that it costs more to deliver that type of care, this removes the revenue out of the drug side completely, and puts it into a management fee.
Dr Beed: We have mentioned personalized medicine: what do payers think of this approach? Are they going to only pay for people who have the biomarker? Does personalized medicine direct therapy, or do people just become a little more comfortable or a little less comfortable with what they have decided they are going to do? If all we have done is add the cost of a test, that really did not change anything. Is personalized medicine the Holy Grail? Are we going to test for biomarkers in everyone?
Dr Beveridge: One of the main problems is that with International Classification of Diseases, Ninth Edition (ICD-9) coding, the insurers do not know whether a patient with lung cancer had small-cell or non–small-cell lung cancer, localized versus metastatic, first-line therapy or fifth-line therapy. Payers were in the dark. In the next couple of years, ICD-10 will address some of that.
The inability to compare groups throughout the country has been a fundamental problem in terms of assessing quality in one area versus the other. Peter Ellis, MD, Kathy Lokay, and others first started using pathways, and then the US Oncology Network and McKesson Specialty Health have continued with pathways in the past 7 years. I suspect that some of the criticism that Dr Ellis remembers is that we (the US Oncology Network) were converting over to “cookbook medicine,” that all people were being treated the same, and that the goals of pathways were to render the least expensive care. As we look at the evolution of pathways in the past 6 to 7 years, we recognize that it allows for true personalization of care. Because the treatment algorithm does specify, and probably all pathways in the country do now, whether this is for a patient with breast cancer who is HER2-positive, HER2-negative, estrogenreceptor–positive, triple-negative, and so on.
But this is not the “be all and end all” as far as where we need to go. In terms of truly personalized care, the answer is more than just pathways. We need to begin to take into account radiation, which we have not yet discussed. Today, radiation is a significant component of our cancer care delivery system.
We keep talking about drugs, because most of us are medical oncologists, but hospitalization and emergency department rates are very important in terms of costs.
At some point we need to address the issues of end of life, corporate use of hospice, and futile therapy. These are all crucially important for the management of the patient. That is what we should discuss when we talk about true patient-centered principles and personalized medicine.
Dr Vogenberg: I have recently had discussions with employers about personalized medicine. One of the common responses I hear is, “If you cannot prove or show me that what you are doing is better than what we did before using a newer product versus an older product, then the only difference is cost.” All of these other quality parameters we talk about are considered to be irrelevant by the true payer (ie, the employer as benefits purchaser). We have seen this with CMS, and we are seeing it now with employers. I have never heard an employer say, “I’m not willing to pay.” All employers are willing to pay, but they want to see that this cost is making a difference. That value is going to be key to the focus of personalized medicine.
We are increasingly moving toward personalized medicine, which will be a large conundrum. A key issue will be whether we can differentiate between the performance of one product and all of the surrounding services.
Dr Sprandio: Ultimately we are going to get to a point where we are using genomic diagnosis. But with tumor genomics sequencing, any escape mechanism creates a significant challenge for the oncologist in trying to decide how to use that wealth of information to figure out which therapy is best. And there are probably going to be a number of tumors for which no markers exist.
Dr Deligdish: Payers think there is value in information related to clinical utility, if that information helps the physician to decide on an optimal therapy. But if this information does not help in the decision-making process, then that information has limited value. For example, if the cost of tumor genome sequencing is in the $2000 range in the next several years, and the cost of treatment is between $5000 and $15,000 monthly, the cost of genome sequencing may be a good investment if it impacts the decision-making process. However, if some patients can be cured by the information gleaned from this test, the value may be priceless.
We can most likely all accept that our current system is not sustainable. If we look to the past, at least in the area of cancer treatment, some fairly dramatic mistakes were made when it came to value and cost; one example is the use of high-dose chemotherapy and autologous bone marrow transplantation for women with breast cancer. That treatment was very popular and was even performed at tertiary medical centers and in the community setting. Indeed, billions of dollars were spent before it was recognized that the treatment had no true value and was associated with significant morbidity and mortality. Similar examples occurred in the past couple of years, when treatments that were promoted without proven benefit were approved and implemented on a fairly large scale. We cannot continue to make these types of mistakes, but we continue to make them.
As we move toward healthcare reform, we continue to see consolidation in the number of providers and payers. But as we continue with this consolidation, we are realizing that it does not reduce cost and does not make the system more efficient. Indeed, it makes the system less efficient and more costly. Unless we, as a society, are willing to address those issues, we will end up in a place where there probably is only 1 payer, because no other payers will be able to afford to be in that space, and patients may no longer be able to afford treatment.
We need to think about where we are going to be in 5 years, because, as Dr Newcomer said, we cannot afford to be where we would end up in 10 years unless some of these issues are addressed.
Dr Beed: If we are going to spend more money on cancer, how are we going to finance it? Are we going to spend less money on other diseases, hold down innovation, or borrow from the next generation? What are the appropriate steps, that could be accepted by providers, that we have to take as a society?
Dr Silver: I have heard at an oncology panel that we spend more money on testing now than we do on chemotherapy drugs and radiation.
Dr Newcomer: That is correct, and that is on genetic testing alone. However, genetic testing includes all diseases. When you do parse it down by cancer, it will come down. But, the point is that when we look at spending, it is clear that we are enamored with genetic testing, and we are spending large amounts of money on it. In fact, we are seeing significant spending for genetic testing in areas without evidence to support use yet.
Dr Silver: Hematologists have been working with these kinds of advanced personalized medicine tests for the past 20 years. These tests have been getting more specific and more advanced. But, it is much easier to obtain tissue in hematologic malignancies than in solid tumors, which is why this has been in the forefront.
Overuse in the very expensive diagnostic testing for personalized medicine typically takes place for several reasons. First, there are many perverse incentives financially for doing this. Second is the lack of impediments for use. For example, Oncotype DX for women with breast cancer does decrease the amount of adjuvant chemotherapy when used appropriately. However, we have reviewed cases when the Oncotype DX showed a high-risk score and the women did not receive adjuvant chemotherapy, or when women had a low-risk score and received adjuvant therapy.
One question that I cannot answer is who was ordering the tests. We can have an 85-year-old woman who should never get adjuvant chemotherapy, but if the test is being ordered reflexively by a pathologist (who does not meet the woman), is that reasonable? We end up spending $5000 on a test that no one is going to look at.
We have excellent tests, but they have to be used in the right context. We are ordering too many things, because we can do it. There are certainly overuse issues that must be addressed.
Dr Beed: Does the PCMH model address the need to keep the pathologists from ordering something when they have never seen the patient?
Dr Sprandio: It does. Among the goals of the PCMH model is assuming complete control of everything related to cancer diagnosis, orchestrating the appropriate response, and coordinating patient care.
Dr Beed: With regard to negative consequences of personalized medicine on drug innovation—should we continue to develop the same innovative products that only approximately 4% of the population will benefit from?
Dr Benson: One of the hallmarks of selecting a marker to become a clinical test is that in the end it must benefit the patient. Part of the problem with technologic innovations—whether it is a diagnostic test, an interventional radiology procedure, or now positron-emission tomography (PET) or magnetic resonance imaging (MRI)—because of the way they are being developed and get reimbursed, they enter the marketplace without the true testing that is required to understand their clinical benefit. We have been talking about drugs, but if we look at PET scans alone, the amount of money being spent on them is mindboggling, and most of it is inappropriate use.
There are important clinical trials addressing the use of PET scans, but they are lagging behind in defining the optimal use of this technology, which is also an evolving technology. Because PET scanning is readily available across the country, it impedes the prospective collection of the essential information needed to answer, “Does this work, or does it not work?”
Part of the perversity of our reimbursement system is the way these products get on the market. They are paid for before we clearly define the appropriate clinical use. We need to make a better investment and say, “We are going to order this test, but the only way the test is going to get reimbursed is that it will be entered into a database or a prospective clinical trial, to try to define what the real use should be.”
How to get around that is a major problem. For example, I was on the ASCO panel that was going to create a guideline for radiofrequency ablation use for patients with colon cancer and liver metastasis. This is a procedure used every day around the world. It has been used for years. In doing a systematic review, we found that there was not even 1 prospective randomized trial on this. It was, therefore, impossible to create a guideline, and we were left defining a group of research questions that are yet to be addressed. But there are currently no incentives for these type of trials to be done; there are certainly no reimbursement disincentives.
Dr Beed: When a new diagnostic test or a new therapy becomes available, we would expect that the coverage will be contingent on these data points, so that we get an answer and can revisit it with real data. Does that appeal to an insurer? Is it doable?
Dr Klein: In many ways, CMS has punted on much of this by saying, “We will give coverage with evidence development.” But the majority of coverage decisions are made at the local or the regional level. Therefore, from a national perspective, CMS has not necessarily taken the bull by the horns.
At Aetna, we have tried to approach some of the more egregious misuses of technology through clinical policy bulletins and sometimes on claims table edits. Nobody likes to get a precertification for an MRI. And yet, because we have collected systematic data, we know that radiology benefit management works and formulary management works. Benefit design changes do drive behavior on the member side.
We have already been doing this, but not enough. If we (ie, payers) are always being looked at as those who are not doing something or are withholding something, then it is hard to get our message out that we are really trying to address overuse. As with the Oncotype DX, for example, there is overuse, underuse, and misuse.
There are 2 ways we can reduce cost. First, we can make the individual practices more efficient as entities that have a product. Second, we can better connect the system with data. The discussion about overuse of genetic tests speaks to a fragmented system in which nobody talks to one another. If we all talked to one another, we would have much less misuse and overuse of testing, because the communication will solve these problems. Those are 2 areas where we may have a glimmer of hope in reducing the cost of cancer care.
Dr Beed: Do we have mechanisms to encourage collection of data and connection of practices?
Dr Vogenberg: Yes, companies have gotten involved very rapidly, particularly the larger Fortune 500 companies. They are looking at the sharing of information that Dr Klein talked about, and are trying to promote better use of information technology, as well as linking it to benefit design. From the employer’s perspective, all they can do is look at changing some of the coverage parameters through their benefit design. They are not going to get involved, and they are not going to “touch” patients. But they are trying to drive many types of behavior.
There are gaps in their success; up until a few years ago, most employers did not pay attention to any of this, because drugs were cheap. Now that drugs are expensive and are getting more expensive, employers are now looking at the cost of healthcare.
What they still do not see is the spending on diagnostics and devices, which is going to force even more interest around the integration of information and a better coordination of care delivery. We are seeing that with some of the larger employers.
Dr Beed: What will the reimbursement mechanism be for employers who invest money and provide better care?
Dr Vogenberg: We are back to how employers perceive performance of their vendors—the health plans and other providers of services. It is rare to see a very clearly differentiated performance, which is why I think we keep going back to this cost issue. Some employers are trying to drive the discussion, as well as partner and collaborate with their plans and with their other vendors to encourage that kind of innovation. Employers are looking to drive innovation with their partners.
What has changed is the cost. Whether it is the patient or the employer as a payer, the dollars at stake today are so significant that it makes it worth their while.
Dr Beed: How does cancer prevention fit into our valuebased care discussion?
Dr Vogenberg: Employers have invested quite heavily in wellness and in preventive services. They are now “getting it,” and they are seeing the return on their investment in employee satisfaction and in the difference in employee performance and productivity as a result of catching disease earlier.
Dr Beed: Are we going to prevent cancer? Is that something we should be investing our money in?
Dr Beveridge: I do not think that medical oncologists see patients early enough to prevent cancer. The numbers of my patients who start a “brown rice and broccoli diet” after they have had colon cancer are quite a few, but they started it approximately 25 years too late. Breast cancer genetics programs are going to be increasingly widespread. I think we should continue to invest in some expensive tests in the relatives of our patients with breast cancer as a preventive measure.
Dr Malin: I would not say that prevention focused specifically on cancer is how we are going to bend the cost curve; rather, the issue of obesity, and it is affecting not only cancer prevalence rates but also all other diseases.
Dr Newcomer: I could come up with a list of 10 things that we could get rid of without hurting the patient. For example, 20% of our patients with stage I breast cancer get PET scans. Why? There is no reason to do that, as Dr Benson mentioned.
I would look for that “list of 10,” and then try and make sure that that is consistently applied throughout my entire practice, that every one of my partners signs off on this list, and that we monitor it carefully.
We will pick our 10 habits, and get rid of them. As soon as we are consistent on those 10 things, we will move to the next 10 things. We have to start finding things today that we are going to stop doing, so that we could leave some funds for therapies that truly work. The last thing any of us wants to say to a patient with Hodgkin lymphoma is, “There’s not enough money in the system to treat you anymore.” That would be a travesty.
There is plenty of waste in the system that we could and should start eliminating, and that budget could be very viable.
Dr Owens: No matter where we went in our value discussion, the elephant in the room remains cost. Every road leads back to cost. So the reality is that value and costs are inextricably linked together. We cannot take them apart. But one of the things that we can do vis-à-vis action is take waste out of the system. Another thing we can do is change the way care is delivered. The only true mistake is to keep doing the same thing we always did, and hope to get a different outcome.
Dr Beveridge: It is worth noting that Dartmouth College has a new institute, the Center for Health Care Delivery Science. Its goal is to change the way we think about the importance of the actual execution of care. The president of Dartmouth College has been quoted as saying that the real rocket science in medicine over the next 10 years is in the delivery of, and in the execution of, care. This is true. It is one way of wringing out the waste.
Dr Owens: The other piece of this is certainly changing the payment methodology. In oncology in particular, a lot is driven by drug cost and the way we reimburse for drugs. To change the system, we have to change the way we pay for it. That is another action item. We cannot be afraid to experiment with new payment methodologies and truly change them.
John Fox, MD, MHA: One of the things that Dr Beveridge brought up was the need to have processes for advance care planning. The National Comprehensive Cancer Network (NCCN) guidelines and the ASCO guidelines document the importance of advance care planning and trying to understand, through conversations with patients, what their goals, preferences, and priorities are for care at the end of life. Yet we don’t do advance care planning, not because we do not think it has value, but because it is uncomfortable, and we do not know how to integrate it into practice. The hard work this requires is not rocket science, just social science. Oncology practices need to achieve consensus among the oncologists, develop the competence to have these discussions, and implement them through a process into the practice.
Waste, and often harm, occur when we provide treatments to patients that they would not have wanted had they known all their options. There is compelling evidence that having advance care planning discussion improves the patient experience in the healthcare system, improves health outcomes, and reduces cost of care—not because of denied care, but because we have provided the care that patient truly wanted.
Dr Silver: My practice is an outlier, in that many of the patients I see have very low-prevalence diseases. Therefore there are no guidelines, and there are no randomized trials for these diseases. In addition, because I am in Michigan, many of my patients work for ERISA corporations, and their insurance company is an administrator. Because the insurance says no, and because there is no evidence for the treatment, when I see a patient I call the human resources (HR) department of that patient’s company to check on insurance. We end up having a good conversation about their employee, about what the employee needs, and about this rare disease. The HR department, the patient, and I feel good about the next step, but I think that is unusual.
Dr Newcomer: At UnitedHealthcare, we are piloting a 6 medical group experiment that starts with an episode payment. We have created an episode payment that covers all of the patient care involved. As the physician group gets more effective in improving an outcome or in reducing the total cost of care, their amount of the episode payment will increase. And we are trying to reduce cost by having one third of the savings go to the group, roughly one third goes to UnitedHealthcare, and one third goes back to the employer in the form of a lower premium.
By now we know that we can get enough data to start figuring this out. The 6 groups just met recently. We looked at 64 measures of cost, quality, and operational efficiencies and came back with some projects that we think will begin to reduce care in a number of areas. The reason I mention this is that it is an example of a payer sitting with 6 medical groups, and I never saw a finger pointed at someone the entire day. The questions were, “I wonder how we can fix this. We saw this problem in the data. Now, how can we solve it? What do you do over here? How did you get such good results? What are you doing here that maybe makes you so variant, and how can we bring it down?”
The point is that we are all going to have to roll up our sleeves and probably work together. I do not think any of us can do this alone. We need to start more of the little projects that we have heard all day about—literally a dozen different attempts at trying to do this, but we need 100. Some of them will fail. A lot of them will succeed. But we are going to have to start faster and begin to collaborate and look at each other as partners. We need to figure out how we can get it done.
Dr Beed: How do we integrate survivorship services into the definition of value?
Dr Benson: This is a fundamental element of personalized medicine. It is not just the markers. The NCCN, for example, has created a partnership in terms of employers. What are the fundamental aspects of a cancer program that truly provide comprehensive care? We have talked about the absurd way drugs have been reimbursed for oncologists, but that has created the margin so that programs could integrate many of these other services that are otherwise not paid for. One of the enormous challenges is how do we fairly reimburse these essential services that are hallmarks for appropriate care?
Another challenge is access. Someone who is in a rural community, for example, may not even have an oncologist. How can we make these services available to much larger populations? It is one thing to be a patient at my cancer center, where we have all of these services on the same floor, so the day patients are being seen, they can get insurance counseling, nutrition counseling, or psychosocial support, for example, if they need them. We can quickly get all of that integrated, but that is not true in many situations.
A free-standing practice may not be able to provide all of those services right in its own facility; this gets us into the referral mechanisms. It is also an issue when we talk about accountable care organizations and bundled payments for care. Not everyone needs all of the services we may have available, such as nutrition services or psychological support. But how do we make sure, when we are talking about episodic care reimbursement, that we are integrating all these other essential services that we now have to pay for, often in a way that comes from philanthropy or from other sources, but not because we are billing for these services?
Dr Beed: Is this just another sign of the toxicity of our à-la-carte system of billing?
Dr Malin: This gets at the issue we were getting at earlier of early detection versus surveillance. When we talk about survivorship care, we need to think about who are the people at risk in the transition. It should not be one-size-fits-all survivorship care anymore than it should be for active treatment. We need to think about when to integrate people back into primary care, who are the people who truly need the extra support, and what are the best models to deliver that kind of care.
Dr Sprandio: One of the things that would be very helpful is if we all agreed on a core set of services that are inherent to providing good patient-centered care that everyone would reimburse. That is not restricted to cancer only. Any patient who has a chronic disease would benefit from that. There is no patient mechanism for that today, because there are no margins on drugs.
That is one challenge, and I agree that when we look at whole person care, there are a set of services that we could all agree are important to that patient’s care but that we do not pay for today.
Dr Beed: Can anybody argue that there are not significant dollars to be saved by removing futile care?
Dr Beveridge: We have had 9 pilot studies in the US Oncology Network looking at end of life. We have piloted putting palliative care physicians in practices. We have trained nurses in palliative care. We have trained nurse practitioners. We have a large number of palliative care physicians in our 1400-physician network at this point. But the common theme is—when providers can speak to the patient with metastatic disease early, which our data clearly support, there is a higher use of palliative care; there is a greater use of appropriate hospice treatment; the days in hospice increase significantly; and the number of days using a drug therapy decrease.
The number of physicians trained in palliative care in the country is so small that they cannot even begin to take up the slack. The average reimbursement for a palliative care physician across the country is less than $210,000, and on the East Coast, in Washington, DC, in particular, the average compensation for a palliative care physician is less than $145,000.
The other thing we have not talked about is the waste that exists when we continue chemotherapy in the face of progressive disease. I have had many oncologists say, “The patient wants to continue treatment, even though their disease has progressed.” Is that a waste, because we are not willing to have a conversation with the patient?
Therefore, not only do we need guidelines on disease surveillance and disease progression, but we also need clear guidelines on when to stop therapy, what to do when patients progress, and whether we should add fourth- or fifth-line therapy.
Dr Deligdish: We do not need guidelines about treating or not treating a patient when a patient’s disease is progressing. This is intuitively obvious, and it is a very small minority of oncologists, I hope, who are treating patients with the same treatment when a patient’s disease is progressing. If we cannot accept that, then the problem is that we do not have access to data that can demonstrate that either way.
We need to be very clear on this; we have talked a lot about advanced illness programs and a little bit about palliative care programs. It is very important to differentiate between these 2 entirely different approaches.
Advanced illness programs have been around for more than a decade. We have all heard about the number of successful pilots that have been conducted. However, there are many examples in which payers have adopted these programs, but they have failed to educate patients or providers as to their availability. In these situations, the programs are not utilized.
Palliative care is quite different. Palliative care can extend survival in patients with cancer, in addition to enhancing QOL.
Call to Action
In closing, the amount of information that we covered is very impressive. We did our best to define value from the different stakeholders’ points of view, and we decided it is inextricably linked to cost and action.
Based on this discussion, I have formed the following 7 actionable steps needed to develop a highly valuable cancer care system: