Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 Improving Value in Oncology Practice: Ways Forward VALUE-BASED INSURANCE DESIGN: INITIATIVES OUTSIDE OF ONCOLOGY Dr. Michael Chernew of Harvard Medical School described value as a reflection of both cost and quality, making clear that value is not synony- mous with either high quality or low cost alone. Measures of value and value-based initiatives must consider both jointly. Value-based insurance design (VBID) focuses on patient incentives and charges patients less for services that are of high value. The basic idea is simple: when we know a health care service is beneficial and of high value, we want to put as few financial barriers as possible between it and patients who it will benefit. By the same token, patients should bear a greater financial burden for care that has marginal value so as to encourage efficiency. VBID recognizes that patient demand and preferences should play a role in the care delivered and that patients should share some of the financial burden, but it also recognizes that standard economic demand theory should not be blindly applied. In this way, VBID is a hybrid of economics and incentive design, with the recognition that the health care market does not operate according to standard market rules. Also, it is important to distinguish VBID from value-based purchasing, which tends to focus on the way providers are paid and incentivized (Chernew et al., 2007, 2008; Fendrick et al., 2001). 85
86 assessing and improving value in cancer care Insurance Theory Insurance lowers prices at the point of service and alleviates risk. The problem is that insured consumers buy services they would not otherwise buy if they were fully informed and had to pay the full price. This problem is termed moral hazard, and it drives people to consume excess care and leads to high premiums when patients are insulated from the full costs of care. Models of cost sharing should not be designed to lower premiums but rather to improve patient incentives and reduce excess use while encourag- ing price shopping. Optimal insurance would balance moral hazardâs ten- dency for overuse and risk aversion, meaning that it would reduce co-pays in situations where the benefit of care justifies the expense and increase co-pays for care whose value does not justify the expense. This encourages patients to consume services that are high value and reduce use of inappro- priate services. But patients do not respond to cost sharing as economists would like. Instead, greater cost sharing leads to patient reductions in the use of appropriate and inappropriate services alike, and this leads to worse outcomes (Siu et al., 1986). Dr. Chernew reviewed VBID in areas of medicine other than cancer. An important dimension of a VBID program, he said, is the services it tar- gets for lower co-pays, such as important medications for chronic illnesses. Pitney Bowes has become the model for targeting chronic illness services (Freudenheim, 2007; Fuhrmans, 2004). The University of Michigan has gone one step further to discount co-pays not just for important services such as diabetes medications but also for belonging to certain patient pro- grams. A second important dimension of a VBID program is its scope: will the program simply lower co-pays for high-value services or also raise co- pays for low-value services? It is important to recognize that VBID tends to cost more money if co-pays are only lowered because it leads to greater use of services and a greater employer share of spending for high-value services that would be used whether or not VBID were in place. Advocates of VBID say that we can pay for VBIDâs extra spending because it leads to fewer adverse events and emergency department visits. The question is how many fewer adverse events will we have? VBID design will be more successful and the cost of VBID can be further offset if the patients targeted are at high risk for adverse events. The cost of VBID is further offset by benefits in patient productivityâimportant benefits that should not be overlooked. Another way to offset the costs of VBID is to increase the co-pays for low-value services or to implement a relatively small increase in cost over all services that are not deemed as having a high value. Dr. Chernew recalled a quote
improving value in oncology practice 87 from David Meltzer that said, âIf you are only doing VBID in situations where it saves money, you are not doing enough of it.â We want to provide high-value care, and high-value care will not necessarily save money. Results from the Literature on Value-Based Insurance Design In a study of the impact of VBID on adherence to medications for chronic illnesses, co-pays for ACE inhibitors, beta-blockers, blood glucose control medications, statins, and inhaled corticosteroids were reduced sig- nificantly by a large employer. Co-pays for generic drugs were reduced to $0 from $5 and brand-name drug co-pays were reduced by half. As a result, adherence to all of the medications improved significantly (Figure 8-1). After implementing VBID, Pitney Bowes claimed to save 6 percent in overall diabetes costs, or over $1 million total (Mahoney, 2005). A VBID program from Asheville, North Carolina, reduced annual diabetes costs per participant by $1,200 to $1,872 (Cranor et al., 2003). Another program for public employees in California increased co-pays and found that they were able to save 20 cents for every dollar spent on extra drugs; the program 14% Decrease in Non-Adherence 12% 10% 8% 6% 4% 2% 0% Ace/ARBS Beta Diabetes Statins Steroids Blockers Drug Class FIGURE 8-1â The impact of value-based insurance design upon adherence to ACE inhibitor, beta-blocker, glucose control, statin, and inhaled corticosteroid therapies. SOURCES: Chernew presentation, February 10, 2009; Chernew et al., 2008. R01506
88 assessing and improving value in cancer care was also able to save up to 50 cents for every dollar among a subset of less healthy individuals (Chandra et al., 2007). In addition to increased use of high-value services, these studies show that cost savings to patients and employers are possible using VBID. Dr. Chernew said that he believed a well-designed VBID program could save money from a societal perspective as well, especially if it raised co-pays for low-value services. The productivity gains and savings from reduced disability would give even greater savings. Furthermore, better targeting through personalized medicine would have the potential to save money as well, especially in cancer care. Applying Value-Based Insurance Design to Cancer For VBID in cancer care, Dr. Chernew suggested that the first dollar co-pay be waived. Charging for the first dollar of services only serves to tax those unfortunate enough to be diagnosed with cancer. Instead, increase cost sharing on treatments that are of less value. Next, Dr. Chernew sug- gested that co-pays should be kept low for appropriate cancer screening services. In addition, if low-value care services can be identified, cancer patients should be charged more for those services. Finally, we can gain considerable value by pursuing personalized medicine, and there is great potential for personalizing cancer care. Health care costs cannot grow at the rate they have grown historically. The question is how many people are going to be adversely affected when we slow health care spending growth? As a society we should make sure to slow health care growth in a way that minimizes adverse consequences and allows people access to high-value care. If we do not have the courage to determine what has high value and what does not, we will simply discour- age use of services indiscriminately, and that would risk driving patients to lower-value care instead of promoting health care that is of high value. GENERATING EVIDENCE OF VALUE POST-FDA APPROVAL: IS THERE A ROLE FOR HEALTH INSURERS? Reducing Uncertainty to Improve Value Dr. Tunis explained that his presentation would be focused on under- standing uncertainty in clinical benefits, a topic infrequently discussed while considering value. Though value can be described in many different ways, every description rests on the existence of some measure of benefit
improving value in oncology practice 89 to someone, whether through improved outcomes, hope, opportunity, or innovation. Furthermore, every measure of benefit has some degree of uncertainty about the magnitude of the benefit. How confident we are in what we know about the benefits of medical care directly affects how we talk about value, but too little consideration is given to our uncertainty. Whether in extrapolating overall survival from time-to-progression data or in determining the precise extent of clinical benefit for cost-effectiveness decisions, reducing this underlying uncertainty around clinical benefits can help us assess the value of care. We also fail to get the best value from our health care system in part due to uncertainty, Dr. Tunis said, and we avoid discussing just how little we actually know about things that are important in care. Critical evidence gaps are common. In 2007, the number of randomized controlled trials (RCTs) published in the medical literature was 18,000. Despite all of this evidence being generated, the conclusion of every systematic review seems to be the same: the available evidence is limited by poor quality and further research is necessary. With 18,000 trials published every year and every review concluding that the evidence is limited, one begins to wonder how we could miss the target so often. How is this possible, and why does it happen? One would expect that we could find good evidence at least once in a while. We all recognize that there are information gaps in health care that lead to uncertainty and problems improving value, but, Dr. Tunis joked, one cannot really explain anything these days without describing its molecular basis. Dr. Tunis then shared his discovery of the molecular basis of uncer- tainty (Figure 8-2). Turning to Figure 8-2, Dr. Tunis explained that the uncertainty path- way begins with intellectual curiosity that feeds clinical research, which in turn produces published studies and information that slowly travel by passive diffusion (knowledge translation route 1 [KT1]) to reach policy decision makers inside the cell cytoplasm. Alternatively, health technol- ogy assessment can act as an active transport mechanism (knowledge translation route 2 [KT2]) to deliver information to (intracellular) policy makers. Because the evidence that reaches policy decision makers is often limited and imperfect, there are many gaps in the evidence, but a defective knowledge translation mechanism (KT3) prevents the information needs of decision makers from reaching the clinical research enterprise. As a result, we have a clinical research enterprise trying to push information into the cell but no feedback loop to transmit what decision makers really need to
90 assessing and improving value in cancer care Low affinity receptors for decision makers Low affinity receptors for evidence INTELLECTUAL CURIOSITY CLINICAL RESEARCH Defective ENTERPRISE transport KT3 PUBLISHED GAPS IN EVIDENCE EVIDENCE KT1 Slow diffu sion HEALTH DECISION KT2 TECHNOLOGY MAKERS ASSESSMENT Active transport Policy, decision making Scientific evidence FIGURE 8-2â The molecular basis of uncertainty. SOURCE: Tunis presentation, February 10, 2009. know back out into the research enterprise. The communication difficul- ties are compounded by low-affinity receptors for evidence that surround policy makers and low-affinity receptors for policy makers that surround clinical research. Clearly, what we need is a targeted therapeutic to reduce uncertainty, Dr. Tunis said. How do we create a better link between the evidence-generating enter- Figure 6-2 new portrait prise and the policy makers who use that evidence to make the informa- tion delivered more responsive to existing needs or gaps? The Centers for Medicare and Medicaid Services (CMS) has a lengthy history of involve- ment in generating better evidence for policy decision making, from the investigational device exemption for coverage of certain devices in clinical trials (1996), to coverage for routine costs for patient care in clinical trials (2000), to ad hoc efforts to work with the National Institutes of Health (NIH). These efforts and many others reflect Medicareâs transition from an agency that simply paid bills in the past to one with a much more important role of informing decision makers with evidence. CMSâs definition of what is clinically reasonable and necessary also drives research focused on policy decision making. For decades, Medicare had no standard definition for what it meant for a health care intervention to be clinically reasonable and necessary. In the year 2000, CMS began to
improving value in oncology practice 91 define reasonable and necessary health care as that which has adequate evi- dence to conclude that the item or service improves net health outcomes. This was adopted on the basis of recurrent use, and it marked the first time that services CMS paid for would have to be supported by evidence that they improved outcomes. CMS then went further to say that the outcomes improved should be those that patients experience, such as quality of life, functional status, and so on. In effect, this explicit definition of reasonable and necessary health care also strongly influences how product developers and clinical researchers generate studies, pushing them toward reducing uncertainty around the issues that decision makers and patients actually care about. Building Evidence for Health Care Technologies and Services After FDA Approval The accretion of information and evidence for a technology or service occurs in a continuous fashion over time, Dr. Tunis said. From the time of its discovery, there is a certain amount that one knows about the technol- ogy by the time of its Food and Drug Administration (FDA) approval, and, it is hoped, more is known by the time it is covered by payors. But even then there remain many unanswered questions about the technologyâs long-term effects, risks, and other outcomes that will not be apparent until the technology becomes standard of care. The trouble is that further studies of a particular technology are discouraged once Medicare and commercial payors have made the decision to cover the technology. How then can we efficiently continue to learn about technologies and reduce uncertainty after insurers decide it should be paid for? Coverage with evidence development (CED) is Medicareâs attempt to reduce uncertainty by linking reimbursement for emerging, promising technologies to a requirement for prospective data collection on those tech- nologies through clinical research studies. Medicare retains the authority to approve the design to those clinical studies based on whether they are suf- ficiently robust to allow Medicare to decide whether the technology is clini- cally reasonable and necessary and improves health outcomes for patients. For instance, CED was applied to off-label uses of drugs for colorectal cancer and stipulated that Medicare would only pay for the drugs if patients were enrolled in National Cancer Institute (NCI)-sponsored clinical trials. CED has also been applied to other emerging technologies, from FDG-PET (fluoro-D-glucose positron emission tomography) scans in oncology to
92 assessing and improving value in cancer care implantable artificial hearts as an alternative to heart transplant. The results of CED have been modest, but there remains considerable policy interest in the CED concept. Take for example CED for implantable cardiac defibrillators (ICDs), devices implanted in the chest wall to deliver an electrical impulse in the event of sudden cardiac arrest. While each ICD costs roughly $50,000, only about 20 percent of implanted ICDs ever need to fire. Coverage for ICDs was expanded in 2005 only for patients enrolled in a large national registry, and data submission to the registry was required for ICD reimbursement through Medicare. The number of patients in the registry has grown to 250,000â300,000, and there is hope that studying the registry data will help identify high-risk patients and predict the 20 percent of ICDs that will ulti- mately fire. In order to risk-stratify ICD patients in the future, the research- ers will need baseline data (present in the registry) and follow-up data on which patientsâ ICDs fired. To date, the follow-up firing data has yet to be collected, chiefly because there is controversy about where this data should be stored. This is just one example of a hurdle that CED has faced. CED has faced a number of other challenges. The first is timingâ when coverage is under review, it may be too late to begin discussing CED. Second, it is difficult to design randomized studies for a treatment that is covered by insurance. Third, while payors see CED as an opportunity for coverage expansion for certain products, manufacturers see the oppositeâa coverage limitation. Fourth, it is unclear how to fund the research CED promotes. Dr. Tunis emphasized the importance of removing CED con- siderations from the context of any specific coverage or payment decision and beginning the CED from early in the evolution of a technology. By the time the technology is considered more than just a promising new treatment modality, it is probably too late to restrict its use to clinical trials. These challenges are not insurmountable, said Dr. Tunis, and there is work underway to build evidence for emerging technologies already on the market. Groups such as Dr. Tunisâs Center for Medical Technology Policy (CMTP) and others are pursuing this work in areas where treatment methodologies have rarely been studied side by side, and significant gaps in clinical knowledge still exist, such as in the treatment of clinically local- ized prostate cancer (Wilt et al., 2008). The CMTP convened a variety of stakeholder groups around FDA-approved prostate cancer treatments and was preparing to study the various modalities head-to-head to gain better evidence of comparative effectiveness. Understanding value depends on reliable information, Dr. Tunis concluded. To get this reliable information
improving value in oncology practice 93 and improve value, we are going to have to find better mechanisms for generating the information that policy makers are looking for. COMMUNICATING EFFECTIVELY WITH CANCER PATIENTS ABOUT THE BENEFITS AND RISKS OF CANCER TREATMENTS On the one hand, communicating with cancer patients about the risks, benefits, and costs might seem easy. Physicians can tell someone how an intervention might influence their survival and their quality of life and can present the risks, including financial costs, so that people should be able to decide how to weigh all of these risks and benefits to arrive at a decision. It seems straightforward, but clearly it is not, said Dr. Peter Ubel of the Uni- versity of Michigan. Patient comprehension is not perfect, especially during discussions with significant emotional content, such as in cancer prognosis or end-of-life discussions. Furthermore, even with perfect comprehension of information, there may be other factors that influence how patients arrive at decisions. Improving Comprehension When conveying risks and benefits, clinicians should start by assessing a patientâs comprehension. One major barrier to comprehension is innu- meracy. For instance, if 10 percent of 1,000 people have a particular disease, only one in three Americans can calculate how many people actually have that condition (100 of the 1,000 people). Therefore, one should not assume that patients understand the concept of percentages. In fact, it may be better to assume the opposite. When helping people make complicated decisions, information should be reinforced in multiple ways. Percentages should be reinforced with frequencies, and any numerical information should be reinforced visually. Pictographs improve comprehension better than other visual aids, such as bar graphs or pie charts, because pictographs help people quickly grasp the information relevant to their decision. For instance, the difference between outcomes of hormonal therapy with and without chemotherapy for breast cancer is more understandable when presented as a pictograph rather than a bar graph (Figure 8-3). In Figure 8-3 it is clear from the pictograph that adding chemotherapy translates into the survival of 2 more patients in every 100 even without the accompanying explanation. The pictograph presents the difference in risk, as opposed to a presenting the viewer with levels of
94 assessing and improving value in cancer care Bar Graph Hormonal Therapy 77 of 100 women are alive in 10 years. 23 out of 100 women die because of cancer. 7 out of 100 women die of other causes. Chemotherapy and Hormonal Therapy 2 out of 100 women are alive because of additional therapy. Pictograph Chemotherapy and Hormonal Therapy Hormonal Therapy 77 of 100 women 2 more women out are alive in 10 years. of 100 women are 23 out of 100 women alive because of additional therapy. die because of cancer. 7 out of 100 women die of other causes. FIGURE 8-3â Statistical representation of the added mortality benefit of chemotherapy for breast cancer in addition to hormone therapy using a bar graph versus a pictograph; statistical benefit shown asFigure 1-3 redrawn.eps 2 more women out of 100 alive because of additional chemotherapy. SOURCE: Ubel presentation, February 9, 2009. risk for each treatment which would then need to be compared. Perform- ing this comparison for the viewer eliminates a computational step and improves comprehension and understanding.
improving value in oncology practice 95 Comprehension Is Not Enough Comprehension is not always enough to ensure that the right decision is made. Even with perfect comprehension, a number of unconscious biases can lead people to decisions that do not follow from the information they have been provided. For instance, consider what happened when a group of women tried to decide whether to take tamoxifen for breast cancer preven- tion. Dr. Ubel and colleagues presented one group of such women with a pictograph showing the risk of developing cataracts as 11.3 out of 100. He presented a different group the same statistical information using different denominators (one showing the same cataract risk of 113 out of 1,000). These latter women were more worried about cataracts because they saw more boxes shaded in (113 compared to 11.3) even though the percentage had not changed. The pictograph with a denominator of 1,000 made the women âfeelâ as though the risk was greater, even though the risk was no different than the previous pictograph with a denominator of 100. So how can providers align what patients feel with the statistical information they must use to make decisions? Human beings evolved without the written word and without num- bers, explained Dr. Ubel. Instead, our ancestors told narrative stories to convey information, and this is how we are wired. Can stories or testimoni- als be used to improve comprehension of statistics? To illustrate the effect of stories on decisions involving statistical comprehension, Dr. Ubel described a study he had performed in which subjects were asked to imagine they had severe angina and were to decide between bypass surgery or angioplasty for treatment. Subjects were told that bypass would require open surgery, have a long recovery period, and a 75 percent chance of curing the angina, while angioplasty would require minimal surgery, a shorter recovery period, and a 50 percent chance of cure. In addition, subjects were provided two short testimonials for each treatment, one story of someone who was cured and one story of someone who was not for both bypass and angioplasty. Twenty percent of subjects chose bypass. Next, Dr. Ubel changed the experiment to reinforce the statistics with the testimonials. For bypass, subjects were pro- vided with three stories of cure and one story in which the angina persisted (3 out of 4 or 75 percent) which reinforced the 75 percent success rate sta- tistic subjects were given for bypass treatment. For angioplasty, two stories told of a cure and two did not, matching the 50 percent success rate given for the treatment. With the stories better aligned with the statistics, subjects chose bypass twice as often as before, or 41 percent of the time. When sub- jects were then given statistics and pictographs with or without testimonials to match the statistics, there was almost no difference in the number who
96 assessing and improving value in cancer care chose bypassâ38 percent and 44 percent chose bypass, respectively. Why was there no added benefit to adding testimonials once the pictographs were provided? Were both the testimonials and the pictographs effectively changing the way the statistical risks felt to subjects? Because the statistics were simple enough for high comprehension even without a pictograph and the choice of treatment was not associated with the level of numeracy of the subjects, Dr. Ubel concluded that, indeed, the pictographs helped subjects both comprehend and feel the difference in treatment success rates. In conclusion, Dr. Ubel turned his attention to the social context of cancer care and clinical discussions of cost. He asked, is there a harder con- text to discuss costs than when faced with a life-threatening illness such as cancer? Comparing risks of mortality and quality of life can be challenging enough. Adding cost considerations seems almost inappropriate because most people have no way to conceptualize trade-offs between survival and cost. Physicians may have good methods available for communicating medical risks and benefits to patients, but the next big challenge is learning how to help patients factor in financial costs as well. CONCRETE IDEAS FOR INCREASING VALUE IN ONCOLOGY CARE: A VIEW FROM THE TRENCHES Dr. Smith began by saying that his presentation would address a number of topics, including unrealistic demands for treatment benefit, low reimbursement for cognitive care, high reimbursement for chemotherapy or infusions, the high-income expectations of oncologists, variable quality, and stress and burnout. Oncology Practice Problems, Their Explanations, and How They Relate to Value The United States spends twice as much money as any other country for the same results in treating cancer (Meropol and Schulman, 2007). While the system rewards oncologists for giving more and more treatment, rather than taking the time to explain to patients when more treatment will not help, the expectations for success in oncology are impossible to meet. As a result, 16â20 percent of patients receive chemotherapy within 14 days of death, hospice stays are too short (one third are only 3 days or less), and we spend 25 percent of Medicare dollars in the last months of life (Harrington and Smith, 2008). Meanwhile, there are at least 100 new cancer drugs in
improving value in oncology practice 97 phase III clinical trials, all will be expensiveâ$5,000 to 7,000 per month (Hillner and Smith, 2009)âand many may be very useful. Allowing these same financial incentives to persist will not be sustainable, Dr. Smith said. Unrealistic Demand for Benefit Most people have unrealistic expectations of benefit from cancer care because they are told the goals of care but they are not told if they are going to die from the disease, how long they have to live, and the anticipated benefit from chemotherapy versus supportive care. The absence of decision aids for treatment of metastatic disease, or other tools to help clinicians weigh toxicities and cost when cure is not the goal, also enables unrealistic benefit expectations. As a result of this avoidant behavior, it is intellectually simpler and more financially rewarding to continue giving chemotherapy until the end is near obvious. By Dr. Smithâs calculations, oncologists make three to five times more money by giving chemotherapy than by talking with people. Low Reimbursement for Cognitive Care Oncologists are paid based on relative value units (RVUs). At $37 per work-related RVU, an oncologist would have to generate over 16,000 work-related RVUs yearly to earn the average U.S. oncologistâs salary. Most oncologists can do between 6,000 and 9,000 RVUs. Even if an oncologist could generate that many work-related RVUs, that would only cover his or her individual salary. How would the rest of the practice be supported? Other clinicians (nurses, social workers, counselors) and integral services (billing, electronic medical records, capital expenses, and so on) would not be supportable financially. Realistic workloads for long clinic days would lead to a decrease of roughly $200,000 per year in the average oncologistâs income if they were based solely on cognitive services, and there would be no way to pay oncologistsâ staff. High Reimbursement for Chemotherapy and Infusions The majority of community oncology practice reimbursement comes in the form of noncognitive care. In 2006, the median oncologistâs yearly salary was $358,000, the mean was $523,000, and the 90th percentile made over $1 million. Oncologists do not give chemotherapy just to make
98 assessing and improving value in cancer care money, Dr. Smith said. According to one published study, reimbursement does not influence the decision to give chemotherapy or not, but the more generous the reimbursement the more expensive the chemotherapy oncolo- gists use. For every dollar increase in reimbursement for the chemotherapy there is a $23 increase in overall chemotherapy costs (Jacobson et al., 2006). Throughout the last 10 years, oncology salaries have increased 86 percent while oncology visits have only increased 12 percent (Bodenheimer et al., 2007). High-Income Expectations, Stress, and Burnout in Oncology Oncologists have high-income expectations for doing their job. For many, it is a 24-hour-a-day, 7-day-a-week, 365-day-a-year responsibility to patient care. Furthermore, the looming 40 percent shortage of oncologists in this country will prevent any brisk reductions in salary in the near future. Within the current oncology payment structure, there is little reward for having a patient discussion on withholding chemotherapy because doing so is harder emotionally and will reduce reimbursement for the practice. Regarding our current incentives, Dr. Smith said: It is a lot easier and less angst producing to just give fourth-line chemotherapy than to sit with a 63-year-old guy with mesothelioma, as I did last Wednesday, and say âYour performance status is four,â which he doesnât understand, âbut you are in bed or the chair all the time. You are very short of breath. I know you desperately want to live longer, but thereâs no chemotherapy that has been shown in any randomized controlled trial .â.â. that is going to make you live longer or better. It would be different if you were much healthier.â He is in tears. His wife is in tears. Three sons in the background are saying âWell, thatâs not what we read about on the Internet.â Whereâs the reward for that? There isnât one except maybe doing a good job. It is hard to take care of sick people. No one likes to break bad news. Add to that breaking bad financial news, and it becomes even more dif- ficult (McFarlane et al., 2008). Dr. Smith said that he is very sympathetic to oncologists who would just give chemotherapy to please the family and patient, and give them some hope, rather than fight the uphill battle to get the patient into hospice at that point. There are no successful models for managing all of these expectations that oncologists have that do not require major shifts in lifestyle, incentives, and income.
improving value in oncology practice 99 Variable Quality of Care The quality of cancer care is extremely variable, with many practices giving treatments that are not supported by evidence, such as fourth-, fifth-, or sixth-line chemotherapy for non-small cell lung cancer (NSCLC) or those with severely impaired functional status. Furthermore, the relative merits of treating actively versus hospice care are unclear. For instance, a recent study of Medicare beneficiaries with pancreatic and lung cancers enrolled in hospice found that those who received chemotherapy closer to their enrollment in hospice died sooner, consistent with death from side effects (Connor et al., 2007). This argues for early hospice referral, but there is no reward for oncologists who refer to hospice sooner, just lost income and difficult conversations. Lastly, there is little interest in performing non- inferiority trials to show treatment equivalency, though it could save society considerable amounts of money. Potential Oncology Practice Solutions Dr. Smith addressed each problem raised in the paragraphs above with specific suggestions to improve value in cancer practice. Unrealistic or Uneducated Demands for Benefit One thing we can do to reduce unrealistic expectations for cancer treatment benefit is revise the National Cancer Institute (NCI; www.cancer. gov) and American Cancer Society (ACS; www.cancer.org) websites to give truthful prognosis and treatment information to patients and their provid- ers. Dr. Smith said that he and his colleagues were constructing an online resource for patients to show them data on prognosis with various treat- ment options in metastatic cancer. One example is shown in Figure 8-4. He explained that this information was not meant to dishearten patients, but to help them plan for all outcomesâcure or otherwise. Consent forms for chemotherapy should ensure that oncologists consider and document their patientâs Eastern Cooperative Oncology Group (ECOG) performance status (Oken et al., 1982), document his or her prognosis, specify the anticipated patient co-pay, specify the line of therapy planned, and specify the anticipated benefits and risks (Box 8-1). Insurance companies could then incentivize oncologists and patients who document these factors and then consider them in treatment decisions for
100 assessing and improving value in cancer care Breast cancer, fourth-line chemotherapy What is my chance of being alive at 1 year if I take chemotherapy or do best supportive care such as hospice? Without chemotherapy, about 5 of 100 women would be alive at 1 year. With chemotherapy, about 10 of 100 women would be alive at 1 year. The average woman with breast cancer treated with âfourth-line chemotherapy â lived about 5 months. Half will do better, half worse. Chemotherapy 10 90 Alive at 1 year Dead at 1 year Best supportive 5 95 care, or hospice 0% 20% 40% 60% 80% 100% Chance of being alive in 1 year What is my chance of my being cured by chemotherapy? FIGURE 8-4â Example of truthful patient information regarding metastatic cancer: fourth-line therapy for breast cancer. SOURCE: Smith presentation, February 10, 2009; from GO8 NLM009525 Smith T (PI), Truthful Information about Prognosis and Options for People with Advanced Cancer. third-, fourth-, or fifth-line treatments. However, criteria for decisions to limit treatment should be determined by society; doctors should not have to limit care in this way at the bedside without any external guidelines or support, Dr. Smith said. Why bother bringing up discussions of death? Holly Prigersonâs group prospectively followed several hundred patients who eventually died of their cancer (Wright et al., 2008). Only 37 percent of them ever had a discussion with their oncologists about the fact that they were dying from their cancer. She found that those 37 percent who were told that they were going to die from their cancer showed no difference in their mental health or worry, and were far less likely to want heroic measures, receive ventilation, or be admitted to the ICU. They were far more likely to admit being terminally ill, complete a do-not-resuscitate order, to use hospice, and to have oppor- tunities for life review. In a companion article, the authors calculated the
improving value in oncology practice 101 Box 8-1 Example Informational Consent Form for Treatment I, __________________ give __________________ permission to administer the following chemotherapy drugs to myself, _____ ___________________________________________________ I understand the risks of chemotherapy and have been provided the following information: ( ) drug-specific information ( ) disease-specific information. I ( ) do ( ) do not want specific information about the benefits of chemotherapy (see below). The goals of this therapy are (doctor should check the appropriate area) ( ) cure ( ) to help me live longer (my disease is not likely to be cured) ( ) to help me feel better Chemotherapy for my _______________ cancer as 1st-, 2nd-, 3rd-, 4th-, ____-line treatment is expected to, on average: â¢ Increase my chance of being alive at 1 year from ______% to ______% â¢ Increase my survival compared to no treatment from ________ months to _______months â¢ Prevent my cancer from getting worse (progressing) by ______months â¢ Shrink my cancer in half in ___ of 100 people like me â¢ Shrink my cancer completely in ___ of 100 people like me â¢ Keep my cancer stable for ___ months in ___ of 100 people like me â¢ Improve my cancer symptoms in ___ of 100 people like me Patient signature: _____________________________________ MD signature: ________________________________________ Physician documentation of patient ECOG Performance Status, 0 to 4 (circle one): 0 â no symptoms 1 â symptoms but normal activity 2 â symptoms but still functioning 3 â in bed or chair â¤ 50% of the day 4 â in bed or chair â¥ 50% of the day
102 assessing and improving value in cancer care costs of care for those who had the discussions were 36 percent lower than the cost for those who did not (Zhang et al., 2009). If one adds this to the 50 percent savings shown for switching from oncology services alone to oncology plus palliative care-based deaths in the hospital, the potential savings are considerable (Smith et al., 2003). Low Reimbursement for Cognitive Care Low reimbursement for cognitive care requires a significant revision of the current system toward rewarding good practice, not unsustainable entrepreneurship, said Dr. Smith. One solution could be paying all oncolo- gists $50 per work-related RVU for cognitive care. Peter Eisenberg has gone so far as suggesting that oncology should be a salaried profession to reduce the incentive to administer inappropriate anticancer therapy (Eisenberg and Mahar, 2009). Alternatively, if payors think that having a nurse, chaplain, or psychologist is important, they could pay for these clinical staff directly, rather than through the oncologistâs salary, Dr. Smith suggested. If these steps are taken beforehand, then profits from chemotherapy should be reduced, said Dr. Smith, one way or another. High Reimbursement for Chemotherapy or Infusions and High-Income Expectations We can reduce the incentives to overuse chemotherapy. Though it will be very complicated to develop capitated or illness-based episode payments, reimbursement could be reduced for chemotherapy to 6 percent over the average sales price. Furthermore, there should be a reward for oncolo- gists who follow National Comprehensive Cancer Network (NCCN) guidelines (second-line NSCLC and third-line breast cancer) to switch to Â nonchemotherapy-based care when their patientâs performance status reaches 3 or 4 (NCCN, 2009). Increasing government purchasing power to reduce the costs of drugs will be essential, and a U.S. version of the United Kingdomâs National Institute of Health and Clinical Excellence (NICE) may be needed to assess the cancer drugs arriving soon from the pipeline. Oncologists say they want to treat people, not disease, and these changes are all in line with that, Dr. Smith said. Oncologists in other coun- tries seem reasonably happy and productive, Dr. Smith noted, and their patientsâ survival is the same.
improving value in oncology practice 103 Stress and Burnout It is incredibly hard to take care of sick people. That is not going to change. In fact, Dr. Smith said, as people live longerâthank goodnessâ they have more complications. Dr. Smith emphasized that these difficulties should not be made worse by forcing oncologists to do the work of rationing that others (payors, government, society) avoid. Reasonable guidance and guidelines, clearly defined societal expectations, or the equivalent of a NICE are required to provide oncologists the decisions and support they need. Variable Quality of Care The American Society of Clinical Oncologyâs Quality Oncology Prac- tice Initiative has successfully assessed oncology practices for adherence to published guidelines. Could this model be applied to assess the number of practices with appropriate counseling and social work staff? Could they be used to assess what percentage of practices delivers third-, fourth-, or higher-line chemotherapy, for which cancer types, and for patients at what performance status? Could they be used to track the percentage of practices that give chemotherapy or refer to hospice within the last 2 weeks of life? Dr. Smith said that all of these were achievable, but there was no reward for doing themâyet. ADVOCATING FOR VALUE IN CANCER CARE Ms. Ellen Stovall of the National Coalition for Cancer Survivorship said, with her 16 years of cancer-related advocacy in Washington, DC, she still believed that the job of citizen activists and advocates is to press for reforms and hold accountable the systems that organize health care, finance health care, finance cancer research, and regulate drug products in order to truly bring patient-centered care to people with cancer and their families. Cancer patients, for the most part, are treated outside any system of quality assurance or required accountability for cancer care. Many people diagnosed with cancer do not know that, and they may not want to know. They want to believe the care they receive is of value. Everyone hopes for a cure for cancer. Every family wants this. Our society wants this. Our lawmakers want this. But we live in a society that is uncomfortable with something that we are all going to experienceâour own death. The combination of our hope for a cure and our discomfort with mortality fails the patient and all those involved with their care. Research
104 assessing and improving value in cancer care on thousands of cancer survivors over three decades shows that survivors fear pain and suffering, and they fear abandonment by their physicians and others. There is now a tremendous opportunity to ramp up discussions of retooling our cancer care system so that it is truly a patient-centered, health care delivery system that values and rewards participation in clinical research or registries, and so that it is a system that values the time health care profes- sionals spend with patients by realigning reimbursement. Clinicians also have the opportunity to promote the implementation of cancer treatment and survivorship plans that clearly spell out goals of treat- ment for each patient. This planning process would focus on treating the whole patient, not only the disease. Creating a cancer care plan after review- ing all the evidence-based choices available for treating each patient provides the foundation or framework for difficult, shared decision making. It also fosters a clinical relationship that is founded in openness and trust, telling the truth, and that values shared decisions, provides support and continuity of care, diminishes fear of abandonment, and accepts all visions of hope that patients and their family members bring to the treatment process. Ms. Stovall concluded by saying that to achieve this vision of cancer care, FDA reform is needed to not only create novel ways of looking at clinical trial design but also to expand access to novel therapies. This should enable treatment knowledge to advance with every patient and provide equity and access to every patient. In addition, payment and financing reform are needed. This can be accomplished by working with private payors and Congress to realize Medicare reimbursement that values cancer care and survivorship care planning as fundamental elements of quality care. Finally, patient and physician education are needed, as well as close collaboration with those in the caregiving community to improve value in cancer care. DISCUSSION Dr. Newcomer asked Dr. Smith to clarify what he meant when he said in his presentation that physicians should not be making decisions about rationing at the bedside on an individual case-by-case basis. Is it not essential that physicians be in the discussion about where therapy should be stopped as a policy and at the bedside? Absolutely, said Dr. Smith, and he explained that oncologists and other physicians have a critical role to play in determining what works. Physicians are perhaps in the best posi- tion to make judgments about when to stop third- or fourth-line therapy.
improving value in oncology practice 105 What should not happen, however, is a situation in which physicians are forced to make decisions for individual patients at the bedside without the support of clinical guidelines or other decision makers who will back them up. Dr. Newcomer asked, who should be writing those rules? Dr. Smith said he believed it should be those involved in using the available health care funds. In the case of UnitedHealthcare, it is all of the patients cov- ered. Dr. Tunis added that Dr. Newcomerâs question at its core was asking, âWho should be responsible for rationing care?â It would be great if a U.S. NICE could remove the discomfort and burden of these decisions from all of us, but Dr. Tunis thought a U.S. NICE was unlikely in the near future. Dr. Chernew added that the fact that those paying into Medicare now are not the current beneficiaries introduces another difficulty in trying to decide who should be involved in rationing decisions. The current Medicare beneficiaries will tend to want Medicare to be much more aggressive than the younger people who are financing Medicare would want it to be. The same might be true of the majority of those paying premiums into private plans now who are not the minority spending the most on care through those plans. Dr. Garrison asked whether NCCN and other guideline developers had a role to play. There was agreement among the panelists that American Society of Clinical Oncologists, NCCN, ACS, and NCI would all need to be involved in delivering truthful information to patients and clinicians about what can be done medically for their illnesses. Mr. Merrill Goozner of the Center for Science in the Public Interest asked whether oncologists are being reimbursed for care that is unproven. Dr. Smith clarified and emphasized that oncologists are not giving unproven therapies, but they are administering therapy that is probably easier to give than having the very difficult discussions about prognosis and chance of cure. The difficulty also appears when helping people make decisions when their values are not shared by the clinicians guiding them. Clinicians provid- ing guidance should be nondirective and help patients make tough decisions while holding an even keel. Ms. Jennifer Hinkelfen of NCCN asked what role clinical practice guidelines have in helping to define what value is. Should cost consider- ations be implicit or explicit in the guidelines? Dr. Tunis responded by saying that it would be a very good idea to explicitly provide economic information on comparative costs of alternative treatments in the guidelines but perhaps not to integrate this fully into the clinical recommendations. In this format the information would be readily available to introduce it
106 assessing and improving value in cancer care into the conversation, but not such that it would directly dictate clinical decisions. Dr. Back suggested that a professional expectation be developed among oncologists that they engage and pursue these end-of-life conversations, guide trainees to do the same, and support one another to create a culture that makes this the norm. Dr. Smith agreed with the suggestion. REFERENCES Bodenheimer, T., R. A. Berenson, and P. Rudolf. 2007. The primary careâspecialty income gap: Why it matters. Annals of Internal Medicine 146(4):301â306. Chandra, A., J. Gruber, and R. McKnight. 2007. Patient Cost-Sharing, Hospitalization Offsets, and the Design of Optimal Health Insurance for the Elderly. Cambridge, MA: National Bureau of Economic Research. Chernew, M. E., A. B. Rosen, and A. M. Fendrick. 2007. Value-based insurance design. Health Affairs 26(2):w195âw203. Chernew, M. E., M. R. Shah, A. Wegh, S. N. Rosenberg, I. A. Juster, A. B. Rosen, M. C. Sokol, K. Yu-Isenberg, and A. M. Fendrick. 2008. Impact of decreasing copayments on medication adherence within a disease management environment: Value-based cost sharing can increase patientsâ adherence to important medications. Health Affairs 27(1):103â112. Connor, S. R., B. Pyenson, K. Fitch, C. Spence, and K. Iwasaki. 2007. Comparing hospice and nonhospice patient survival among patients who die within a three-year window. Journal of Pain and Symptom Managment 33(3):238â246. Cranor, C. W., B. A. Bunting, and D. B. Christensen. 2003. The Asheville Project: Long- term clinical and economic outcomes of a community pharmacy diabetes care program Journal of the American Pharmaceutical Association 43(5 Suppl):S36âS37. Eisenberg, P., and M. Mahar. 2009. A very open letter from an oncologist. http://www. healthbeatblog.org/2009/01/a-very-open-letter-from-an-oncologist.html (accessed April 11, 2009). Fendrick, A. M., D. G. Smith, M. E. Chernew, and S. N. Shah. 2001. A benefit-based copay for prescription drugs: Patient contribution based on total benefits, not drug acquisition cost. American Journal of Managed Care 7(9):861â867. Freudenheim, M. 2007. To save later, some employers are offering free drugs now. The New York Times, February 21, 2007. Fuhrmans, V. 2004. A radical prescription: While most companies look to slash health costs by shifting more expenses to employees, Pitney Bowes took a different tack. The Wall Street Journal, May 10, 2004. Harrington, S. E., and T. J. Smith. 2008 The role of chemotherapy at the end of life: âWhen is enough, enough?â Journal of the American Medical Association 299(22):2667â2678. Hillner, B., and T. J. Smith. 2009. Efficacy does not necessarily translate to cost effectiveness: A case study in the challenges associated with 21st-century cancer drug pricing. Journal of Clinical Oncology 27(13):2111â2113.
improving value in oncology practice 107 Jacobson, M., J. A. OâMalley, C. C. Earle, J. Pakes, P. Gaccione, and J. P. Newhouse. 2006. Does reimbursement influence chemotherapy treatment for cancer patients? Health Affairs 25(2):437â443. Mahoney, J. J. 2005. Reducing patient drug acquisition costs can lower diabetes health claims. American Journal of Managed Care 11(5 Suppl):S170âS176. McFarlane, J., J. Riggins, and T. J. Smith. 2008. SPIKE$: A six-step protocol for delivering bad news about the cost of medical care. Journal of Clinical Oncology 26(25):4200â 4204. Meropol, N. J., and K. A. Schulman. 2007. Cost of cancer: Issues and implications. Journal of Clinical Oncology 25(2):180â186. NCCN (National Comprehensive Cancer Network). 2009. National Comprehensive Cancer Network guidelines: Non-small-cell lung cancer and breast cancer. http://www.nccn.org/ professionals/physician_gls/f_guidelines.asp (accessed March 28, 2009). Oken, M. M., R. H. Creech, D. C. Tormey, J. Horton, T. E. Davis, E. T. McFadden, and P. P. Carbone. 1982. Toxicity and response criteria of the Eastern Cooperative Group. American Journal of Clinical Oncology 5(6):649â655. Siu, A. L., F. A. Sonnenberg, W. G. Manning, G. A. Goldberg, E. S. Bloomfield, J. P. Newhouse, and R. H. Brook. 1986. Inappropriate use of hospitals in a randomized trial of health insurance plans. New England Journal of Medicine 315(20):1259â1266. Smith, T. J., P. Coyne, B. Cassel, L. Penberthy, A. Hopson, and M. A. Hager. 2003. A high- volume specialist palliative care unit and team may reduce in-hospital end-of-life care costs. Journal of Palliative Medicine 6(5):699â705. Wilt, T. J., T. Shamliyan, B. Taylor, R. MacDonald, T. J., I. Rutks, K. Koeneman, C.-S. Cho, and R. L. Kane. 2008. Comparative effectiveness of therapies for clinically localized prostate cancer: Comparative effectiveness review no. 13. Rockville, MD: Agency for Healthcare Research and Quality. Wright, A. A., B. Zhang, A. Ray, J. W. Mack, E. Trice, T. Balboni, S. L. Mitchell, V. A. Jackson, S. D. Block, P. K. Maciejewski, and H. G. Prigerson. 2008. Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. Journal of the American Medical Association 300(14):1665â1673. Zhang, B., A. A. Wright, H. A. Huskamp, M. E. Nilsson, M. L. Maciejewski, C. C. Earle, S. D. Block, P. K. Maciejewski, and H. G. Prigerson. 2009. Health care costs in the last week of life: Associations with end of life conversations. Archives of Internal Medicine 169(5):480â488.