Cancer community weighs in on comparative effectiveness research
Recognizing that more than half of healthcare treatments today are delivered without clear evidence of effectiveness, the US Congress last year appropriated $1.1 billion toward accelerating comparative effectiveness research (CER) as part of the American Recovery and Reinvestment Act (ARRA) of 2009.
More recently, healthcare reform legislation passed by Congress created provisions for developing a CER infrastructure, along with a mechanism for funding and coordinating such research. Each of the bills creates a trust fund to support this research through a combination of appropriations and a per-capita fee on Medicare and private insurers. A major difference between the Senate and House bills is their governance structure; those differences had not been resolved at press time. A report issued by the Institute of Medicine defines CER as, “the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat, and monitor a clinical condition or improve the delivery of care.”1 The main goal of CER, the report concludes, is to help patients, clinicians, insurers, and policy makers make better informed healthcare decisions.
“CancerScope” asked several members of the oncology community to comment on how new government emphasis on CER will affect cancer research and treatment.
Collaboration and Communication Key
A Washington, DC-based advocacy organization, Friends of Cancer Research, hosted a forum on Capitol Hill in May 2009, in conjunction with its publication of a white paper, Improving Medical Decisions Through Comparative Effectiveness Research: Cancer as a Case Study.2
“We tried to step back and look at how this would fit into the overall landscape of cancer research,” says Jeff Allen, PhD, executive director of Friends of Cancer Research. “The oncology community can help lead the way because we routinely collect biospecimens with diagnoses—which can help generate additional information. We also have a unique national infrastructure to conduct these broad, multi-institutional studies and look at electronic data sources “Because clinical trials must be conducted in small populations, it's important to understand the underlying biology of why certain treatments work in order to pose additional questions for future studies that apply in “real world” situations, he adds.
Not only is it important to build a network of electronic data sources to help generate hypotheses, but it is also essential to ask the right questions. In addition, the scientific community must determine how these study results can be used and communicated in the most effective way. “That's difficult to codify into law, and it's likely to vary,” Dr. Allen says. “Patients and physicians will have to use results as an added tool to their decision making—and each individual's situation will have to be evaluated on its own merits.”
Dr. Allen stresses that learning more about therapies and products after they are already out on the market will present challenges because most pharmaceutical companies and medical device manufacturers do not have an incentive to compare their product to that of a competitor.
Ensuring that these studies get done will require some type of public-private collaboration, he says. “Asking the questions and identifying the gaps will be easy—designing and conducting the studies will be challenging.”The advocacy community must stay involved in the discussions and ensure that CER is sufficiently funded to complete some of these important studies, Dr. Allen concludes.
New Ways of Looking at Data
Because phase 3 clinical trials are so costly ($100 million or more each) and take so long to complete, CER requires additional methods of research, such as looking at databases, according to Donald Berry, PhD, head of the quantitative sciences division at the University of Texas M. D. Anderson Cancer Center in Houston. Dr. Berry designs clinical trials and works on population-based studies.
“We need to develop databases and ways of sharing information across groups, including pharmaceutical companies,” he says. “We need to milk the information that's out there. The problem is, we don't always know if we can milk it, or if the milk has poison in it.”
Because CER seeks to learn how various therapies will work in the real world rather than just in the limited population of a clinical trial, the biggest question is where researchers will find databases that represent the world, Dr. Berry notes. “Sometimes the data is not there or the biases are so profound that you just can't use it,” he says. “That's part of CER—asking what the data is and whether we can exploit it to address these important questions.” He adds that, although scientists may not be able to obtain those answers in 2 years, they may be able to develop an efficient way to pose the questions in that period. Much of the data will have to come from modeling by using currently available information.
For example, Dr. Berry and colleagues used modeling to compare the benefits of biennial versus annual mammography screening in a paper published in the Annals of Internal Medicine.3 Developing models based on national data on age-specific incidence, competing mortality, mammography characteristics, and treatment effects, the authors concluded that biennial screening achieves most of the benefit of annual screening with less harm. No one had ever performed a study on this comparison, he adds.
Although models are only as good as their assumptions, they can generate answers much faster than the 10 or more years it takes to complete a phase 3 clinical trial, Dr. Berry notes. And, when performed appropriately, modeling can tell scientists what they do and do not know about a particular therapy or screening. However, if performed incorrectly, it can result in misleading conclusions, he says.
Still, Dr. Berry is optimistic that in 10 years, scientists will identify key questions and have a big impact. However, he laments the lack of a centralized national database for vital information on all cancer patients and their treatments. With only 3% to 5% of adult patients with cancer on clinical trials, some 95% of patients have enormous amounts of information that researchers cannot obtain. “Almost 100% of cancer patients have information somewhere on their disease, treatment, and outcome, but even in the same institution, we don't have all that information put together,” Dr. Berry says. “I think CER will lead to an awareness of the need to do that.”
He points to another problem with the current system: if a pharmaceutical company runs a drug trial and it fails, the US Food and Drug Administration is not allowed to release that information. As a result, another company may run a trial on the same drug not knowing it has already failed. “Wouldn't it be wonderful if that data had been in a bigger database, so the second company could have known and not wasted patient time and resources? Of course that would take a congressional act,” he says. He hopes CER funding will help create the infrastructure and ability to develop and access such databases.
Congressional healthcare reform legislation created provisions for developing a Comparative Effectiveness Research (CER) infrastructure as well as funding and coordination mechanisms.
The oncology community can help lead the way in CER because of its experience with multi-institutional clinical studies and biospecimen collection.
CER will require database networks to provide additional information that is collected by clinical trials and will incorporate modeling to generate conclusions from available information.
A continuing challenge will be communicating study results and getting the medical community and patients to abide by key recommendations.
Culture Change for Medical Consumption Needed
Deriving answers to some of the most pressing questions in oncology is not going to be easy, notes Otis Brawley, MD, chief medical officer of the American Cancer Society, pointing out that studies comparing radical prostatectomy to external-beam radiation therapy have been attempted and failed 3 times over the past 30 years.
“Men weren't willing to be randomized by a computer to decide whether or not they got surgery versus radiation,” he says, noting that women, on the other hand, were willing to be randomly assigned to mastectomy versus radiation and lumpectomy to determine the most effective treatment.
Not only must these important scientific questions be asked, but the medical community and patients must begin to abide by the answers once they're determined, he adds. “We tend not to be scientific and rational in the way we consume medicine in the United States,” he says. “We already have studies that show we do way too many x-rays and MRIs, but we don't listen to them. People still get screened for lung cancer with chest x-rays, even though studies done in the 1970s showed they don't save lives.”
Although some critics of healthcare reform worry that reform will lead to rationing of medicine, Dr. Brawley notes the real issue is the need for “rational consumption” of medicine. “There's a mistaken belief in this country that consuming more is better when it can actually be harmful,” he says. “We need to change the culture of how we consume medicine.”
He points to recent guidelines issued by the American College of Obstetrics and Gynecology recommending that women be screened for cervical cancer with pap smears every 3 years. Critics are concerned that's not frequent enough, “even though there is data that almost all women who were diagnosed with cervical cancer had never had a pap smear,” he says. “Researchers had found that some women were getting pap smears far more frequently than they needed, and those who needed them weren't getting them at all. We're trying to bring some rationality into this.”
In another example of recent CER, the European and American studies of prostate cancer screening showed different results. Although the European study found that screening reduces the lifetime mortality risk by 20%, it had a 0.98 confidence level. Meanwhile, the American study found that although screenings find many incidences of prostate cancer, the screenings do not actually save lives.
“I think men do have more information now than they did a year ago in helping them decide whether or not to get screened,” Dr. Brawley says. “The studies clearly show that screening diagnoses a lot of men who don't need to be diagnosed and causes treatment for a lot of men who don't need to be treated. It's a subjective decision based on objective evidence. If I'm afraid of prostate cancer or if my father died of it, then maybe I should get screened. If not, maybe I shouldn't.” Dr. Brawley hopes to see collaboration among the National Cancer Institute, the American Cancer Society, and other large nonprofits that fund cancer research to define and publish white papers on some of the key questions that need to be addressed by CER.
Hard Choices Ahead
In 2008, Leonard Zwelling, MD, MBA, professor of medicine and pharmacology and Chancellor's Health Fellow in Comparative Effectiveness at M. D. Anderson Cancer Center, got a firsthand view of the healthcare reform process when he became a Robert Wood Johnson Foundation Health Policy Fellow serving in the office of the ranking member of the US Senate Health, Education, Labor, and Pensions Committee, Sen. Michael Enzi (R-WY).
“We learned that CER is not just a clinical trial—it's a continuation of the results you acquire in a clinical trial because a clinical trial usually comes to an unambiguous answer,” he says. “In CER, the degree of certainty is less—it could be a meta-analysis or prospective data in a nonspecified population.”
Two key questions in CER that Dr. Zwelling says he does not know the answer to are how good the data need to be to change physician behavior and who will decide whether the research is good enough. A case-in-point is the recent US Preventive Services Task Force controversial recommendation to not conduct routine mammography screening on women in their 40s. “It's a question of how much money we're willing to spend to save 1 life,” he says. “In England, they're willing to have that debate all the time. But we won't even have the debate here. We think everyone should get what they want. The problem is, who will pay for it?” Although CER can hint at how much better or worse a medical decision may be, it will not make that decision for us, Dr. Zwelling notes.
In addition, paying for clinical research is a challenge. When Dr. Zwelling served as associate vice president for research administration at M. D. Anderson in the mid 1990s, about 25% of trials were paid for by pharmaceutical companies. Now that figure is closer to 60%. “If they're paying for it, they will structure it to their advantage—the whole tenor of research has changed,” he says.
At the same time, some people may not want to know whether one particular therapy is better than another. Patients want to know which therapy is the best treatment for their prostate cancer, for example. However, urologists and radiation therapists may not want to know, if there is a chance that the results will affect their livelihood, he says. How the results of CER are communicated to the public is as critical as the research itself, Dr. Zwelling adds.
“In the end, it's the doctor and a patient in a room, and the doctor wants to give that patient the best information. That's hard to do when the research is based on data that is not as tightly controlled as a clinical trial,” he says. “It will force us into decisions we've been unwilling to discuss thus far.”