The Federal Challenge of Responding to the Need for a RLHS
Healthcare advances are often fostered by federal policy and led by federal agencies. As Dr. Wallace noted, policies that limit smoking in public places have substantially impacted tobacco use rates, and “if you want to make major changes in population health, it is not just the healthcare system [you need to focus on], but public policy.” Etheredge called on the HHS to lead the effort to create a RLHS for cancer. HHS is the umbrella organization for a number of federal agencies that determine healthcare policy for cancer, including FDA, NIH, NCI, the Cancer Clinical Trials Cooperative Groups, NCI-designated National Cancer Centers, AHRQ, CDC, CMS, and ONC. “If HHS does not lead, the chaotic state of adult cancer care will likely continue, or get worse,” said Etheredge.
He expanded that statement by noting that one HHS agency in specific, CMS, which runs both Medicare and Medicaid, is the largest payer for cancer care in the country. Between 5 and 10 percent of Medicare’s budget is spent on cancer care, “so it will be good to get those incentives aligned with the kind of quality we want and the kind of learning system we want,” Etheredge said. “In a sense we are running a large national experiment through Medicare by paying for tens of billions of dollars of cancer care, but we never collect any data, and we know it is of very uncertain quality,” he added. He called for using the purchasing power of Medicare to support the healthcare learning agenda and better-quality cancer care.
In his discussion of how Medicare could further the development of
a cancer RLHS, Etheredge outlined a series of steps CMS could take to influence and support a learning agenda for cancer care, including the following:
Cover new cancer therapies subject to evidence development reporting requirements to learn as rapidly as possible about their best use for personalized care.
Require expanded reporting of cancer clinical data and quality measures to national cancer registries, and withhold payment to providers until after such information is submitted and passes review.
Fund and set standards for EHRs, with Medicare cancer care modules for Medicare patients.
Shift Medicare cancer payments from paying for anything and everything that is done to “pay for performance” in “preferred provider” contracts, thereby paying more for effective, high-quality care, as measured by guidelines. This would incentivize participation in learning healthcare networks.
Inform Medicare patients and physicians about cancer CER, best practices, and quality performance. By posting information on the Web, patients can choose the best cancer providers and treatments.
Use innovation funds to support new models and incentives for patient-centered, high-quality care.
As Etheredge noted, much of the national funding for EHR subsidies will be from Medicare so it makes sense that this agency should set the standards for those systems. “It would be a huge travesty to have Medicare paying for EHRs and not be able to have the meaningful use requirements include reporting the data that we need to advance the quality and the information base of Medicare cancer care,” Etheredge said. He emphasized the need for an organized, coherent national cancer strategy in which “HHS is going to put all of the pieces together rapidly” to advance to a new era of evidence-based health care. Yet it can be challenging to carry out some of these Medicare objectives, noted Dr. Bach, who detailed the difficulties Medicare had with its coverage with evidence development (CED) program in which additional data are gathered in the course of care. This program may withhold payment for a specific treatment or diagnostic unless the intervention is accompanied with data physicians submit for their patients that receive the intervention. Within this program, Medicare may also agree to pay for an intervention, but it will pay providers more if
they submit more patient data. There are two types of CED that have been implemented. One type is dependent on treated patients being enrolled in clinical trials of the intervention being given. Another type of CED is registry based, in which coverage is given only if physicians provide additional information about the patients given the diagnostic or treatment. For example, Medicare’s coverage of PET scans for many oncology indications initially was not guaranteed unless ordering physicians provided information to the National Oncologic PET Registry on patients’ clinical status and expectations and potential responses in light of PET scan results. The analysis of registry data on 23,000 scans revealed that about one-third of clinical decisions are altered by PET results, and nearly three-quarters of biopsies are avoided by using PET (Hillner et al., 2009). The findings led Medicare to lift the CED requirement and cover PET scanning for all of the indications that were examined.
Dr. Bach noted a number of lessons that were learned and challenges that were insufficiently met with Medicare’s first CED experiences. The registries used for CED are not housed at Medicare, but rather with a third party, for a number of reasons, and this has been a major impediment. “There is no such thing as a disinterested third party,” Dr. Bach said, and it was also challenging to update funding for such registries. The ethics of the human testing that occurs in registry-based CED also is not straight-forward, he added. Another limitation of this approach to foster learning is that there is no ability for Medicare to fund analyses or efforts to acquire follow-up data, Dr. Bach noted. He also pointed out that CMS was rightfully criticized for oversampling in its observational clinical studies, noting that the large sample sizes available in certain registries and other databases strictly speaking are not needed to study certain simple issues, especially if there is high prevalence of an event being assessed. “Why do you need 23,000 patients to figure out if a PET scan changed therapy?” he asked, pointing out that inefficiencies and excessive costs occur when overly large sample sizes are used in clinical research. The agency was also accused of not asking the right questions (e.g., it should have asked whether outcomes rather than clinical decisions were affected by PET scans).
Medicare’s CED that uses clinical trials also has had some challenges. “We got tremendous pushback from manufacturers because, in many cases, the sponsors were about to launch the trials anyway and they did not like the inefficiencies,” Dr. Bach said. In addition, there were problems with adequately blinding participating patients because the copayments were so
much higher for the tested treatment so patients were aware of which treatment they were receiving.
Medicare also has had problems with its 2006 Oncology Demonstration Project, in which physicians were paid $23 to submit additional information, such as cancer stage, treatment adherence to guidelines, and the focus of the visit. As described previously, because of basic misunderstandings or lack of awareness of what was being asked of them, physicians often did not enter the required information properly, which limited the usefulness of the data, Dr. Bach noted.
The FDA is another key federal agency with a major role to play in fostering a nationwide RLHS because its approval of drugs and devices is needed for these products to enter the market. “We set the standards required for medical products to get on the market. This is a real hard incentive,” said Dr. Woodcock, who is the director of FDA’s Center for Drug Evaluation and Research. She added that the evidence of safety and effectiveness FDA requires for medical interventions is not the only evidence needed to show that these products are truly safe and effective in clinical real-world settings over the long term. Dr. Woodcock also pointed out medical products and uses for which there are no FDA-generated evidence requirements. These products and uses include procedures, certain laboratory-developed tests, and off-label uses of interventions, which are common in oncology and often lack a formal evidence base. In addition, because cancer is a life-threatening disease, the FDA has granted many “accelerated” approvals for cancer treatments that enable them to enter the market without meeting FDA’s more rigorous requirements for safety and effectiveness but require manufacturers to verify that these standards are met in the postmarket arena by conducting ongoing clinical trials. Until that verification is achieved, FDA can restrict the distribution of medical products. FDA also plays a role in evidence dissemination by regulating what sponsors can say about their product on the label.
The FDA Amendments Act, which was passed in 2007, enables the FDA to require postmarket studies and surveillance activities from sponsors. It may require all patients or only a subset of patients receiving a medical treatment to be enrolled in a registry to assess safety issues of the treatment. In addition, FDA conducts its own postmarket surveillance of product safety, an activity that has recently been enhanced with its newly launched Sentinel initiative. This initiative is intended to help FDA more rapidly learn about safety outcomes from distributed healthcare data of
up to 100 million patient records collected from a network of federal and private data sources.
Dr. Woodcock noted that these efforts to collect more postmarket safety and effectiveness data are not sufficient, and there needs to be more of a bridge between the development realm of clinical trials and the healthcare realm. As she pointed out, “One day drugs are unapproved and investigational, and CMS does not pay for them, and then the next day they are approved and you can use them for anything. If you designed a rational system that would not be it, because there is a continuum of learning and we really do not have the opportunity to match use according to knowledge—the level of evidence that there is about a drug.”
There also needs to be more effective translation of new science to the clinic, Dr. Woodcock stressed, and given the unacceptable time and effort to do clinical trials, she added, “The current paradigm of how we do trials and execute them and get results limits our capacity in this country.” She called for more widespread community participation, as discussed earlier, to narrow the gap between research and practice and to accelerate learning.
The FDA and Medicare wield tremendous regulatory power in health care, but they are just two of several federal agencies that can spur the development of a cancer RLHS. If federal agencies cooperated, even more could be accomplished, Etheredge noted. He envisions such cooperation, coupled with collaborations with the private sector, leading to the establishment of a three-year national research plan for “learning as much as possible as soon as possible” about the best use of new technologies for personalized cancer care. This could be achieved if such a program was hinged to FDA approvals or Medicare payments for new drugs or devices. In addition, Medicare’s coverage with evidence development, along with similar incentives from private payers, could generate the data for the much-needed CER studies. A national data registry system or clearinghouse for CER data is also needed, along with effective means to disseminate the CER results to patients and physicians, Etheredge said. With the collaboration of all HHS agencies, “we have the ability to move forward with a system that could capture a lot of learning very quickly, and help inform physicians and patients,” he added.