A Clear Path to Reimbursement: Comparative Effectiveness Research

While the recent influx of funding for comparative effectiveness research (CER) from the American Recovery and Reinvestment Act (ARRA) has brought heightened attention to the field in the United States, such well-designed studies have been effectively implemented in Europe, Australia and Asia for the purposes of evaluating molecular imaging technologies and techniques. These efforts have resulted in government agencies recognizing the clinical value and cost effectiveness of these methods through reimbursement, leading the United Kingdom, France, Germany and Australia to create CE government-run entities as a by-product of national healthcare reform and a desire for cost containment.

What makes CER effective?

The American College of Physicians defines CER as “the evaluation of the relative (clinical) effectiveness, safety and cost of two or more medical services, drugs, devices, therapies or procedures used to treat the same condition.” However, properly designing and conducting prospective trials with these goals can be challenging and expensive.

As a result, researchers often rely on retrospective data, which requires a complex, branching study design. For instance, if a meta-analysis examines PET, it should combine data from the literature using a treatment algorithm tree that includes the incidence of disease, sensitivity and specificity of the PET/CT studies compared with conventional imaging—creating a branching structure where the number of correct diagnoses can be compared with the number of mistaken diagnoses. Also, the researchers should define and assign weights to the various steps along the way.

While SNM President Michael Graham, MD, PhD, acknowledges that this conventional analysis methodology can get “fairly messy,” the process allows physicians to evaluate both the costs and clinical effectiveness of the strategies. This method has proved particularly effective for the staging of lung cancer, and PET ultimately proved its cost effectiveness compared to conventional imaging.

One such example is an Italian study, conducted by Manseuto et al, which sought to evaluate the economic impact of the introduction of PET in the clinical management of patients with well known or suspected lung cancer through a cost-effectiveness analysis of different diagnostic strategies, such as CT alone versus PET, versus PET for all.

The Italian researchers found that the introduction of PET in the clinical management of all 75 patients with known or suspected lung cancer previously evaluated with CT is cost effective, and allows patients to gain 2.64 life years at an annual cost of about €415. The authors of the 2007 study, published in the Quarterly Journal of Nuclear Medicine and Molecular Imaging, were able to compare each branch. They decided PET as a result of inconclusive CT was not the best choice despite being more cost effective than PET for all, because it showed lower clinical effectiveness for life expectancy (1.58 vs. 2.11 life years).

In Europe, agencies “usually draw the line at approximately €33,500 per year of life saved to deem a procedure cost effective, and this study shows the annual cost of PET at a little less than €200 per year of life saved in the clinical management of lung cancer patients,” says Graham, who also is director of nuclear medicine at the University of Iowa Carver College of Medicine in Iowa City, Iowa.

Despite the high and often unchangeable mortality rate in lung cancer, CER has the opportunity, by evaluating molecular imaging techniques used in determining the extent of metastatic disease, to reduce cost of overall treatment. Another European study is seminal in the field of cost-effectiveness and was led by van Tinteren et al of the Netherlands, and published in the Lancet in 2002. The team found that one in five patients with non-small cell lung cancer could avoid unnecessary chest surgery if given a PET scan to determine the extent of metastatic disease. The cost analysis, which showed that although the cost of using PET had more expensive up-front costs, the molecular imaging test realized savings by reducing unnecessary surgery. The authors concluded that PET reduces futile thoracotomies by 50 percent without any net additional cost.

Based on these findings, a separate group of Dutch researchers conducted the 2004 PLUS study, published in the European Journal of Nuclear Medicine and Molecular Imaging with 188 patients, in whom the cost price of PET varied between €736 and €1,588, depending on the hospital setting and procurement of 18-FDG commercially or from on-site production. The average overall hospitalization costs per patient in the conventional work-up group were €9,573 and in the PET group, €8,284. The authors noted that the “major cost driver was the number of hospital days related to recovery from surgery.”

Due to these cost-effectiveness analyses, the findings were influential in PET gaining reimbursements for the diagnostic work-up of lung cancer patients in various European countries, according to Wolfram H. Knapp, MD, president of the European Association of Nuclear Medicine (EANM) and director of the nuclear medicine clinic at Hannover Medical School in Hannover, Germany.

In Europe, reimbursement is decided on the national level, so each country varies drastically on the fees paid. For instance, the socialized systems of Italy, Spain, France and Switzerland cover all PET indications, according to Knapp, but in Germany, PET is nationally reimbursed only for pulmonary tumors. “In Germany, only the private insurers cover PET for more indications,” he says.

“Across Europe, we have a spectrum of coverage [for molecular imaging technologies],” Knapp says. “Countries with socialized medicine, such as Italy or France, will look to policies in countries, like the United States, to determine their reimbursement determinations.” However, he adds that some well-accepted procedures, such as SPECT for oncologic purposes, have received widespread reimbursement throughout Europe.

Another CER study, conducted in Australia and published in 2005 in the European Journal of Nuclear Medicine and Molecular Imaging, helped to unequivocally prove PET’s value—and justify its costs on that continent—in the staging of lung cancer. Yap et al found that routine FDG-PET scanning with selective mediastinoscopy will save AU$2,128 per patient, and potentially reduce inappropriate surgery. They wrote that these “cost savings remain robust over a wide range of disease prevalence and FDG-PET costs.”

However, CER hasn’t always produced such positive results for molecular medicine, as the use of FDG-PET in some cancers has proven ineffective . “In certain slow-growing cancers, we are unable to visualize the disease,” Graham explains. As a result, the U.S. Medicare system doesn’t reimburse for the use of FDG-PET in these cancers, such as prostate cancer.

Also, questions remain about the efficacy of using CER to assess radiopharmaceuticals, particularly whether the data “can be robust enough and gathered quickly enough” for a government agency to decide on reimbursements, according to Barry A. Siegel, MD, co-chair of the NOPR working group, chair of the American College of Radiology Imaging Network (ACRIN) PET Imaging Core Laboratory and chief of nuclear medicine at the Mallinckrodt Institute of Radiology at Washington University in St. Louis.

While a great deal of CER is beginning to emerge, Graham acknowledges that there is an “obvious publication bias,” meaning if researchers discover a technique is less than cost effective, they may be “less than enthused” about sending it off to be published. Therefore, there is not an abundance of clinical studies which disprove the effectiveness of various imaging technologies.

CER gains popularity in U.S. with administration shift

More than 15 years ago, the U.S. Office of Technology Assessment published a report discussing the benefits and challenges of creating a CER initiative, but it has struggled to gain footing. Yet, when the Obama Administration signed the ARRA into law in February, allotting $1.1 billion for CER, the field clearly got a shot in the arm.

Out of the 100 priority CER projects identified by the Institute of Medicine, diagnostic imaging was named in 11, highlighting the use of PET for the diagnosing, staging and monitoring of oncology patients.

Along with the ongoing healthcare debate in the U.S. Congress, this funding has led policy leaders and invested parties to debate just what constitutes proper CER programs and projects, and which type of entity should assess the data. A series of New England Journal of Medicine editorials have deliberated the merits and drawbacks of randomized controlled trials compared with patient registries, and expressed concern about how U.S. government agencies will respond to the data.

Rep. Michael Rogers, R-Mich., criticized CER that will be used to “deny or ration care,” especially if there is “no restriction on how the federal government can use this research…This is a dangerous open-door policy that will almost certainly lead to comparative-effectiveness research being used to make coverage decisions,” he says.

In response to the clamoring surrounding U.S. government-sponsored CER, Graham says it is an “important direction for research,” stressing the importance of study design, but cautioning that the “results of any one study should not be utilized for absolute determination for reimbursement. This is where the concern arises.”

Currently, one problem with evaluating medical imaging is that the technologies are assessed in isolation. “The government only sees that medical costs are increasing, and seeks to put a stop to it, but they are not examining how it is potentially saving money through preventing futile surgery,” Graham says. “For instance, surgery costs the U.S. healthcare system between $30,000 and $50,000, which could be prevented through a $2,000 medical imaging study. The government agencies are looking at the patient care cycle in segregated silos. These evaluations need to be done in a coordinated manner, and the data need to be properly disseminated.”  

Despite these concerns, U.S. CER projects are already underway, such as the $4 million granted to University of Washington School of Public Health to evaluate the effectiveness of cancer diagnostics, such as PET/CT and blood- or tissue-based biomarkers, to determine the extent of disease and plan treatment.  

Also, the National Oncologic PET Registry (NOPR) is a “remarkably successful patient effort, mainly due to the large numbers of patients,” that numbers more than 100,000, according to Graham.  

The “practice-based evidence” collected through NOPR was influential, in conjunction with other data, in CMS’ decision to “substantially expand” PET coverage for the initial evaluation of many more cancers, says NOPR Co-Chair Siegel. In early April,

CMS issued a final national coverage determination to expand coverage for initial testing with PET for Medicare beneficiaries who are diagnosed with and treated for most solid tumor cancers.

Since then, the objectives of the NOPR protocol have broadened, according to Siegel, as the researchers attempt to reach beyond a change in clinical management to more closely relate the PET results with a change in patient outcomes.

For instance, the investigators will assess how PET plays a role in changing a patient’s care plan from active treatment to palliative treatment, in trying to compare a cohort of patients managed without PET information with a cohort of patients from the NOPR database after PET became reimbursed. “We will be looking to see if there was any improvement in patient outcomes once PET became available for their management,” Siegel explains.

He acknowledges the limitations of using registry data, as this does not provide an appropriately definable control group, and therefore it is difficult to prove improved patient outcomes. “Historical controls are not strictly comparable to current patients,” he says. “However, very few of our clinical decisions in medicine are based on large, properly-powered, randomized, controlled trials.”

Currently, NOPR is collaborating with investigators at Dartmouth, which received a two-year grant for a CER project that will “ideally lead us to closure with CMS, achieving even broader PET coverage, with whichever exceptions may be stipulated by the agency,” Siegel notes.

With all the data being collected, Siegel questions whether this will lead to changes in practice on its own, or will lead to changes in government policy—the latter of which could force the changes in practice. 

Lessons learned

While the United States is just launching CER programs, other countries have well-established CER entities which serve a variety of purposes:

• United Kingdom: The National Institute for Clinical Excellence, established in 1999, provides guidance on public health, healthcare technologies and clinical practice to “promote clinical and cost effectiveness” throughout NHS, by means of guidelines and audits.
• Germany: The German Institute for Quality and Efficiency in Healthcare, which initiated operations in 2005 as part of national healthcare reform to identify quality standards by assessing clinical effectiveness. In 2007, the institute expanded its role to incorporate cost-effectiveness data.
• Australia: Pharmaceutical Benefits Advisory Commission was the first reimbursement committee globally in the late 1980s. Currently, it approves all new drugs, and advises the government on appropriate prices.
• France: The French High Health Authority, established in 2005, provides accreditation, develops guidelines, promotes health IT tools and informs the national insurance systems about listing and reimbursement. Since 2008, the entity also has assessed the adoption of cost-effective technologies.*

However, simply because these entities are firmly established doesn’t mean all the quandaries are solved. In fact, Knapp says the “restrictive” national German policy will only reimburse a procedure if a prospective, randomized trial proves it is cost effective. Due to the costs and time required with these trials, various organizations have been more focused on collecting data on real-life clinical practice. 

It takes a (global) village

Molecular imaging organizations have the capability to present a platform in which various healthcare facilities, across country and state lines, can collect large amounts of data. The universal understanding of nuclear medicine is already getting broader, exemplified by the international representation at annual conferences. For instance, at this year’s SNM conference, less than half of the presentations came from the United States—with Japan, Germany and Korea accounting for a combined 24 percent.

Certain CER studies have expanded their findings to assess the cost of procedures in other countries. One German CER study found that using PET in patients with enlarged lymph nodes on CT raised the incremental cost-effectiveness ratio to €36,667 per life years saved, but the costs of PET were almost balanced by a better selection of patients for beneficial cancer resection. Dietlein et al wrote that “[w]hen PET or CT were positive for mediastinal lymph nodes, the exclusion from biopsy confirmation led to cost savings that did not justify the expected reduction in life expectancy” (Nuklearmedizin 2001;40(4):122-8). Then, the authors also assessed the economic data from the U.S. and Japan to show that, in those settings, PET-based algorithms were cost effective for the management of lung tumors.

Increasingly, organizations are pooling data gathered from various facilities, especially through patient registries. However, unifying the data does require appropriate quality control, such as phantom studies in each facility, allowing researchers to correct for individual deviations.

“If you want to compare data with those from Brussels, Stockholm and Berlin, it is not possible nowadays because of the different machines, software and reconstruction algorithms. As a result, a phantom must be in place to quantify the data collected,” Knapp explains.

However, Siegel questions whether patient registries, like NOPR, could work on a multinational scale, due to the variations in PET reimbursement policies. He suggests that large, prospective trials—randomized or not—would work better on a multi-national level, adding that ACRIN is actively working to enroll sites in Asia and Europe for its trials.

While Siegel does not see the phantom as an impediment to conducting multinational trials or registries, he cautions against over-standardization because it is not reflective of real-world clinical practice. “Also, running those phantoms requires expenses to the trial, and effort for the participating facilities,” he says.

Yet, EANM has established a platform for a European-wide neurology patient registry, so now researchers could easily launch multicenter oncologic registries because the phantoms are already in place. The neurology study, which is concluding at the year’s end, sought to define normal values for the uptake of a radiopharmaceutical for the diagnosis of Parkinson’s disease.

Also, the European Organization for Research for Treatment of Cancer, along with the EANM, is planning a study to assess whether C-11 choline or C-11 acetate are effective in the staging or restaging of prostate cancer.

The SNM also is seeking to organize a patient registry for reimbursed PET oncology studies that would allow the organization to supplement the database with outcome information after a couple years, in order to establish which procedures are most cost-effective. For this endeavor, SNM will employ its Clinical Trials Network, which currently has approximately 220 U.S. sites enrolled with approximately 10 sites qualified.

If the organization receives Agency for Healthcare Research and Quality funding, SNM will seek to organize a conference, addressing CER in molecular imaging in oncology that would tentatively take place in the spring or summer of 2010. 

The potential

In his closing remarks at the 2009 SNM conference, Henry N. Wagner, MD, from Johns Hopkins Bloomberg School of Public Health in Baltimore, urged attendees to become involved “not only in basic and clinical science, but also in the economics and politics of healthcare. We must examine and publish cost and effectiveness data as well as data about the clinical value of molecular medicine.”

Wagner added that CER, “will develop at an accelerating rate, [and] provide us with enormous opportunities to advance the promise of molecular medicine… Increasing knowledge is expensive, but making correct decisions decreases the cost of caring for each individual patient, thereby making it possible to increase productivity.”

*Information from “Comparative Effectiveness Research: International Experiences and Implications for the United States,” Kalipso Chalkidou, MD, PhD, July 2009.