www.cmtpnet.org Building a Strategic Framework For Comparative Effectiveness Research in Oncology NATIONAL LEADERSHIP SUMMIT ON CER PRIORITIES, METHODS AND POLICY Final Report Center for Medical Technology Policy September 2011 NATIONAL LEADERSHIP SUMMIT ON CER PRIORITIES, METHODS AND POLICY
64
Embed
Building a Strategic Framework For Comparative ... · [Type text] DRAFT – DO NOT CIRCULATE [Type text] Building a Strategic Framework For Comparative Effectiveness Research in Oncology
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
[Type text] DRAFT – DO NOT CIRCULATE [Type text]
www.cmtpnet.org
Building a Strategic Framework For Comparative Effectiveness Research in Oncology
NATIONAL LEADERSHIP SUMMIT ON CER PRIORITIES, METHODS AND POLICY
Final Report
Center for Medical Technology Policy
September 2011
NATIONAL LEADERSHIP SUMMIT ON CER PRIORITIES, METHODS AND POLICY
About CMTP ................................................................................................................................................ 10
Learning Healthcare Systems .................................................................................................................. 39
Engaging Community Practice ................................................................................................................ 40
The Economic Imperative ....................................................................................................................... 40
Keynote 3: Reimbursement Reform, Dr. Peter Bach ................................................................................. 43
Session 6: Value and Costs in Oncology ..................................................................................................... 43
Patients’ Role in Value Decisions ............................................................................................................ 44
The Role of Cost ...................................................................................................................................... 45
Developing Costs and Value Information ............................................................................................... 46
CER and Value-Based Pricing .................................................................................................................. 46
such as the variations in treatment efficacy among different patient populations, will be necessary
to maximize the effect of treatments. In the public arena, architects of the CER enterprise will need
to allay concern that CER will be used to ration access to genetic testing and novel targeted agents.
Pharmacogenomic studies illustrate the importance of a partnership between stakeholders in CER.
Pharmacogenomic tests do not have a defined approval process and evidence of clinical utility is
often lacking. The CER enterprise will promote co-development of agents and genetic tests and help
to ensure incorporation of patients’ goals, values, beliefs, and expectations into study design and
analyses. Incentives motivating test developers may not be aligned with patients’ values, requiring the
research community to play a proactive role in identifying promising candidate tests for CER and in
enlisting biomedical partners in generating evidence.
5. CER results must be translated into clinical practice.
In evidence-based healthcare, findings from up-to-date, high-quality research inform clinical care.
The research effort will be wasted unless obstacles to evidence translation are addressed. Improved
methods with which to promote translation will increase the ability of clinicians to implement
change. Clinical practice guidelines and quality initiatives provide potential mechanisms to integrate
emerging data into practice, yet their development requires additional funding and coordination.
Attention to various aspects of evidence development will facilitate translation of CER results. First,
the generation of reliable data requires the enrollment of patients from community settings who are
representative of real-world clinical populations, as the vast majority of patients receive care there.
Gathering data in community settings is challenging due to wide variations in practice patterns,
limited systematic data collection, and a lack of financial incentives for community-based clinicians
to participate in research. Despite these issues, initiatives such as the NCI Community Cancer
Centers Program (NCCCP) have made progress.
9
Second, CER translation depends upon infrastructure development. Re-alignment of payment
incentives would support the development of the information technology (IT) infrastructure.
Similarly, the development of methodological and data standards would provide a coordinated
framework for data acquisition. The IT infrastructure must accommodate patient-reported
outcomes (PROs), as these data are increasingly recognized as a central feature of CER. Feedback
loops that connect researchers, guideline panels, and practitioners must be “hard-wired” through IT
pathways to ensure that information flows bi-directionally to support formulation of clinically
important research questions, hypothesis generation, study conduct, translation of results, and
evaluation of impact. This is the starting point for the development of rapid learning systems in
oncology.
6. The CER enterprise must address cost and value.
While Congress limited consideration of cost and value in ARRA and PPACA, they will eventually
demand attention. Defining cost and value is not straightforward. Even if the effect of a given
treatment on survival is known, its true value to a patient is often much more difficult to define. At
the same time, direct and indirect costs are not captured in a systematic way, leading to difficulties
in analyzing cost drivers. It is important that physicians, researchers, payers, and policy makers avoid
substituting their own value judgments for those of patients and that costs not be narrowly defined.
Learning how to capture patients’ values and to incorporate this aspect of treatment evaluation into
research requires improved methods and a continuing public conversation. Ultimately, CER will need to
incorporate nuanced metrics of patient value that include measures of patient experiences, out-of-
pocket costs and other cost and value dimensions. New systems will need to be developed to
consistently and reliably capture these data.
The Summit succeeded in initiating a conversation surrounding the direction of CER in oncology, and
included a broad range of stakeholders with diverse visions in the conversation. Convergence was
seen on a strategic framework that has the potential to accelerate CER in oncology and, thereby, to
improve the quality of cancer care.
This conversation has just begun and the stakeholders involved in the process must continue the
dialogue. CMTP will do its part to foster this conversation. The first strategy is to appeal for
reactions to this report and request suggestions for further enhancing CER in oncology. As the
discussion evolves, we will make it accessible to the public through our website (www.cmtpnet.org).
10
About CMTP
The Center for Medical Technology Policy (CMTP) is a private, non-profit organization that provides a
neutral forum in which patients, clinicians, payers, manufacturers, and researchers can work together to
design and implement prospective, real-world studies to inform health care decisions. CMTP focuses on
a range of methods for evaluating comparative effectiveness, including pragmatic trials, adaptive
designs, clinical registries, and other study designs that generate evidence to provide patients, clinicians
and payers with a reasonable level of confidence in their decision-making. The primary goal of CMTP is
to improve the process for generating reliable and credible information about the real world risks,
benefits, and costs of promising new medical technologies.
Project Staff
Sean Tunis, Director
Robert B. Giffin, Senior Research Director
Laura Esmail, Senior Project Manager
Sharon Murphy, Senior Advisor
Daniel Mullins, Research Fellow
Bradford Hirsch, Research Fellow
Russ Montgomery, Research Associate
Swapna Karkare, Research Associate
Merianne Spencer, Research Associate
Julie Simmons, Event Planner
Acknowledgements
Through their dedication and hard work, many individuals helped to make this meeting and the resulting
report a success. We would like to first acknowledge the many individuals who participated in the
advisory committee that guided the program before the meeting and helped to shape this final report
(see Appendix A). We would also like to acknowledge the federal agencies and private organizations that
played an important role in providing advice and expertise as the project developed. A huge thanks goes
to all of our speakers, moderators, response panelists and workgroup leaders whose willingness to share
their ideas led to an incredibly rich discussion at the summit (See Appendix C). Finally, we extend a very
special thanks to the sponsors of this program, without whose commitment and support, the Summit
would not have been possible. They include Amgen, the Blue Cross and BlueShield Association, Eli Lilly,
Genentech, GlaxoSmithKline, Humana, Pfizer, and Sanofi-Aventis.
11
Introduction
In 2011, instead of being based on solid evidence regarding the harms and benefits of interventions, the
practice of medicine in America is still, to an alarming degree, based on beliefs and traditions that have
evolved as a result of clinical experience and expert consensus,. While there has been undeniable
progress, there are examples of the negative impact of failing to base patient care on robust evidence,
such as the use of high-dose chemotherapy and bone marrow transplant for metastatic breast cancer.
Clinicians utilized this aggressive and costly intervention until randomized controlled trials (RCTs)
demonstrated a lack of superiority to conventional chemotherapy. This and similar examples clearly
underscore the importance of basing treatment decisions on sound evidence. Unfortunately, there is
limited evidence to guide physicians and patients for many clinical scenarios. For example, despite a
number of treatment options for early-stage prostate cancer—such as surgery, external beam radiation,
and watchful waiting -- little is known about which is most effective and the impact of each on a
patient’s quality of life.2
Advanced as a paradigm for effective healthcare since the 1990’s, evidence-based medicine (EBM)
requires the availability of sufficient, clinically relevant data to inform healthcare decisions.3 With the
United States (US) healthcare system in crisis, more and possibly different types of evidence are needed
to enable truly evidence-based care. The 2010 Patient Protection and Affordable Care Act (PPACA)
included the goal of increasing the systematic creation and use of evidence in the delivery of care.4 Its
approach builds on the already well-established fields of health technology assessment and EBM. PPACA
provided for the creation of the Patient Centered Outcomes Research Instituted (PCORI), which
commissions research meant to help patients and their healthcare providers make informed decisions.
PPACA and PCORI advance the federal investment in evidence development beyond that of the
American Recovery and Reinvestment Act (ARRA) of 2009, which appropriated $1.1 billion for
comparative effectiveness research (CER).5
The concept of CER was born out of the desire for robust evidence that is directly relevant to real-world
clinical decision-making. The Federal Coordinating Council has provided the following definition of CER,
which has been endorsed with minor variations by the Institute of Medicine (IOM) and other national
agencies:
CER is the conduct and synthesis of research comparing the benefits and harms of
different interventions and strategies to prevent, diagnose, treat and monitor
health conditions in “real world” settings. The purpose of this research is to
improve health outcomes by developing and disseminating evidence-based
information to patients, clinicians, and other decision-makers, responding to their
expressed needs, about which interventions are most effective for which patients
under specific circumstances.1
12
CMTP held a two-day conference on November 1-2, 2010, to address opportunities, challenges and
approaches to CER. Participants represented academic medicine, professional/medical associations,
federal agencies, industry, and the non-profit sector. The Summit was organized into six sessions
focused on the following topics: (1) Introduction/Setting the Stage, (2) Engaging Patients in CER, (3)
Improving Clinical Evidence for Oncology, (4) The Critical Challenge of Genomics and Personalized
Medicine, (5) Translating Evidence into Practice, and (6) Value and Cost in Oncology. The proceedings
presented herein adhere to this structure and present the main themes of each session.
Session 1: Setting the Stage
Sean Tunis, Founder and Director of CMTP, opened the Summit by discussing the importance of CER and
the need for concrete, actionable goals that can move the field forward and guide the PCORI. He
described its transformative potential, highlighting the dire need for actionable evidence today. As an
example of the lack of reliable evidence with which to make clinical decisions, he highlighted the recent
systematic review of treatment modalities for early-stage prostate cancer performed by the Agency for
Healthcare Research and Quality (AHRQ).6 (Table 1) Despite the availability of hundreds of clinical trials
on this topic, the investigators conducting the review could not find sufficient evidence to recommend
specific modalities.
Table 1: Comparison of strength of evidence for radiation treatment options in clinically localized prostate cancer (adapted from the AHRQ Technology Assessment, August 2010).
6
Comparisons Disease specific
survival
Freedom from
biochemical failure
GU/GI toxicity
No treatment Insufficient Insufficient Insufficient
Figure 2: Criteria defined by GRADE working group to score the quality of evidence While RCTs start with high evidence and observational research with low, the grade can be recalculated according to the defined criteria.
SOURCE: Reprinted with permission from the BMJ Publishing
Group Ltd: Atkins D, et al. Grading quality of evidence and
strength of recommendations. Grade working group. BMJ
2004; 328:1490.
Several initiatives are also attempting to strengthen the design of OCER studies. In 2007, AHRQ developed a
guidance document entitled Registry for Evaluating Patient Outcomes: A User’s Guide.11 It was meant to
establish quality standards among registries and to provide a checklist of good and bad practices. While not
specific to CER, it effectively addressed many of the important issues. The Good Research for Comparative
Effectiveness (GRACE) Initiative, also started in 2007, seeks to enhance the quality of OCER by outlining
good practices.12 In addition to these activities, the statute that created PCORI charges its Methodology
Committee with developing a translation table to guide the selection of the most appropriate methods to
address different CER questions.
Oxaliplatin for Stage III Colon Cancer: Example of a Registry Study
Registries are being used effectively for OCER, but various challenges must be overcome. To illustrate
the issue, Deborah Schrag discussed her experience conducting a registry-based CER trial. She sought to
evaluate the comparative effectiveness of adding oxaliplatin to adjuvant chemotherapy for stage III
colon cancer. She defined for the audience the steps that she found critical to the success of her trial: 1)
identifying a priority question; 2) reviewing the efficacy data; 3) framing the question; 4) assessing if the
intervention is in use; 5) examining the trends in outcomes; 6) framing the analysis (PICOTS); 7)
assembling the data sources and obtaining approvals; 8) defining the cohorts; 9) ascertaining the
26
treatment groups; and 10) evaluating the similarity of the cohorts. She then described each in the
context of the trial.
The first three steps involved developing the research question. In Dr. Schrag’s case example, review of
the available data from earlier trials showed that adding oxaliplatin to a 5-fluorouracil (5-FU) backbone
resulted in superior efficacy as compared to 5-FU alone. There were, however, substantial tradeoffs in
terms of toxicities13 and it was unclear if the prior study results were representative of real world
practice. After reviewing the evidence, the research team engaged stakeholders to help frame the
analysis and to develop the research question. The goal of this inclusive process was to ensure that the
concerns and values of decision-makers would be adequately addressed by the resulting study.
After framing the question, an assessment of the study feasibility was needed to determine whether the
data contained in available registries were sufficient to answer the question. Dr. Schrag used data from
the Medicare- Surveillance, Epidemiology and End Results (SEER) database to evaluate patterns of care
and to ensure that oxaliplatin-based regimens were used widely enough to make conclusions. She also
examined trends in outcomes and identified possible covariates to include in the analysis. A trend was
seen toward better survival, providing positive initial feedback on the hypothesis. Following this
background work, Dr. Schrag framed the analysis using the AHRQ-specified PICOTS framework and the
GRACE principles.
Assembling data sources and obtaining approvals proved to be the most difficult and time-consuming
step in the study. These steps accounted for approximately 85% of the overall project time. A high
degree of political skill was required in dealing with the many relevant stakeholders, numerous IRB, and
many other individuals whose consent was required. Ultimately, the project included trial data from the
Accent Database; efficacy data from Medicare, the Cancer Care Outcomes Research and Surveillance
(CanCORS) Consortium, and the National Comprehensive Cancer Network (NCCN); and state-level data
from California and New York.
With data sources and approvals secured, the next step involved defining the cohorts and ensuring the
application of consistent criteria to each dataset, in order to maximize comparability and transparency.
This step required engagement with the curator of each dataset to understand the unique attributes of
each. The previously defined covariates were used to assess the similarity of cohorts as well as
heterogeneity across cohorts, using techniques such as propensity score matching.
The development of a statistical analysis plan appropriate to the specific question at hand is a critical
step in the design of any high-quality research study. Dr. Schrag used an array of methods to analyze the
data, including univariate and multivariate analyses, propensity score weighting, and inclusion/exclusion
of subgroups. Selection of analytic methods was followed by a priori specification of a sensitivity
analysis, considering factors such as time horizons, treatment assignments and outcomes.
Upon completion of the analysis, the final steps were to share the results with the stakeholders, gather
their responses, and then reanalyze the assumptions and results. In the end, the study revealed a
27
survival advantage that became apparent approximately two years after the initial treatment. The
benefit was small but statistically significant. Dr. Schrag admitted that “the stakeholders knew what they
were talking about when they wanted us to ask this question.” Nonetheless, the study accomplished an
important purpose in helping to distill the steps needed to complete a registry-based CER study in
oncology.
Reflecting on the project, Dr. Schrag stressed the importance of being able to leverage existing data. She
cautioned that the conduct of such research is difficult and time-consuming, leading some to question
the utility of large-scale, registry-based studies unless the process can be streamlined. Systems to
facilitate efficient data collection and analysis and to ensure a high level of quality while maintaining the
privacy of patients, could enhance the role of registry-based CER. Despite concerns, the promise of
OCER to increase the quantity and quality of available evidence is very high.
Development of Efficient Data Systems to Support CER
Electronic data systems provide an unprecedented research opportunity. If data quality, accessibility,
interoperability and governance issues are addressed, progress could be made toward implementing a
rapid learning health system in the US. Peter Yu outlined the steps required to allow the use of
electronic databases for CER. He described the electronic medical record as “an opportunity to redesign
how we approach the collection of data.” In an effort to make physicians comfortable with the transition
from paper-based to electronic capture of clinical data, vendors have tried to replicate the paper chart.
Data captured this way are idiosyncratic, consisting of personal notations, abbreviations and
inconsistent context which limit its usefulness and reliability.
Data interoperability (i.e., the ability to coordinate data from disparate and heterogeneous datasets) is a
key challenge in developing systems that can support CER. Interoperability requires changes in existing
information systems, such as standardization of data formats so that it can be easily pooled and
interpreted. The research community must also agree on the common data elements to be collected,
with standardized terminology and values for each element such as disease stage, patient comorbidities,
chemotherapy details and response criteria.
This work has already begun. The American Society of Clinical Oncology (ASCO) and the NCI have agreed
on many of the data elements that are important to oncology. The Certification Committee for Health
Information Technology (CCHIT), created by the Bush Administration with the mission of accelerating
the adoption of health IT, is currently conducting an assessment of the oncology certification process,
with a report expected in 2011. The IOM and the Office of the National Coordinator (ONC) also
identified three critical elements needed to allow progress in the field, two of which were addressed
through health reform legislation. These include requirements for the adoption of electronic health
records nationally and programs to enable semantic interoperability between systems. The final step
will be to define a way to aggregate and analyze data. Addressing issues of privacy, data ownership and
consent will be critical to success.
28
Dr. Yu concluded by outlining two approaches to issues of privacy and data ownership arising with
interoperable electronic data systems. In a cooperative group model, local data is extracted and
maintained in a centralized data repository. While this approach avoids many of the pitfalls of data
sharing, it is resource intensive, can result in poor data quality due to inconsistent methods of
extraction, and may lead to complexities with the release and sharing of information. The alternative
approach, a distributed data network, avoids redundant systems as the data is held, curated and
controlled through a local provider network. Security and privacy are maintained by provider
organizations but queries can be sent to a provider’s database. Dr. Yu advocated the latter, but
acknowledged that there is a lack of universal agreement on the best approach.
As electronic health records mature and interoperable, standardized data systems evolve, the
opportunity to conduct efficient and reliable observational studies will increase. Progress toward this
future state will require significant effort, funding and collaboration.
29
Discussion and Key Themes – Improving Clinical Evidence for Oncology
The following key themes were distilled from the presentations, panel discussions, and breakout
sessions, with input from the multi-stakeholder planning committee:
1. In oncology, there is an urgent need for an evidence development system which focuses
on the comparators of importance, outcomes of interest and the types of trials that
should be conducted. There has been an explosion in the number of available agents to
treat patients as well as an exponential rise in the number of trials assessing treatment
regimens. The research community must prioritize studies and methods that will lead to
improved clinical care.
2. Before beginning a CER study, its relevance and ability to answer the question at hand
must be established. Sponsors and research teams should be required to address two
questions: What is the probability that the study will definitively answer questions that
matter to patients? Will the study fill an evidence gap, such as those highlighted through
systematic reviews? This must be a stakeholder-driven process.
3. RCTs alone are not adequate to address the enormous evidence gaps in oncology. RCTs do
not always reflect real-world experience, as they are conducted in tightly controlled
environments. Because of their expense, complexity, resource-intensity, and timelines, they
can only address a limited number of questions. Observational studies can potentially fill
the gaps left by RCTs, but stakeholder skepticism regarding this less rigorous design must be
overcome. Defining research and evidentiary quality standards so as to ensure good practices
and reliability could help allay concerns.
4. Electronic health records should be designed to facilitate CER. Currently, the format and
content of electronic health records do not simplify or expedite the tasks entailed in CER.
The data captured in the current electronic health records are idiosyncratic, consisting of
personal notations, abbreviations and inconsistent organizational structures. To move
forward, researchers must define the data elements that need to be captured and
standardize the relevant terminology and values.
5. Conduct of large, registry-based studies will require a culture of collaboration. Oftentimes,
the rate-limiting step in registry-based CER revolved around assembling the data sources
and obtaining approvals. Greater efficiency and an overlying system to facilitate analysis is
needed.
30
Session 4: The Critical Challenge of Genomics and Personalized Medicine
Genomics could transform the delivery of medical care by allowing patients to receive only those agents
likely to provide benefit according to their unique profile. This vision is consistent with the goals of CER.
The potential benefits of genomically-based approaches to clinical care are widely acknowledged, but
important questions remain unanswered. For example, how will genomics be integrated into clinical trial
design without driving the costs of research prohibitively high? How should tissue be banked for future
study? When are biopsies appropriate in clinical trials? This session addressed these topics.
An Evolving Framework for Drug Development
The current approach to drug development involves identifying a molecule, demonstrating that it works,
conducting trials in humans and after establishing clinical activity, trying to identify which sorts of patients
are most likely to benefit. According to Neal Meropol, this approach is no longer viable and should be
replaced by a system in which drug development and diagnostic development proceed hand in hand. Dr.
Meropol also noted that the IOM identified the role of biomarkers in oncology as one of the top quartile
areas in need of CER. Stephen Eck reinforced this point with the observation that many treatments
delivered today do not benefit patients to the extent one might hope because patient populations are not
uniform and a “one size fits all” approach fails to account for these differences. Based on genetic variations,
patients are likely to respond differently to the same drug. 14
In addition, it has taken decades to improve treatment options for many types of cancer and a better
understanding of genetics may expedite this process. Dr. Eck described how treatments evolve. When a
new disease is discovered, it is initially treated with “non-technology” in which patients receive only
supportive care and symptom control. At the outset of his fellowship, treatment for multiple myeloma was
largely supportive. Next, “half-way technologies” are introduced. These new and exciting interventions tend
to be overused, and entail uncertainty about their best use. Though they postpone death, their mechanism
of action is poorly understood and their benefit is limited. Eventually “high technology,” the ultimate
objective, is developed. With these treatments, the effect size can be large and patients may be cured. In
this stage, we know which treatments should be provided to a given individual. Different cancers rest at
different positions along this developmental trajectory. In kidney cancer, tremendous progress has been
seen over the past decade, yet we have a limited understanding of the order in which to use the available
treatments and it is difficult to differentiate between indolent and aggressive disease. This disease remains
at the halfway point in the developmental framework. Hopefully, better molecular approaches will
accelerate this change.
31
New Research Strategies to Advance Drug Development for Genomic Medicine
Genomic research will require new methods that are capable of handling the intrinsic complexity of the
field. Dr. Meropol noted that it is often difficult to answer CER questions in oncology with only phase III
trials. It is also challenging to find financial support for the development of tests that identify which patients
will benefit from a specific drug, as industry does not want to limit the potential pool of candidates for a
given treatment.
Interpreting the results of studies is also difficult. Scott Ramsey provided an example of the complexity
clinicians face when applying the results of genomic studies to practice. In a recent study, researchers
claimed that they had identified a group of genetic sequences which predicted an increased risk for prostate
cancer with an odds ratio of 9.46.15 While the claim sounded impressive, after accounting for age and family
history, the predictive power of the test only identified a 0.02% increase in the risk of cancer. Hence the test
is not useful. This illustrates the importance of sophisticated interpretation of genomic studies.
The use of pooled analyses, decision modeling and other emerging approaches can help make sense of this
complexity and will move treatments in various diseases toward the “high technology” state. Two relevant
observations are that (1) it is often difficult to gather enough information in a single study to reach a sound
conclusion, and (2) there is increasing availability of tissue with excellent clinical annotation with which to
study various patient populations. Dr. Meropol used the example of Rat Sarcoma (RAS) and Rapidly
Accelerated Fibrosarcoma (RAF) mutations to elucidate the benefits of pooling studies and leveraging tissue
samples across repositories. RAS mutations predict resistance to treatment with Epidermal Growth Factor
Receptor (EGFR) inhibitors. But a downstream gene, RAF, also plays a role. The question was asked, should
patients with a wild-type RAS but a mutated RAF gene still get EGFR inhibitor therapy? RAS is mutated in
40% of colon cancer patients, while RAF is only abnormal in 8% of patients. To look at RAF is difficult enough
in a prospective RCT; an even larger study is needed to look at RAF in combination with RAS. Yet, one can
take tumor samples and outcome data from multiple trials, study genetic sequences across populations and
then pool the results. Such a study has been conducted and the results demonstrated that a RAF gene
mutation does not confer complete resistance to EGFR inhibitors.16 A more recent study further shows that
the location of the mutation on a RAS gene might affect outcomes with an EGFR inhibitor.17
Incorporation of Patient Values into Genomic CER
Patient decision-making is a critical consideration in genomically-based personalized medicine and thus, in
CER, meant to inform it. Dr. Meropol raised several real-world questions, such as “How do we develop
evidence and build decision models that take into account a patient’s goals, values, beliefs and
expectations?” A continuum exists of how patients value different factors such as survival and quality of
life. According to Dr. Meropol, patients “weigh decisions based on side effects, cost, inconvenience of
therapy and each patient may have a different end point that matters to them. Is it living two months to
see my daughter get married?” The traditional focus purely on survival needs to be modified to factor in the
relevant considerations.
32
Dr. Meropol discussed stage II colon cancer as an example. While many stage II colon cancer patients can
be cured with surgery, about 20% will relapse. Adjuvant chemotherapy may reduce the risk of relapse, but
only for 3% to 4% of patients. This relatively small benefit does not justify universal treatment. 18 Who
should receive the six-month course of chemotherapy? Dr. Meropol and his collaborators are developing a
genomic-based recurrence score to help identify who will benefit and who will not. A decision model can
also be built that takes into account more than just overall survival, helping to account for a patient’s
perceived gains and losses. CER will play an important role in supporting this type of complex analysis,
modeling, and personalized clinical approach.
Critical Role of Genetic Diagnostics in Genomic Medicine and CER
While genetic tools for widespread screening are not yet a reality, prognostic and predictive tests are
being introduced with increasing frequency. Dr. Ramsey provided an example in colorectal cancer in
which the sub-population of patients with a specific genetic marker (KRAS wild-type) are more likely to
benefit from cetuximab, an EGFR inhibitor, than those with a mutation in the gene.19 This finding allows
targeted treatment of wild-type patients, increasing the drug’s efficacy and the associated clinical
outcomes. Despite isolated areas of progress such as this, the field is in a state of flux. Thousands of
tests will come to the market over the next decade and it is unclear who will drive their uptake, how
clinicians will know when to use them and how insurers will decide when to pay.
In genetic testing, an established framework is used to help interpret the quality of available evidence
for a given test. Dr. Ramsey summarized the key attributes of the framework:
Analytic validity refers to how accurately and reliably a test reveals the presence of a specific
gene or trait. It is the “low-hanging fruit” as it is the minimum that should be expected of any
test.
Clinical validity refers to how often a test estimates a given outcome. For instance, if a test is
intended to provide prognostic information about prostate cancer, how often does it do so
accurately?
Clinical utility denotes how likely a test is to significantly change or improve a patient’s
outcomes compared to no testing. If a genetic test predicts response to chemotherapy, how
often does using it to guide therapy improve overall survival?
Value is the ultimate goal. This does not refer only to financial savings, but also to how often a
test provides value to a patient according to his/her priorities by improving quality of life or
other outcomes of interest.
33
Currently, diagnostic tests are often deficient in one or more of these criteria and there is limited
understanding of the quality of evidence underlying the tests used in practice. In a review conducted by
Regence Blue Cross and Blue Shield of 96,500 submissions for reimbursement of genetic tests, the total
cost to the insurer was $85,000,000 – nearly $1,000 per test.20 Despite these enormous costs, the
review found evidence of medical necessity for only five of the tests as shown in Table 4. "Why,” he
asked, “is it that we have so many tests that are developed and ultimately reimbursed without any real
understanding of their value?”
Table 4: Genetic tests with established medical necessity according to Regence Blue Cross and Blue Shield21
Test Indication
KRAS Predict response to cetuximab (Erbitux) & panitumumab (Vectibix) in
patients with metastatic colorectal cancer
BRCA1 & BRCA2 Detects inherited susceptibility for breast/ovarian cancer
Oncotype DX Determines breast cancer recurrence risk
APC, MYH, MMR Detects inherited susceptibility for FAP, Lynch Syndrome
MEN 2A, MEN 2B Detects inherited susceptibility for medullary thyroid cancer
The answer lies, in part, in the regulatory framework and its lack of rigor. Tests are often developed in ways
that do not answer questions that are relevant to patients and physicians. Once a promising genetic test is
discovered in an academic laboratory, it is likely to be licensed to biotechnology companies for clinical
validation. In the absence of a regulatory requirement to demonstrate analytic or clinical validity prior to
marketing, shortcuts are often taken to minimize expenses. Published results are likely to deal only with the
discovery phase, with marketing following soon after publication, as shown by the blue arrow in Figure 3.
Tests often bypass the FDA and are marketed under the Clinical Laboratory Improvement Amendments
(CLIA), which imposes minimal requirements. As a result, the clinical validity and utility of genetic tests
often remain unclear.
Figure 3: The current model of cancer genomic test development, including the relevant stakeholders, outcome at each stage, and limitations
SOURCE: Scott Ramsey presentation (November 1, 2010).
Reprinted with permission from Scott Ramsey.
34
Co-development of drugs and diagnostic tests is essential to the advancement of our understanding of
novel therapeutics. Dr. Shak discussed the co-development of HER2 testing with trastuzumab as an
example. Trastuzumab’s development involved investigators in fifteen countries, many of whom were
skeptical about the feasibility of drug and test co-development. Advocates helped design the trials and, in a
subsequent evaluation by statisticians, it was seen that if HER2 testing had not been conducted to better
understand response, the phase III trials would have been entirely negative. In Dr. Shak’s opinion, “this very
effective drug would never have been seen to have any activity…in the absence of the selection of
appropriate patients for testing.” This, he feels, is an important lesson for investigators. The results of the
correct test can greatly enhance drug effectiveness and utility. The encouraging example of trastuzumab,
however, does not yet represent the norm.
Dr. Shak also highlighted the Oncotype DX recurrence score, marketed by his employer, Genomic Health.
Unlike HER2, it was developed independent of a specific drug but is intended to help guide therapy. A
landmark study in the 1990s established the benefit of chemotherapy plus hormonal therapy, compared to
hormonal therapy alone, in early stage breast cancer.22 Because 85% of women were cured with surgery
and hormonal therapy alone, it was unclear whether it was necessary to expose all early-stage patients to
the toxicity of chemotherapy to benefit a few. Researchers heard from physicians, advocates, and patients
that guidance was needed and they set out to develop a test to fit this very specific purpose. Oncotype DX
helped to answer the question. Dr. Shak pointed out that one must be prepared to do multiple studies in
order to be convinced that a test is doing what it claims. Despite thirteen prior studies, OncoType DX is still
being evaluated in the TAILORx trial to define the benefit of chemotherapy in patients who fall in the
moderate risk category. 8
According to Dr. Eck, in the context of new drug development in a competitive market, CER could be used to
distinguish between true game changers and modest or incremental improvements and to clearly
distinguish economic value. Increased market pressures, with higher standards for market survival, will
encourage drug developers to shelve those agents that are marginally important and return to the lab
earlier in order to produce something superior. Dr. Eck also predicted an increase in head-to-head
comparisons because, with so many different mechanisms and side effects, the risks and benefits will need
to be truly understood. He cautioned, however, that there is a real potential for CER to require many more
studies and substantial additional investment in order to get drugs to market, thus adversely impacting
innovation and limiting the therapeutic options available to patients.
Collaboration to Advance Genomic and Personalized Medicine through CER
Advancement of genomic and personalized medicine through CER will require partnerships between test
developers, patients, regulators and clinicians in order to ensure that the right questions are being asked
and that the development proceeds in a coordinated way. Test developers will continue to be involved in
discovery, verification and standardization of assays. The research community must play a proactive role in
identifying promising candidate tests and in acquiring biomedical partners. Researchers must guide the
35
design and implementation of studies since the incentives of test developers are not always aligned with
those of patients. As shown in Figure 4, the health care system is a critical component as the source of both
patients and outcomes data. With cooperation among test developers, researcher groups and the health
care system, drug and test development will be far more likely to succeed.
Figure 4: Test developers, research groups and the health care systems must all play specific roles in order to improve the quality of tests in development.
SOURCE: Scott Ramsey presentation (November 1, 2010). Reprinted with permission from Scott Ramsey.
36
Discussion and Key Themes – The Critical Role of Genomics and Personalized Medicine
The following key themes were distilled from the presentations, panel discussions, and
breakout sessions, with input from the multi-stakeholder planning committee:
1. In oncology, CER must develop robust methods that address patient- and population-level
variations in treatment effects, values and access. Oncology, more than any other field of
medicine, relies on targeted therapies. The appropriate application of these agents depends
upon understanding differences in treatment effects based on patient characteristics. Fully
addressing these variations is expensive and difficult using the research methods currently
available to us. There remains concern that the focus will be on average benefits across large
patient populations instead of on patient-level differences. Creative approaches and vigilance
will be needed in order to ensure success.
2. Methods must be developed and implemented to efficiently deal with complex genetic
variation and the breadth of treatment options. Many questions cannot be addressed through
RCTs and large-scale surveillance studies. New approaches, such as pooled data analyses and
decision modeling, must be developed to address the extensive evidence gaps in oncology with
scientific and analytic rigor.
3. Evidence on the clinical utility of genomic tests is lacking despite its critical role in CER. Most
pharmacogenomic tests are not subject to the FDA approval process, making evidence of
clinical utility, though crucial to the appropriate use of these tests, generally lacking. Co-
development of drugs and diagnostic tests is essential, as the development of better drugs will be
facilitated by a better understanding of their effects on patients with specific genetic variations.
Co-development will necessitate incentives or requirements for the conduct of such research.
37
Session 5: Translating Evidence into Practice
Clinical Guidelines
Bottlenecks in the translation of evidence into clinical practice remain one of the greatest impediments
to evidence-based medicine. While the principal focus of CER is on the development of new evidence to
inform treatment decisions, it is critical that the CER enterprise also address translation so that findings
can impact clinical choices. Session 5 outlined a range of initiatives for developing this field including
activities in the areas of clinical practice guidelines, quality initiatives, learning healthcare systems and
engagement of community practices.
Clinical Practice Guidelines
Clinical practice guidelines serve as a vehicle for translating evidence into practice. Ethan Basch
discussed ways CER could enhance the impact of guidelines. Guidelines are often based on expert
opinion and the best available evidence: however, the evidence is often inconclusive. Over time, CER
could be used to build a more complete and robust evidence base through iterative and mutually
informative processes. Those who conduct reviews and develop guidelines can identify gaps in
knowledge and prioritize research questions, while researchers provide new evidence to inform reviews
and guidelines. By directly engaging clinical practices in this iterative process, the real-world nuances of
clinical practice can guide the prioritization of research topics. The vision of a continuous loop of
evidence generation, guideline update, practice improvement and research prioritization will lead to
improved patient care.
A number of examples of guideline development processes were discussed, with particular focus placed
on the speakers’ experiences with ASCO and NCCN. ASCO initially developed guidelines in 1994, focusing
on questions such as the use of chemotherapy in colon cancer, the use of anti-emetics during treatment
and the role of biomarkers. ASCO disseminates its findings through several outlets including The Journal
of Clinical Oncology, The Journal of Oncology Practice and web-based tools. Its goals are to provide
reliable recommendations and to disseminate knowledge in order to influence clinical practice,
formularies, and coverage. More than half of the top 20 downloads from The Journal of Clinical
Oncology website are ASCO guidelines, demonstrating their popularity.
Dr. Basch stressed the importance of establishing and adhering to evidentiary standards when
developing guidelines, stating that this rigor is equal in importance to the generation of evidence itself.
To him who devotes his life to science, nothing can give more happiness
than increasing the number of discoveries, but his cup of joy is full when
the results of his studies immediately find practical implications.
Louis Pasteur
38
Consistent standards allow reliability across evidence development groups and increased confidence in
their recommendations. ASCO’s process for guideline development relies on a systematic review of the
literature followed by a structured consensus process as outlined in its methodology manual—this
minimizes the sole reliance on expert opinion and makes potential conflicts of interest transparent. The
final step is an extensive peer review process.
Guidelines play a pivotal role in encouraging the appropriate use of tests and treatments, thereby
optimizing the use of scarce resources. By allowing only high quality evidence to influence guidelines, it
puts pressure on academia and industry to improve the processes of evidence generation (following the
Summit, the Institute of Medicine issued a report with recommendations for ensuring the quality and
integrity of Guidelines).23
William McGivney discussed the NCCN’s use of its compendia and guidelines to promote the
dissemination of clinical evidence. The NCCN draws from 21 cancer centers across the US to form 44
multi-disciplinary panels that develop treatment guidelines. Between 100,000 and 140,000 people visit
the NCCN website every month to access its recommendations and over 200,000 people do so per year.
A Genentech report showed that 84% of US oncologists use NCCN guidelines in their practice, while 60%
utilize ASCO guidelines, 27% internal guidelines, and 8% payer guidance. 24 In addition, payers such as
United Healthcare, Aetna, and CMS use them in making coverage determinations. These statistics
demonstrate the importance that the clinical community places on guidelines.
Quality Initiatives
Quality initiatives provide a mechanism for improving care while also identifying research questions
through their practice evaluations. As an example, Dr. Basch discussed the Quality Oncology Practice
Initiative (QOPI), which consists of a network of 600 oncology practices serving 900 sites. QOPI
participants conduct chart reviews to determine provider and organizational adherence to as many as
90 quality measures. Data are compiled and the results are used to understand practice patterns,
identify areas of concern and assess guideline adherence, with the ultimate goal of improving care and
addressing gaps in knowledge. Prior identification of deficiencies in three areas— infertility,
chemotherapy administration within two weeks of death and end of life discussions—has led to
education and research efforts to address the gaps. As another example, after ASCO released its
guidelines for the treatment of non-small cell lung cancer, QOPI helped to identify sites that deviated
from the updated guidelines and worked to understand the underlying reasons for non-compliance. Dr.
McGivney also discussed the NCCN collaboration with Ingenix, a large health informatics company, to
translate NCCN guidelines into quality rules that can be applied across tumor types. These systems
encourage adherence to evidence-based practices and help to ensure the delivery of high-quality care.
They must become more robust.
39
Learning Healthcare Systems
The idea of a learning healthcare system is garnering attention and widespread acceptance as a model
for continuous, iterative evidence generation and clinical practice improvement.25,26 The concept of the
learning healthcare system is motivated by three fundamental goals: (1) to generate high-quality
evidence and apply it to individual patients (i.e., to support personalized medicine); (2) to make
scientific discovery “a natural outgrowth of patient care” by using clinical experience to assess
effectiveness and generate hypotheses; and (3) to improve the value and efficiency of healthcare.25,26 In
practical terms, rapid learning healthcare describes a system in which data that are collected during
routine patient care are fed into an ever-growing databank and the system “learns” by frequent analysis
of the captured data. The new insights are used to inform subsequent care.27 CER, as research that
evaluates the relative benefits of available treatments under real-world conditions, is by definition
integral to learning healthcare systems.
A learning healthcare system and the CER enterprise that supports it requires a continuous feedback
loop (Figure 5). New evidence is used to develop or update guidelines which, in turn, improve and
update clinical practice. Multiple feedback loops at each step return relevant information to clinicians
and researchers, fueling further inquiry, discovery and translation, ultimately improving both clinical
outcomes and the functioning of the research enterprise. This continuous process defines rapid learning
healthcare. Dr. Basch highlighted the central importance of patient-reported outcomes (PROs) in rapid
learning systems. PROs add the patient’s perspective to the processes of question formulation,
prioritization and evaluation. Likewise, electronic health records that directly collect data about quality
from individual practices can be used to iteratively assess practice quality, identify research targets and
drive evidence generation.
Figure 5: Evidence generation in a learning healthcare system
Researchers face several challenges in quantifying costs and effectiveness. Steven Pearson outlined a
range of questions on the topic. Is effectiveness measured in terms of survival, surrogate outcome
measures, quality of life or events averted? How do we assess the meaning of time and relative health
to those who know they are nearing the end of life? Dr. Pearson stressed the need for explicit and
transparent elucidation of these values and suggested that a possible role for PCORI is to help define the
methods by which clinicians can elicit and incorporate preferences.
Unit costs and utilization rates represent the basic building blocks of CER analyses, however, they can be
difficult to quantify. The first challenge is to find reliable cost data. The second is to use these numbers
to express value in ways that make sense to patients and other stakeholders. While the cost literature
often refers to cost utility metrics, such as cost per QALY, these metrics may not be easily understood or
appropriately interpreted by patients and physicians. It may be easier for diverse stakeholders to
understand costs if they are reported using cost consequences such as “cost per adverse event
avoided.” However, because these sorts of costs are not always related to a final effectiveness measure,
they do not fully address the payers’ need to quantify the budget impact of a treatment. Dr. Pearson
illustrated this point using cholesterol lowering agents which, while cost-effective, become very
expensive when distributed to a large proportion of the population.
Even for a given test, cost-effectiveness evaluations by different researchers or in various settings can
lead to widely divergent conclusions. Dr. Pearson illustrated this using the example of colorectal cancer
screening. The Institute for Clinical and Economic Review (ICER) conducted an evaluation for the
Washington State Healthcare Authority and found that colonoscopies and CT colonography (CTC) were
equally effective. Since the unit cost for the CTC procedure was twice as high as that of colonoscopy in
Washington, coverage of CTC was denied in that state. By contrast, in Wisconsin, the same CTC
procedure was offered at one third the cost of colonoscopies, making it a high-value service that was
ultimately covered. Clearly, the definition of value is not inherent to a procedure, but emerges as a
product of numerous factors including costs, outcomes, patient preferences and other stakeholder
perspectives.
CER and Value-Based Pricing
Given concern among federal officials and the general public that CER will be used only to save money,
in parallel with the recognition that research is needed to help determine where costs can be trimmed
while still optimizing outcomes, the relationship between cost and value in CER must be more clearly
defined. Dr. Pearson suggested that CER be viewed, not as a method for determining whether or not to
cover a treatment, but rather as a method of generating evidence that can be used to guide pricing
policy. Decision-making related to coverage and payment rates for Medicare have historically been
47
distinct processes. In general, coverage is first decided upon, and then reimbursement rates are set.
This approach can result in high prices for newly covered drugs, despite a lack of evidence of increased
efficacy over the prior standard of care. The US currently lacks a method for relating price and
reimbursement to evidence demonstrating higher value. Dr. Pearson highlighted a number of recent
interventions such as proton beam therapy, erythropoiesis-stimulating drugs and off-label prescribing
that illustrate this point.
Several approaches could be used to embed calculations of value into determinations of
reimbursement, including global payments, episode-based payments, and least cost alternatives. Dr.
Pearson discussed a novel process in which available evidence drives pricing and coverage decisions.33
At the time when reimbursement is approved for a given treatment, the comparative effectiveness of
that intervention would be a primary factor in determining pricing. Lower pricing would be assigned to
a treatment with limited, equivocal or lower-quality evidence until new evidence is generated and
pricing is re-evaluated.
Assessments of value, though always difficult to conduct, are particularly challenging in a complex
interventional area such as oncology. Conducting research that evaluates value, however, is essential in
order to develop systems of care that will ultimately provide the greatest possible healthcare value to
patients and to the public.
48
Discussion and Key Themes – Value and Costs in Oncology
The following key themes were distilled from the presentations, panel discussions, and
breakout sessions, with input from the multi-stakeholder planning committee:
1. It is important for researchers to gain a better understanding of patients’ and
oncologists’ perspectives on cost and value. Congress limited consideration under
PCORI. General consensus holds that the value of a treatment depends on more than
its cost, though costs cannot be ignored. When attempting to assess or define value,
physicians, researchers, payers and policy-makers must be careful not to substitute their
own judgments for those of patients.
2. Methods for understanding cost and value must be further developed. Learning how
to capture and incorporate patient values, given the complexity of clinical practice,
requires the development of new methods and a continuing public conversation. Research
should incorporate nuanced metrics of patient value that include patient-reported
outcomes and out-of-pocket costs. Better systems must be developed to consistently
capture reliable data on treatment costs.
49
References 1. Federal Coordinating Council for Comparative Effectiveness Research. Report to the President
and Congress. Washington, DC: US Department of Health and Human Services; 2009 June 30. 2. Barry MJ. The prostate cancer treatment bazaar: Comment on "Physician visits prior to
What it is and what it isn't. 1996. Brit Med J. 2007;455:3-5. 4. Patient Protection and Affordable Care Act, H.R. 3590, 111th Cong., 1st Sess. (2010). 5. American Recovery and Reinvestment Act In, H.R. 1, 111th Cong., 1st Sess. (2009). 6. Ip S, Patel K, Obadan N, Chung M, Bannuru R, Lau J. Comparative evaluation of radiation
treatments for clinically localized prostate cancer: An update. Rockville (MD): US Department of Health and Human Services, Agency for Healthcare Research and Quality. 2010. Technology Assessment.
7. Abernethy AP, Raman G, Balk EM, et al. Systematic review: Reliability of compendia methods for off-label oncology indications. Ann Int Med. 2009;150:336-43.
8. Sparano JA. TAILORx: Trial assigning individualized options for treatment (Rx). Clin Breast Cancer. 2006;7:347-50.
9. Nice: Guideline development methods. Chapter 7: Reviewing and grading the evidence. London (UK): National Institute for Health and Clinical Excellence. 2004.
10. Atkins D, Best D, Briss PA, et al. Grading quality of evidence and strength of recommendations. Brit Med J. 2004;328:1490.
11. Gliklich RE, Dreyer NA, eds. Registries for evaluating patient outcomes: A user’s guide. Cambridge (MA): Outcomes Sciences, Inc. (d/b/a Outcome); 2007. Contract No.: HHSA29020050035I_TO1. Sponsored by the Agency for Healthcare Research and Quality.
12. Good ReseArch for Comparative Effectiveness. http://www.graceprinciples.org/. 13. Andre T, Boni C, Mounedji-Boudiaf L, et al. Oxaliplatin, Fluorouracil, and Leucovorin as adjuvant
treatment for colon cancer. New Engl J Med. 2004;350:2343-51. 14. Bardelli A, Siena S. Molecular mechanisms of resistance to cetuximab and panitumumab in
colorectal cancer. J Clin Oncol. 2010;28:1254-61. 15. Zheng SL, Sun J, Wiklund F, et al. Cumulative association of five genetic variants with prostate
cancer. New Engl J Med. 2008;358:910-9. 16. Laurent-Puig P, Cayre A, Manceau G, et al. Analysis of PTEN, BRAF, and EGFR status in
determining benefit from cetuximab therapy in wild-type KRAS metastatic colon cancer. J Clin Oncol. 2009;27:5924-30.
17. De Roock W, Jonker DJ, Di Nicolantonio F, et al. Association of KRAS p.G13D mutation with outcome in patients with chemotherapy-refractory metastatic colorectal cancer treated with cetuximab. JAMA. 2010;304:1812-20.
18. Figueredo A, Coombes M, Mukherjee S. Adjuvant therapy for completely resected stage II colon cancer. Cochrane Database Syst Rev. 2008;3:CD005390.
19. Lievre A, Bachet JB, Le Corre D, et al. KRAS mutation status is predictive of response to cetuximab therapy in colorectal cancer. Cancer Res. 2006;66:3992-5.
20. Medical Policy: Genetic and diagnostic testing. 2010. http://blue.regence.com/trgmedpol/geneticTesting/gt20.html. (Last accessed March 1st, 2011).
21. Regence Blue Cross and Blue Shield. Medical Policy: Genetic and diagnostic testing. 2010. http://blue.regence.com/trgmedpol/lab/lab20.html. (Last accessed March 1st, 2011).
22. Wood WC, Budman DR, Korzun AH, et al. Dose and dose intensity of adjuvant chemotherapy for stage II, node-positive breast carcinoma. The New Engl J Med. 1994;330:1253-9.
23. Clinical practice guidelines we can trust. Washington, DC: Institute of Medicine; 2011. 24. The Genentech oncology trend report: Perspectives from managed care, specialty pharmacy
providers, oncologists, and practice managers. 2008-2009 Oncology Trend Report. South San Francisco (CA): Genentech; 2009.
25. Etheredge LM. A rapid-learning health system. Health Aff. 2007;26:107-18. 26. Slutsky JR. Moving closer to a rapid-learning health care system. Health Aff. 2007;26:122-4. 27. Abernethy AP, Ahmad A, Zafar SY, Wheeler JL, Reese JB, Lyerly HK. Electronic patient-reported
data capture as a foundation of rapid learning cancer care. Med Care. 2010;48:S32-8. 28. Elkin EB, Bach PB. Cancer's next frontier: Addressing high and increasing costs. JAMA.
2010;303:1086-7. 29. Neuss MN, Desch CE, McNiff KK, et al. A process for measuring the quality of cancer care: The
quality oncology practice initiative. J Clin Oncol. 2005;23:6233-9. 30. Nadler E, Eckert B, Neumann PJ. Do oncologists believe new cancer drugs offer good value? The
Oncologist. 2006;11:90-5. 31. Neumann PJ, Palmer JA, Nadler E, Fang C, Ubel P. Cancer therapy costs influence treatment: A
national survey of oncologists. Health Aff. 2010;29:196-202. 32. Greenberg D, Earle C, Fang CH, Eldar-Lissai A, Neumann PJ. When is cancer care cost-effective? A
systematic overview of cost-utility analyses in oncology. J Nat Canc Inst. 2010;102:82-8. 33. Pearson SD, Bach PB. How Medicare could use comparative effectiveness research in deciding
on new coverage and reimbursement. Health Aff. 2010;29:1796-804.
51
Appendix A: Advisory Committee
Amy Abernethy, MD, Program Director, Cancer Care Research Program, Duke University Medical Center
Michael del Aguila, PhD, Senior Director, Biometrics, Health Outcomes and Payer Support, Genentech
Jeff Allen, PhD, Executive Director, Friends of Cancer Research
Naomi Aronson, PhD, Executive Director, Technology Evaluation Center, Blue Cross Blue Shield Association
Ethan M. Basch, MD, MSc, Oncologist, Outcomes Researcher, Memorial Sloan-Kettering Cancer Center, Chair,
ASCO Guidelines Committee
Roy Baynes, MD, VP, Oncology, Global Clinical Development, Amgen
Josh Benner, PharmD, ScD, Research Director, Brookings Institution, Engelberg Center for Health Care Reform
Timothy D. Birner, PharmD, MBA, Director, Global Evidence and Value Development, Medical Affairs, Global
R&D, Sanofi-Aventis
Amy Bonoff, MBA, Advocate, National Breast Cancer Coalition
Kim Caldwell, RPh, Director, Competitive Health Analytics, Humana
Stephen Eck, MD, PhD, Vice President, Translational Medicine and Pharmacogenomics, Eli Lilly
Erin Karnes, MPH, MBA, Research Associate, Brookings Institution, Engelberg Center for Health Care Reform
Muin Khoury, PhD, Director, National Office of Public Health Genomics, Centers for Disease Control and Prevention
William Li, MD, President and Medical Director, Angiogenesis Foundation
Robert McDonough, MD, Head of Clinical Policy Research & Development, Aetna
Sharon Murphy, MD, Scholar-in-Residence, Institute of Medicine
Joshua Ofman, MD, MSHS, VP, Global Coverage & Reimbursement, Global Government Affairs; VP, Global Health Economics and Global Development, Amgen
Andrew Perry, Senior Director, Oncology Payer Marketing, GlaxoSmithKline
BW Ruffner, MD, President, Tennessee Medical Association
Mary Lou Smith, JD, MBA, Co-Founder, Research Advocacy Network
Josephine Sollano, PhD, Head, Oncology Medical Communications and Global Health Economics and Outcomes Research at Pfizer Pharmaceuticals
Tamas Suto, Vice President and Head of Medical Affairs, Oncology, Sanofi-Aventis
52
Appendix B: Summit Attendees
Name Title Organization
Amy Abernethy Associate Professor, Duke University
Medical Center, Director Duke Cancer Care
Research Program
Duke University Medical Center
Gerry Adler Social Science Research Analyst Office of Strategic Planning, Centers for Medicare
and Medicaid Services
Raiz Ali Director Avalere Health
Jeff Allen Executive Director Friends of Cancer Research
Kirsten
Anderson
Chief of Staff, Office of the Chief Medical
Officer Aetna
Marietta
Anthony Director, Women's Health Critical Path Institute
Naomi Aronson Executive Director Blue Cross Blue Shield Association
Leonard Arzt Executive Director National Association for Proton Therapy
Tom Ault Principle Health Policy Alternatives
Dan Ball Research Advisor Eli Lilly and Company
Tracy Baroni
Allmon Executive Director, Health Policy Novarits Oncology
Charlie Barr Director of Patient Registries Genentech
Andrea
Baruchin Director, NIH Relations Foundation for NIH
Roy Baynes Vice President, Global Development Amgen
Robert Beltran Physician Executive Registry of Physician Executives
Marc Berger Vice President, Global Health Outcomes Eli Lilly
Nathan Blake Director of Legislative Affairs Jenkins Hill Consulting
Giselle Bleecker President Medical Market Strategists
Diane Bodurka Professor University of Texas M.D. Anderson Cancer Center
Amy Bonoff Advocate National Breast Cancer Coalition
Jonathan Bor Senior Editor Health Affairs
Pamela Bradley Associate Director of Science Policy American Association for Cancer Research
Otis Brawley Chief Medical Officer American Cancer Society
53
Name Title Organization
Steven
Brotman Senior Vice President Advanced Medical Technology Association
Paul Brown Research Associate Professor University of North Carolina
Susan Brown Director, Health Education Susan G. Komen for the Cure
Suanna Steeby
Bruinooge Director, Research Policy American Society of Clinical Oncology
Sarah Butler Senior Associate Avalere Health
Tanisha Carino Senior Vice President Avalere Health
Kenneth Carson Assistant Professor of Medicine Washington University School of Medicine
Steven Clauser Chief, Outcomes Research National Cancer Institute
Vivian Coates Vice President ECRI Institute
Perry Cohen Project Director Parkinson Pipeline Project
Deborah
Collyar President Patient Advocates In Research
Mitra Corral Associate Director Bristol-Myers Squibb
Martin Corry Director of Federal Health Policy Buchanan Ingersoll & Rooney PC
Eva Culakova Biostatistician III Duke University
George
Dahlman Senior Vice President, Public Policy The Leukemia & Lymphoma Society
Nancy
Davenport-
Ennis
Chief Executive Officer National Patient Advocate Foundation
Chris Dawe Health Staff Senate Finance Committee
Michael Del
Aguila Senior Director Genentech
Jessica
DeMartino Manager, Health Policy Programs National Comprehensive Cancer Center
Claude
Desjardins
Professor & Director of Clinical Scholars
Project University of Illinois Medical Center
Anthony Dias Managing Director Blue Cross Blue Shield Association