6
Final Remarks
The primary goals of this workshop were to discuss the types of evidence needed by the various stakeholders involved with genomic diagnostics and to identify mechanisms to obtain high-quality evidence more efficiently. From the regulatory perspective, decisions by the FDA to clear or approve medical devices (including genomic diagnostic tests) for marketing are based on the safety and effectiveness of the product. From the payer perspective, demonstration of clinical utility or evidence of improved health outcomes is required for decisions to provide coverage. Evidence-based review groups also take contextual issues into account and look for overall net benefit (the balance of benefits or potential benefits versus harms or potential harms) when making a recommendation for or against the use of a genomic application. From a provider perspective, the focus is on value for the patient and on improving individual outcomes by identifying the most appropriate treatment for a person’s situation (e.g., increasing the chances of cure, survival, or palliation or decreasing exposure to toxicity from unnecessary or inappropriate therapy).
With these perspectives and those given by other stakeholders during the workshop in mind, participants reviewed a variety of approaches to evidence generation, such as clinical trials, retrospective analysis of archived specimens, coverage with evidence development, and chains of evidence and analytic frameworks, and offered ideas and strategies that could help move evidence generation for genomic diagnostic test development forward.
CHAIR’S SUMMARY
Workshop chair Debra Leonard concluded the workshop by highlighting key questions and topics for further discussion (perhaps in future workshops facilitated by the IOM roundtable) and potential action items identified from the discussions.
Evidence
-
Consider how to close the gap between FDA and payers’ evidence requirements. (The FDA, payers, evidence-based review groups, and providers all expressed a willingness to come together with IOM facilitation for further discussion.)
-
Conduct an analysis of the cost-effectiveness of an “analytic framework” process (constructing a chain of evidence from pieces that together provide adequate supporting data) versus conducting one or two high-quality, prospective randomized controlled clinical trials.
-
Define what constitutes “adequate evidence.” Perfect evidence is unattainable. What level of certainty will allow the transition of a genomic intervention into clinical practice?
-
Educate the public on the need for evidence to support clinical tests and clinical practice. Create a public demand for evidence and reduce the demand for tests simply because they are available or new.
Reimbursement and Coverage
-
Discuss new economic/reimbursement models that place value on tests that can help identify when a particular treatment will not be beneficial and thereby prevent unnecessary costly therapeutic interventions.
-
Discuss implementation of a system that does not pay for treatment if a prognostic or predictive genomic test is available and the results of that test do not support treatment in the patient.
Medical Practice
-
Consider whether safety and efficacy (as determined by FDA review and approval/clearance) is sufficient to support the clinical use of a new genomic test in the context of medical practice relative to an individual patient’s situation or whether large amounts of population data should be required.
-
Explore the disconnect between medical practice and evidence-based review recommendations: Why do physicians tend to ignore evidence-based review recommendations and how can uptake of recommendations be enhanced?
Clinically Focused Research
-
Foster a patient-centric research system to focus diagnostic test research on clinically important questions.
-
Develop a cooperative arena for identifying the top 10 clinically important questions and the resources or mechanisms to generate that evidence collaboratively. Consider convening a large annual meeting to analyze the available data that can be used to answer clinically significant questions.
-
Discuss with journal editors the importance of transparency in the reporting of diagnostic test validation studies toward establishing a strong, unbiased evidence base. Journals should publish only validation studies that meet study design quality criteria, and results should be published regardless of whether the outcome is positive or negative.
Access to Clinical Trial Specimens and Data
-
Establish a single index of annotated clinical trial specimens and closed health systems (e.g., Medco) that have the ability to conduct genomic test development projects. Explore whether the mechanism used by GAPPNet could be used to achieve this goal.
-
Consider adopting across all institutes of the NIH the central repository model of NIDDK, which requires that specimens from NIDDK-funded clinical studies be submitted to NIDDK’s sample and data repositories, thereby facilitating controlled storage conditions and resource sharing.
-
Continue to develop ClinicalTrials.gov to be more complete with regard to trial information posted on that site and, in particular, to facilitate the reporting of trials with negative outcomes.
-
Develop models of data sharing for genomic tests and test development.
Academia
-
Engage academic medical center leadership in discussions about how faculty contributions are valued; change the academic promotion and reward systems to more highly value and to better
-
reward clinically important research outcomes and collaborative efforts, rather than rewarding solely on the number of grants and publications.
-
Develop a link between academic research and development and the health-care teams at academic medical centers, with the shared goal of improving the health of patients. Identify the incentives for aligning these stakeholders in order to leverage specific clinical experiences and develop novel research initiatives.
-
Train clinical investigators in diagnostic test development and in study design options and optimization.