Basic Concepts in Device Development
David R. Holmes, MD
Anthony DeMaria, MD
Nabil Dib, MD, MSc
Spencer King III, MD
David G. Reuter, MD, PhD
Despite advances in medical care strategies including health care delivery systems, pharmaceuticals, and devices, cardiovascular disease remains the leading cause of death and disability worldwide. This is becoming an increasingly important clinical issue worldwide, in low- and middle-income countries with changing lifestyles, and also in resource rich countries due to the fact that patients are living longer and, in some areas, comorbid conditions such as diabetes and obesity are increasingly prevalent. To continue improving the outcome of therapy for cardiovascular disease will require multiple parallel steps and the integration of multi-stakeholder expertise. This document will focus specifically on the translational pathways for new devices but many of the key principles will also be germane for the pharmaceutical industry.
Translational science, research, and clinical medicine are integrally related. The opportunity to refine and revolutionize clinical care for patients is the inspiration that drives all innovation in medicine. The scientific starting point is the study of physiological processes and events typically beginning at the molecular level, then working up the chain to benchtop testing, in vitro modeling, in vivo modeling, and finally to testing and application in the human arena. Following this, there need to be widespread methods to validate the concepts, comparative effectiveness, and then implementation of the specific approach to broader patient care in an efficient, safe, and practical strategy. For this to occur, multiple parallel processes are required (Table 1) and multiple stakeholders must be involved (Table 2).
Table 1. Development of New Devices
- Recognition of unmet clinical needs
- Scientific study
- Pathophysiology of the “disease”
- Epidemiology of the “disease”
- Discovery of physiological mechanisms
- Innovation with development of new approaches/strategies/devices/relevant animal models
- Preclinical testing
- In vitro
- In vivo clinical models
- Early feasibility studies
- Pivotal trials in selected patient groups
- Regulatory approval
- Application in broader patient groups (registries)
- Comparative effectiveness research
- Professional society guidelines and appropriate use criteria
Table 2. Stakeholders in Device Development
- Clinical scientists
- Basic scientists
- Biomedical engineers
- Industry scientists
- Patient advocacy groups
- Regulatory agencies
- Reimbursement agencies
- Legal administration
- Hospital and physician groups
- Professional societies
These processes are lengthy, time consuming, expensive, and increasingly challenging. This relates to several factors: science has become more complex; product development exists in a world of increasing globalization of research and commercialization; and there is a progression of regulatory safeguards and legal issues that must be navigated. Other crucial issues relate to the politics of delivery of health care including the definition of what is health care, which population group is the recipient, who delivers it, and how is reimbursement organized. All of this exists in a changing global environment where there is intense competition for limited resources. The development of Health Technology Assessment groups adds a further layer of complexity particularly as this affects product innovation, development, and testing.
Political and social issues will play an increasingly important role by virtue of the emphasis on issues such as personalized medicine, patient responsibility for health maintenance, and expectations related to concepts such as quality of life versus emphasis on survival at all costs, and primary and secondary prevention to prevent the development and progression of disease requiring high cost therapy.
In light of the dramatic changes underway, a document for product development and iteration is necessary to fill current knowledge gaps and show a clear translational pathway that can expedite the science from discovery through refinement to clinical application. In short, given the existing rules, regulations, policies, and the exigencies of today’s changing health care industry, what is the best recommended course from discovery through Food and Drug Administration (FDA) approval, reimbursement, and guideline incorporation? This document was developed with the input of academia, clinical and basic scientists, industry experts, the FDA, and the Centers for Medicare and Medicaid Services (CMS), and organized by the International Society for Cardiovascular Translational Research (ISCTR). The ISCTR provides an environment for collaboration across these groups to expedite scientific discoveries into clinical application.
Preclinical and Clinical Trial Design and Endpoints
Strategies for trial design and endpoints must always start from a patient advocacy perspective. What data are required to make an informed decision regarding whether to pursue a given therapy? Scientific and clinical considerations include:
1) the specific statistical methods to be used: frequentist or Bayesian;
2) the specific safety and efficacy boundaries for device performance compared with the standard of care;
3) superiority versus noninferiority trial goals;
4) specific endpoints for both safety and efficacy; and
5) ground rules for continued surveillance by the data and safety monitoring board for the endpoints of device performance to facilitate notification and stopping rules during clinical trials.
These and other points are best addressed by very close and frequent collaboration between the proponents (industry and trial leadership) and the FDA.
The topic of performance criteria encompasses important implications and multiple challenges to be considered. Some of these boundaries may not have been defined because there is no predicate device and so they are arbitrary based upon what is felt to be reasonable; this may be particularly germane for Bayesian analyses with ‘priors.’ Additionally, the specific absolute or relative boundaries are important for assessment of performance. These boundaries should be set to identify statistical as well as clinical significance. An important consideration in these specific issues is in-depth discussion with the FDA regarding the appropriate statistical methods to be used.
The specific endpoints or metrics are of critical importance. Composite endpoints are commonly used to decrease the size of patient enrollment required but the hierarchical nature of these must be kept in mind. Relevant examples were seen in the trials of coronary artery bypass graft surgery (CABG) versus left main coronary artery (LMCA)/complex multivessel disease stenting. In the pivotal SYNTAX (Synergy between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery) trial, repeat revascularization was part of the composite endpoint, which also included death, myocardial infarction (MI), and stroke. In the subsequent EXCEL (Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) trial, repeat revascularization was not part of the composite endpoint and the trial conclusions were different. Another relevant example would be the endpoint of mortality; are investigators evaluating all-cause mortality or cardiac mortality? While both are important, devices may not affect noncardiac mortality, which might affect the trial results of new devices aimed at the treatment of cardiovascular diseases. However, the complexity of adjudicating cardiac mortality should not be underestimated.
Device Development Considerations
For device development, preclinical studies are essential and entail resolution of multiple issues. Appropriate animal models must be chosen based on the size of the animal, the vasculature in which the device will be tested, and the degree to which the animal’s anatomy best correlates with human anatomy. Some specific animal models might be selected for characteristics such as the ability to create cardiac lesions similar to the human anatomy.
Other characteristics include the length of the study and whether it is chronic or not. For example, if a porcine model is used and is intended for a 6-month endpoint, a mini swine may be chosen because of the practical nature of studying a fully developed adult pig. The design of these studies should be well thought out and developed in concert with the FDA so that important safety issues such as device thrombosis or device degradation with biopolymers can be assessed.
Finally, during animal testing, the device’s procedural performance must be able to be assessed, modified, and refined as needed. For first-of-a-kind technologies, preclinical studies offer the most efficient and cost-effective environment to iterate a novel therapy in preparation for human trials. Despite one’s best effort to select the ideal animal model, the heterogeneity of human disease provides the optimal environment to fully test a novel therapy. The Early Feasibility Study (EFS) guidance documents created by the FDA enable refinement of novel technologies in the clinical setting to minimize patient risk, enable continued iteration, and lay the foundation for the investment in a robust clinical trial designed to assess safety and effectiveness. For the final submission as part of the premarket approval (PMA), the device design needs to be fixed. (The guidance documents and more information about early feasibility studies are available from the FDA.)
Regulatory and Market Approval
There are three categories of devices for market and regulatory approval. Class 1 devices have the lowest risk and general controls alone are sufficient for establishing and maintaining safety and effectiveness. Classes 2 and 3 are intermediate and higher risk respectively and require incrementally more data to provide reasonable assurance for safety and effectiveness.
For Class 2 devices, proponents are trying to establish for the FDA substantial equivalence to a “predicate device,” one that has already been approved and legally marketed. Assessing the applicability and comparability with the “predicate device” is of central importance in this regard. This might apply, for instance, to a change from one generation to another generation of a specific device. For Class 3 high-risk devices, the bar is higher, including a PMA that includes both preclinical and clinical data to support safety and effectiveness. These PMAs are subject to quality system regulation inspection before approval. Additional components to PMAs include manufacturing quality evaluation, strategies for training, and post-approval studies, among other items.
Central to the field of developing technology, reimbursement is in the purview of CMS, and many issues need to be considered. Coverage of investigational devices (IDEs) includes categorization of that specific device into Category A (experimental/investigational) and B (nonexperimental/investigational); this is developed in conjunction with the FDA. Another important consideration is reimbursement not just to cover the device cost, but also the ‘routine costs of care’ related to the device. If ancillary tests are required for a specific device trial that are not part of routine care, reimbursement is not available; accordingly these costs must be born either by private insurance (unlikely), the patient (even more unlikely), or the sponsor. Resolution of these issues will be important to optimize the process of new technology development.
Guideline development and promulgation codify the role of new approaches. While increasingly central to the optimization of health care delivery, it must be remembered that guidelines are more fixed in time and sometimes lag behind actual practice patterns, which are fluid and depend on newly evolving data presentation and publication. Specific policies for the process have been implemented by professional societies based upon objective data and upon evaluation of the level of evidence used (for example, randomized clinical trials, registries, etc.). The practical issue for the guidelines – keeping up with the data in rapidly changing fields – is an important one, and processes for rapid iterations that include current updates must be implemented to provide “just-in-time” information needed for practice. For some insurance carriers, presence of a specific indication for the device in question is used to formulate coverage decisions. It is critical that guidelines become a living document to prevent inappropriate delay in the treatment of patients.
An important feature of this ISCTR document, which should influence guidelines, is that it is dynamic. As our understanding of disease processes evolve, new clinical trial endpoints may replace old parameters to provide clinicians and their patients the most relevant data for shared decision making. Indeed, the beauty of innovation is that new technologies enable improved disease management, which subsequently allows scientists to better understand fundamental mechanisms of disease. When the product life cycle is a catalyst for new scientific discovery, next-generation therapies should be tested with refined clinical trial endpoints that represent our current understanding of disease. The ISCTR document is designed to evolve as science matures so that its core principles and guidance are relevant for the longitudinal development of translational therapies.