OR WAIT 15 SECS
On time and on budget are always necessary for a clinical trial, but medical devices offer nuances.
What do bandages, drug-eluting stents, dental floss, infusion pumps, and medicinal leeches have in common? To the vast majority of people, the answer to that would be an emphatic "absolutely nothing." But to many of the readers of
Applied Clinical Trials
, the answer is obvious—they're all medical devices of one kind or another.
Medical devices, an extremely heterogeneous group of health care products, can use any combination of mechanical, electronic or chemical biochemical action(s) to achieve their purpose.1 Although a medical device doesn't achieve its primary intended effect by pharmacological, immunological or metabolic means, it can be assisted by such means. Two examples of this include drug-eluting stents and bone cements containing antibiotics; both are classified as medical devices based on their primary intended effects: holding open narrowed arteries and attachment of a prosthesis. These are also two examples of combination products that include a drug component that assists in achieving the primary intended effect of the device by pharmacological means.
Medical devices and drugs can both be developed for the purpose of treating the same disease. For example, electroconvulsive therapy devices and serotonin-reuptake inhibitors have both been approved for the treatment of depression. However, the regulatory pathway to approval for each would be quite different. Not unexpectedly, these differences in regulatory requirements trickle down to the clinical studies that are conducted. It is vital to understand the unique aspects of medical device development before planning a clinical study of a new device.
The purpose of this article is to highlight the operational and regulatory differences between medical device development and drug development, identify the unique aspects of clinical studies involving medical devices, and highlight best practices for planning and tracking the progress of these studies.
Classification of risk
Unlike the situation for new drugs, it is not necessary to conduct clinical studies for every new medical device before it is approved and marketed. The first step in determining whether or not clinical studies will be necessary is to determine to which category of device your product belongs. Regulatory authorities (in the United States, the EU, and elsewhere) recognize different categories of medical devices based on design complexity, device use characteristics, potential for harm, and other factors. Although each country or region defines these categories in somewhat different ways, they each include an evaluation of risk. For example, the categories of medical devices recognized in the United States are as follows:
In the United States, the FDA has published the categorization of medical devices at www.fda.gov. Risk categories are similar in other countries, although some differences do exist. Importantly, in 2010 all medical devices marketed in the EU (regardless of classification) must have a clinical evaluation on file with the regulatory authorities. Without this clinical evaluation, which may include information from clinical studies, the device may not be marketed or sold.
In the United States, the next step in determining whether or not a clinical study is necessary is to determine whether or not the device is "substantially equivalent" to already on the market. Substantial equivalence means that a device has: the same intended use and the same technological characteristics as a marketed device, or the same intended use and different technological characteristics but is as safe and effective as the marketed device and does not raise different questions of safety and effectiveness. If a Class 2 or Class 3 product is not deemed substantially equivalent to another approved product, the submission of clinical data showing safety and effectiveness, although not necessarily efficacy, will be required. 2
For those steeped in drug terminology, medical devices that are "substantially equivalent" to an already approved device are almost never the device equivalent of a generic drug (which must be chemically identical to the innovator drug). Substantially equivalent devices often include fairly significant design changes, but as long as these technological differences don't raise questions about safety or effectiveness, the device will be considered substantially equivalent. In contrast, similar changes made to a generic drug would mean it was no longer a generic product, but a new chemical entity (NCE) subject to the regulations governing NCEs. The terminology varies, but there is a similar determination of equivalence (or not) to already marketed products in other countries as well.
Obtaining regulatory approval
A new drug cannot be marketed in the United States without an approved New Drug Application (NDA). In EU member states, either a Community Authorization or a National Authorization must be obtained. A full discussion of the process to obtain regulatory approval for a new medical device is outside the scope of this article, but as stated previously, due to the heterogeneity of the types of medical devices, there are multiple pathways to obtaining clearance or approval to market a device.
In the United States, the FDA Center for Devices and Radiological Health is responsible for medical device market clearance or approval. Most Class 2 medical devices require the submission of a 510(k) premarket notification to FDA. The FDA will review the information included in the notification, and if it's acceptable, will clear the product for marketing. Most of the information submitted in a 510(k) premarket notification is of a nonclinical nature, and currently only 10% to 15% of 510(k) submissions are submitted with clinical data to demonstrate substantial equivalence.3
High-risk medical devices or Class 3 and Class 2 medical devices that are not substantially equivalent to another approved product require premarket approval (PMA) prior to marketing. All PMA submissions contain clinical data, but the volume of data included is generally much smaller than that included in a similar application for a new drug. A recent review of the summaries of safety and effectiveness data for 78 PMAs for high-risk cardiovascular devices showed that 65% of the FDA approvals were based on a single study that included an average of 300 patients.4
Examples of U.S. Classifications
Although the regulatory approval process is quite different in the EU, the same high-level principle applies: Before a device can be marketed, the developers must demonstrate the new product is safe and effective. As in the United States, the type and amount of data required for review varies by type of device as well as risk; however, every medical device sold in the EU (regardless of classification) must have a clinical evaluation report in its technical file.
In the EU, unlike the United States, organizations called Notified Bodies are responsible for assessing conformity of medical devices with the relevant directives. Notified Bodies are not governmental agencies but are appointed by EU Member States to undertake specific tasks identified within EU directives. They will inspect and audit manufacturers to ensure that all requirements are met. In the United States, these activities are performed by the FDA.
Obtaining regulatory approval for a medical device can take anywhere from months for many substantially equivalent devices to 10 years (or more) for a high-risk, innovative device. Not surprisingly, it is difficult to determine a single average cost for development of a medical device.
Operational planning of studies
Accurately planning for any clinical study requires a thorough understanding of the operational work required and the level of effort that will be necessary. In many ways, planning and conducting a study of a medical device is very similar to planning and conducting a study of a drug.
Activity-based planning, which is sometimes referred to as a "project-management approach" or a "bottom-up approach," is an effective method for estimating the budgetary and headcount needs for medical device studies. This approach—a thorough discussion of which is outside the scope of this article—starts with determining the level of effort for a specific resource (person) to perform a given activity, which is typically calculated using an algorithm derived from experience and analysis of the cost and time drivers that affect each activity. After calculating the effort (in work hours), the cost for that specific activity can be derived by factoring in the hourly cost of the person(s) who perform that task. The compilation of costs for all tasks in the project is the total project cost.
Most of the activities performed for a medical device study are virtually identical to those performed for a drug study. For example, both types require a study protocol (which is also called an investigational plan), informed consent forms, case report forms, qualified study sites, appropriate site and vendor contracts, IRB review and approval, monitoring and drug/device accountability, collection and cleaning of data, and analysis and summarization of study results. Both require the input of statistics, data management, clinical operations, project management, and medical, quality, and regulatory staff to successfully complete.
However, as highlighted in the rest of this article, there are important differences between drug and medical device studies that must be taken into consideration.
In the United States, the conduct of medical device studies is regulated under the Investigational Device Exemption (IDE) regulations (21 CFR Part 812). For studies that meet the definition of "significant risk," which means it has the potential for serious risk to the health, safety or welfare of a subject, an IDE must be submitted to and approved by FDA before the study can begin. The FDA has 30 days in which to review and respond to an IDE application. Unlike an IND, an IDE application must be submitted for every applicable device study, not just the first study of a particular version of a device, and it must be formally approved prior to the study. The safety and effectiveness data collected during a study conducted under an IDE can be used to support a PMA or Premarket Notification 510(k) submission to FDA.
Clinical studies with devices of significant risk must be approved by both the FDA and an IRB before the study can begin, and many IRBs will require that FDA approval be obtained prior to their review. Studies with devices of nonsignificant risk don't require FDA approval but must be approved by the IRB before the study can begin.
In the EU, manufacturers must notify the relevant Competent Authorities prior to the conduct of a proposed study, who have 60 days to review and authorize studies of higher-risk devices. For lower-risk devices, the Competent Authority may immediately issue a letter of no objection as long as ethics committee approval has been obtained. A standard application form for notifying the Competent Authorities that a study is planned has not yet been developed, so the appropriate application form(s) must be obtained from each relevant Competent Authority.
There is no standard form equivalent to a Form FDA 1572 that an investigator participating in a medical device study must sign, either in the United States or the EU. However, for studies conducted in the United States, 21 CFR 812.43(c) outlines the elements that must be included in an investigator agreement that must be signed by each investigator prior to participating in the medical device study.
Payment for investigational devices
The vast majority of the time, sponsors do not charge the participants in a study, their physicians, or insurance companies for the cost of an investigational drug (Treatment INDs are an exception to this; sponsors are allowed to recoup the cost of the drug but are not allowed to make a profit). For most pharmaceutical companies, the cost of manufacturing the investigational product is usually not so high as to represent a barrier to conduct of a study.
For an investigational medical device, the situation can be completely different. Many medical devices being tested under an IDE are complex and expensive to manufacture, yet represent only incremental improvements over approved devices that have already been demonstrated to be safe and effective. There is very little reason to think that these improved devices are not safe and effective, and the expense of providing the new-generation device free-of-charge could be a significant barrier to the development and evaluation of incremental improvements.
To lower this barrier and ensure that their patients have access to improved products, the U.S. Centers for Medicare and Medicaid Services (CMS) will reimburse for the cost of investigational medical devices that meet certain criteria—although they'll only reimburse up to the cost of a currently marketed, similar product. These medical devices are "Category B Investigational Devices," and must be studied under an IDE. The investigator (with the sponsor's assistance) is responsible for submitting a Medicare Request for Reimbursement Letter to verify that the device is reimbursable before conducting the study. Although private insurers are not required to follow Medicare guidelines, they tend to institute coverage policies that are similar to Medicare.
This means that for some investigational medical devices, the sponsor company could charge the investigator for the device, and the investigator would then obtain reimbursement from Medicare or other insurer. The extra time and effort that will be required to obtain these reimbursement decisions must be built into the timeline and budget for the clinical study.
GCP and ISO
However, the current version of ISO 14155 does not address all of FDA's specific requirements for the conduct of clinical studies (e.g., the presence of written SOPs, minimum data retention period, access to source data, definitions of reportable adverse events). For studies that will be conducted under an IDE and used to support a 510(k) or PMA in the United States, it is important to ensure that all of the study will be considered acceptable to the FDA. It is important to make sure that FDA GCP requirements will be met.
ISO 14155 is currently in the process of being updated, and the Draft International Standard (DIS) is currently available. It's very likely that the next approved version will more closely mirror ICH GCP. To obtain a full copy of the currently approved ISO 14155, or ISO/DIS 14155, or any other of the approximately 18,000 ISO standards, visit the ISO Web site at www.iso.org.
Quick—what's ISO 14155, and what does it have to do with clinical trials? Many clinical research practitioners are surprised to learn that ICH guidance documents, including ICH GCP, were developed for medicinal products only, and do not formally apply to medical devices. If you've worked on medical device studies before, particularly in the EU, you're probably already aware that the GCP requirements in the EU Medical Device directives can be met by following ISO 14155: Standards for the Clinical Investigation of Medical Devices. This standard is one of many published by the International Standards Organization (ISO).
Adverse events reporting requirements for medical devices are different from those for drugs, and this can cause a fair amount of confusion for both sponsors and investigators.5 In a study of a drug, all adverse events, regardless of cause, are reportable. There are good reasons for this: Drugs are distributed systemically in most situations, may or may not be metabolized or affect the metabolism or availability of other drugs, and may or may not affect multiple organs. Many of the side effects of drugs occur at a background level in the general population, and it is often necessary to analyze data from large groups of patients to determine whether or not an adverse event is related to a drug.
Adverse Events for Medical Devices
Medical devices, on the other hand, usually act locally and have a transparent method of action, and it is often easy to determine if an adverse event is caused by the device (e.g., a faulty device that delivers an electric shock to the patient). Most medical device side effects don't occur at a baseline rate in the population, so statistical analyses involving large groups of treated and control patients are often not necessary to attribute a particular adverse event to a device.
In the United States, the FDA requires that investigators prepare and submit a report of any Unexpected Adverse Device Effect (UADE) occurring during a medical device study to the sponsor and their IRB within 10 working days of learning of the event. The sponsor is then required to conduct an immediate evaluation of the UADE and submit a report to the FDA, all reviewing IRBs, and all participating investigators within 10 working days after they were notified that the UADE had occurred. Although the acronym UADE used by the FDA to define what reports have to be expedited doesn't include a reference to the seriousness of the effect, the FDA defines a UADE as a serious adverse effect on health or safety (21 CFR 812.3).
The relationships of each category of medical device adverse events are shown in Figure 1. Depending on the type of medical device, it might be advisable to include additional requirements for adverse event reporting in your protocol. Of interest, the current draft of the ISO 14155 standard that is circulating for comment includes a requirement that all adverse events be recorded on CRFs (regardless of their relationship to the device).
It is important to ensure that your investigators understand how to categorize adverse events so that they can record and report them appropriately. It can be useful to provide a tool that will assist them with this—one example seen online recently is an "Adverse Event Wheel," which the user spins until all of the attributes of a particular event are visible on the wheel and the event categorization is provided. The protocol or investigational plan for the study must identify reporting time frames, as well.
In the EU, adverse event reporting requirements are defined on a by-country basis, so the sponsor will need to ensure that they understand the requirements and have implemented appropriate methods to meet them.
The CRFs designed for the study must enable the investigator to record the appropriate information about an adverse event. Medical device adverse event reports will usually include information about what happened to the device as well as what happened to the subject.
Finally, it is important to determine what tool will be used for coding of adverse events that occur during a medical device study. Unfortunately, although the MedDRA coding system is used by most sponsors to code adverse events in drug studies, no such standardization has been accomplished in the device world.6
Training of site staff
It is virtually never necessary to involve an engineer in the training of a site participating in a drug study, and in most cases, the training involved does not involve hands-on preparation or administration of the drug itself. However, in studies of complex medical devices, it may be necessary to spend a good portion of a day training the site staff on the appropriate use of the device (e.g., an updated infusion pump). In some cases, the CRA, who may also be a nurse, will be able to provide the training without assistance; however, in others it is necessary to have someone from the engineering team conduct portions or all of the training.
In addition to potentially requiring more involved training, it is necessary to evaluate whether or not an engineer or other product expert should be available at the site while the medical device is being used during the study.
Tracking project progress is a vital component of overall study management. Many study sponsors today still use budget tracking without evaluating the "earned value" of the work performed to assess project status over time. Unfortunately, simply tracking the budget and a set of major milestones that are achieved can be very misleading. A powerful technique for tracking the status of a clinical study is called Earned Value Management (EVM), and it's a technique that has been used for decades in other industries like construction and defense.
EVM requires an activity-based plan to track progress against, as well as detailed reports from your vendors describing the work performed to date. EVM will highlight both cost and schedule variances at the same time, and will provide early warnings about exactly where and why a project is moving off track. For example, the actual cost of a study could be exactly what had been planned at year-end, however, the study may be several months behind schedule (indicating that unidentified scope changes may have occurred). Likewise, the actual cost of a study could be less than planned, which could imply that the study is behind schedule. Through the application of EVM techniques will you be able to determine if the study is in fact delayed, or if there have been unplanned efficiencies that have resulted in cost savings. Applying EVM to any clinical study provides a level of visibility to ensure you stay on track.
Although clinical studies of medical devices share many similarities with those of new drugs, there are important differences that must be taken into account, starting with the regulatory requirements that depend heavily on the type and level of innovation of the particular device.
The project team must work together closely up front to outline the scope of work of the project. Once all of the required tasks have been identified, the project manager can utilize activity-based planning to accurately estimate the budgetary and head count needs for the study. Study progress can be tracked accurately using earned value management, which provides the ability to hone in on the reasons for delays or budget overruns and identify the best approach to resolve the situation. These techniques can be applied regardless of the type of study you're conducting—and one thing common to all study teams is the need to complete your study on time and within budget.
Molly Blake-Michaels, MS, is Director, Clinical Services, for ClearTrial, Westmont, IL, email: [email protected].
1. J. Tobin and G. Walsh, Medical Product Regulatory Affairs (Wiley-VCH, Weinheim, Germany, 2008).
2. I. Sim, "Trial Registration for Public Trust: Making the Case for Medical Devices," Journal of General Internal Medicine, 23 (suppl 1) 64-68 (2007).
3. A. Yustein, "The FDA's Process of Regulatory Premarket Review for New Medical Devices," in C.W. Howden (Ed.), Gastroenterology and Hepatology Annual Review (AGA Institute Press, 2006) pp. 142-144.
4. S. Dhruva, L. Bero, R. Redberg, "Strength of Study Evidence Examined by the FDA in Premarket Approval of Cardiovascular Devices," JAMA, 302 (24) 2679-85.
5. A. Gertel and N. Stark, "Serving Two Masters: The World of Medical Devices," The Write Stuff, 17 (1) 74-77 (2009).
6. M. Field and H. Tilson (Eds.), Safe Medical Devices for Children (The National Academies Press , Washington, DC, 2005).