MD+DI Online is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Articles from 2004 In August

Sterilization Methods Stand the Test of Time

Karl J. Hemmerich is the manager of plant operations for STERIS Isomedix (Sandy, UT), and John Masefield is executive advisor to STERIS. Jerry R. Nelson, PhD, is director of Nelson Laboratories Inc. (Salt Lake City). 

Over the past 25 years, the medical device industry has changed significantly. Sterilization has evolved as well, sometimes reacting to the industry and sometimes leading the way in new methods, standards, and processes. In the eyes of someone looking forward in time from 1979, many medical devices today are very sophisticated in nature and design while others hardly changed at all. Sterilization technologies, having remained unchanged in the physics and chemistry of microbial inactivation, have undergone a different sort of metamorphosis than the medical products they support.

In 1980, the primary methods for medical device sterilization were ethylene oxide (EtO), gamma, electron beam (E-beam), moist heat, dry heat, liquid-chemical germicide, and other gaseous methods. More than 20 years later, the primary sterilization modalities remain the same. So, has nothing changed?

Each sterilization method kills organisms effectively, and usually one method is more effective or more efficient for a given product than another. However, changes in the methods themselves, the standards, and the development of biological indicators (BIs) and parametric release have led to modifications in the processes and shifts in industry preferences. This article explores those changes and their effects over the past 25 years on the medical device sterilization market.

Sterilization of Medical Devices

In 1980, EtO was used for nearly 90% of all sterilized devices. At that time, only the largest manufacturers used either gamma or E-beam sterilization. Irradiation contractors were just emerging with a limited number of facilities. In fact, the two largest contract irradiators, STERIS Isomedix and Sterigenics, were only founded in the 1970s. Figure 1 shows the usage breakdown for EtO, gamma, and E-beam, as well as for in-house and contract sterilization, from 1980 to 2003. 

One aspect of the medical device industry that is hidden in the data is that an ever-increasing portion of U.S. and Canadian medical device manufacturing is moving overseas. However, many of those foreign-produced products return to North America for terminal sterilization.
The reason for this pattern relates to both the corporate structures and the difference in the capital investment required for manufacturing operations compared with sterilization. Manufacturing often requires minimal costs for both labor and capital equipment, whereas sterilization facilities require substantial capital for equipment. 

The Role of Standards. The development of standards and guidelines has been critical to advances in sterilization. Both the Association for the Advancement of Medical Instrumentation (AAMI) and the International Organization for Standardization (ISO) have developed viable validation guidelines that are relatively easy to execute.

These standards often came in response to FDA and international regulatory requirements related to good manufacturing practices and quality assurance. The Medical Device Amendments of 1976 were followed by FDA's “Reproduction Quality Assurance Planning” guidance documents in 1981 and 1988 and its “Guideline on General Principles of Process Validation” in 1983 and 1987, which changed the way sterility was addressed and validated. These were augmented by the 1990 amendments.

Today, the AAMI and ISO guidelines enable companies to implement a sterilization program without the need for a large and burdensome sterilization technologies department. These voluntary consensus standards have also provided regulatory bodies such as FDA a reference on which to base their determination of product sterility. Table I provides a listing of current AAMI and ISO sterilization standards.

Contract Sterilization. In response to market demands, contract sterilization services have increased significantly. The growth of sterilization contract services has been a result of the drive for medical device manufacturers to decrease costs.

The availability of contract services has played a major role in the development of sophisticated medical device products. Without having to invest in the equipment or labor required, companies have access to the state-of-the-art sterilization facilities that are necessary for such products. 

The convergence of these events fueled the progress of sterilization over the past 25 years. Improving turnaround time is often a requirement for survival. Methods, systems, and guidelines have been developed to fulfill the just-in-time manufacturing methods in widespread use today. Parametric release and instantaneous BIs are just two of the breakthrough concepts being developed to reduce turnaround time and improve the reliability of sterilization for manufacturers.

EtO Sterilization

EtO sterilization has been critical to the sterilization of disposable medical devices since its discovery by Phillips and Kaye in 1944. They stumbled upon it while conducting work on biological decontamination at Fort Detrick, MD, for the U.S. Army. In the 1980s, EtO lost some market share to irradiation sterilization as a result of environmental laws (e.g., Montréal Protocol [1987], California Prop. 65, California Air Quality Management District, and other EPA regulations) that were designed to curtail the destruction of the earth's ozone layer by fluorinated gases.

At the time, many thought that EtO would be phased out; however, the decreased use proved to be just a temporary setback as companies implemented the required changes to switch from 12/88-EtO/chlorofluorocarbon-gas mixture to 100% EtO gas. Since then, EtO sterilization has grown steadily as it benefited from the ANSI/AAMI/ISO Guidelines ST27/11135-1994, existing capital investments, minimal retrofit costs, and contractor availability.

Major advances in EtO sterilization include improved BIs, parametric release, lower gas concentrations, and improved outgassing methods. BIs have evolved from earth- or sand-borne bacteria in EtO's early years to protein-based alternatives with ever-faster response times. In fact, the parametric release process has decreased from 10–14 days, which was common for BI release in the late 1970s, to the current 2 days. Even the dread of dealing with EtO residuals has abated through the use of nitrogen flushes, cycled heat washes, and other processing methods. 

Parametric release, which was implemented in a few in-house 12/88 EtO sterilization facilities 20 years ago, is now in place in many in-house and contract facilities. The advent of intrinsically safe infrared spectrophotometers, as well as computer-controlled systems, has enabled parametric release in 100% EtO processes. Enhanced BI incubation time reduction studies, lower residuals, and parametric release have ultimately led to safer products and shorter product holding times.

Gamma Sterilization

Table I. An overview of global sterilization and related standards (click to enlarge).

Gamma sterilization was a post-WWII technology. During the early 1960s, it was used initially to eradicate anthrax on wool for carpet 
manufacturing in Australia. Gamma was not used to sterilize medical products in North America, however, until 1964. The Atomic Energy of Canada (later MDS Nordion; Ottawa, ON) commissioned the first industrial cobalt-60 sterilization facility to sterilize surgical sutures for the Ethicon division of Johnson & Johnson.

In 1972, Isomedix (later STERIS Isomedix) opened its first contract irradiation facility in New Jersey. Sterigenics opened its first gamma facility in 1979. A few large medical device manufacturers invested in this new technology around the same time.

The development of radiation-stable and enhanced-performance polymers was key to the success of gamma sterilization. Today, PVC and polycarbonate formulations eliminate the brown-yellow color that developed with predecessor products.

Radiation-stable polypropylene has been developed to prevent the brittleness that previously limited the polymer's use in products that would be radiation sterilized. AAMI Technical Information Report 17 was published in 1997 to define materials stability, processing effects, and proper test qualification methods (including accelerated aging). 

Validation methods have also improved. Guidelines have progressed from “just use 2.5 Mrd” to the Kilmer method, to ANSI/AAMI ST32 (1980/ 1990), to AAMI/ISO 11137 (1994), and AAMI/ISO 13409 (1996), and finally, to the current TIR 27 VDmax. In 1974, under the auspices of AAMI, a committee was formed to develop rational process-control guidelines for the sterilization of medical devices.

The validation guidelines that were developed included the development of rational dose-setting methods that took into account actual product bioburden, its distribution on the product, its radiation resistance, and, to a limited extent, the end-use of the product. By 1994, after much work and a series of national and international meetings, the AAMI dose-setting methods received worldwide recognition when included, with little modification, in ISO 11137. 

Use of these rational dose-setting methods frequently resulted in sterilizing doses much lower than the traditional 2.5 Mrd, which further expanded the range of conventional medical-grade plastics that could be sterilized without generating unacceptable color, hardness, or brittleness. It is important to note that the unit for irradiation dosage has changed from megarads (Mrd) to kiloGrays (kGy) in order to comply with System International conventions.

By the mid-1990s, a growing number of contract gamma irradiation facilities were making sterilization services available to medical product manufacturers, making radiation sterilization more cost-competitive. 

A national network of large-scale contract irradiation facilities has meant that healthcare companies no longer need to build in-house irradiators. Many device companies have decommissioned their in-house irradiation facilities in favor of using contract irradiation facilities.

Mathematical Modeling. Manufacturers are increasingly interested in the mathematical modeling of dose distribution in irradiated products. Such modeling may help improve process design, evaluation, and control. Manufacturers hope to better understand the distribution of the absorbed dose within products of heterogeneous density when the products are exposed to ionizing radiation in an industrial irradiator. They also hope to minimize the number of physical dosimeters required.

The use of dosimetry to measure the actual dose applied during processing has also evolved. The use of extensive numbers of thin films and dyed-plastic dosimeters has shifted to the less-frequent use of more-accurate reference alanine dosimeters in production irradiation runs. Predictive mathematical models have been developed for gamma, high-energy E-beam, and x-ray irradiation facilities. As determined by the mathematical modeling, the dosimeters are placed in zones of minimum and maximum dose in the products being irradiated. Other dosimetry developments include Internet-based dosimeter calibration and real-time monitoring of absorbed dose.

E-Beam and X-Ray Sterilization

Figure 1. The bars for each year show the market share for EtO, gamma, and E-beam/
x-ray sterilization methods, for representative years from 1980 to 2003. The overlaid lines show the industry's use of in-house and contract sterilization facilities during the same period (click to enlarge).

The biggest single change in the E-beam market has been the improved reliability of the delivery systems. IBA Inc. revolutionized the E-beam market with its Rhodotron accelerator, while Titan Scan Systems improved the reliability of linear accelerator design. Besides the accelerators themselves, materials-handling systems made a quantum leap to better present the product to the beam consistently and accurately. 

X-ray has only recently been used for commercial sterilization. However, it was actually the first radiation technology for sterilizing medical products in the late 1800s. E-beam found its first medical niche in 1956 as a sterilization method when Johnson & Johnson used it for sutures.
Recent developments have started looking at x-ray for use with 5–10 million electron-volt (MeV) high-power electron accelerators, with the goal of enabling the sterilization of higher-density medical products and the terminal sterilization of heat-sensitive, low-density medical products.

Progress is being made in the development of modular, sealed E-beam tubes, working at voltages around 100 kV. The tubes show promise for low-power use, particularly in the surface sterilization of medical devices and pharmaceutical packaging materials used in conjunction with aseptic filling operations.

Vaporized Hydrogen Peroxide 

Vaporized hydrogen peroxide (VHP) sterilization is increasingly being used for the industrial sterilization of medical devices. As a low-temperature gaseous method of sterilization, it is an alternative to EtO for some products. It does not penetrate as much as ethylene oxide does; however, it does offer some advantages, including excellent material compatibility and short cycle times. And, because peroxide breaks down into water and oxygen, there is virtually no aeration time. VHP is also sporicidal at very low concentrations (typically 1–2 mg/L at 25°C).

Like all sterilization modalities, VHP has limitations. Because it lacks the penetration available with EtO, an open gas pathway must be present, and certain materials (e.g., cellulose) cannot be processed. 

Validation for VHP is similar to that for EtO. Because VHP sterilization has no consensus standard, the concepts that are outlined in ANSI/AAMI/ISO 14937 for general requirements of a sterilizing agent can be used to develop an appropriate validation strategy. Other published EtO standards can be used as well. Several companies have recently obtained 510(k) clearance to use VHP to terminally sterilize their medical devices.

Steam Sterilization

Medical devices manufactured from materials that are thermally tolerant can be processed with steam. The use of steam is increasing due to its inherent low processing and capital costs, direct in-house control, and parametric release; however, it remains a small percentage of the overall market.

Some sterilization contractors now provide steam sterilization services. In addition, a variety of packaging materials are available that can be 

Dry-Heat Sterilization

AAMI recently released a new standard for validation of dry-heat sterilization. Dry heat has classically been considered as 170°–180°C for an hour; however, longer times at considerably lower temperatures also can be validated. An increasing number of medical devices are being processed with dry heat (160C), although its market share is not easily determined because most dry-heat processing is not outsourced. Packaging materials that can withstand the rigors of dry heat are also now available. 

The Sterilization Future

Over the past 25 years, sterilization has evolved to keep pace with the medical device industry. Although sterilization technologies themselves have remained unchanged in their physics and chemistry, some significant developments have advanced the methods and processes. The future promises to offer even more substantial developments. Key developments in the sterilization market as a whole should include:

• Changes in the sterility assurance level (SAL) requirements for various device groups. Internationally, the requirement for a device to be labeled with a “sterile” statement requires a SAL of 10–6. In the future, this requirement may be revisited to allow products with alternate SALs (i.e., 10–3, 10–4, etc.), based on the need for such elevated SALs and on product use. 
• Alternative gaseous-sterilization modalities, such as vapor-phase hydrogen peroxide and plasma vapor-phase hydrogen peroxide, offered on a contract basis.
• An overseas exodus of sterilization similar to that of device manufacturing.
• Modalities that are less energy dependent (i.e., more cost-effective) to combat spiraling energy costs.
• Improved manufacturing controls that reduce bioburden spike issues.
• Documents from the standards-writing groups covering how to deal with bioburden spikes.
• Rapid environmental monitoring techniques for manufacturing. 
• Further harmonization of international methods. 

Major advances affecting EtO sterilization and processing include:

• BIs that offer essentially instantaneous response times.
• Further developments in parametric release for EtO sterilization.
• Evolution of the industry guidelines for validation. 

Developments in gamma sterilization will likely include:

• Final development and implementation of VDmax guidelines at doses other than 25 kGy with validation using this technique from 15 to 35 kGy.
• Evolution of the industry guidelines for validation. In 2004, the next revision of the AAMI 11137 standards should be released: AAMI 11137-1 draft standard, “Process Requirements for Radiation Sterilization;” AAMI 11137-2 draft standard, “Validation Methods for Radiation Sterilization;” and AAMI 11137-3 draft standard, “Dosimetry for Radiation Sterilization.”
• No sterilization dose validation requirement for devices with proven very low bioburdens.
• Predictive dosimetry using mathematical modeling that is based on the application of knowledge and computer power currently available.
• Parametric release for irradiation processing. Cost-competitive pressures and improved science (based on enhanced computer control and real-time dose-measurement systems) may drive this feasible methodology forward.
• Alternative dose-setting strategies, further increasing the options available to the sterilization microbiologist.

Laboratory advancements should include:

• Rapid microbiological methods that reduce the time needed for bioburden assessment from about a week to a day. The technologies already exist, but they will come into common use as regulatory acceptance, awareness, and their utility become better known.
• Genetic-based microbial identification methods as the standard for identification of environmental isolates, bioburden tracking, and failure investigations. 
• Technology improvements such as isolators, cleanroom HEPA-filtered respirators and improved barrier gowns and gown materials that increase the reliability of existing sterility test methods.
• New rapid microbiological methods that replace the existing sterility test procedure.

With these developments, sterilization will continue to be an integral part of the medical device industry. As it has in the last 25 years, we are sure that MD&DI will bring you updates and developments as they happen in the next 25 years.

Copyright ©2004 Medical Device & Diagnostic Industry

Device Regulation: Policies, Practices, and Procedures

Originally Published MDDI August 2004

Regulatory Affairs

The Medical Device Amendments of 1976 changed the landscape of FDA regulation of medical devices. With these changes came new policies and procedures that have transformed the industry.

Jonathan S. Kahan

Jonathan S. Kahan is a partner at Hogan & Hartson LLP and a member of the MD&DI editorial advisory board. 

The editors of MD&DI have asked me to provide some thoughts regarding FDA regulation of medical devices over the last 28 years. Much has happened in terms of device regulation over this almost three-decade period. Since the passage of the Medical Device Amendments of 1976, not only has FDA's regulation of the medical device industry drastically changed—the industry itself has been transformed.

When I started working with the medical device industry in 1974, sales of medical devices by U.S. companies totaled $5 billion. By 1995, sales had grown to $52 billion. Today, sales by U.S. medical technology companies are close to $94 billion, and growing at an annual rate of 6%.

The advances in medical technology have been equally impressive. The industry began as one that manufactured commodity products such as surgical gloves and basic in vitro diagnostics (IVDs). By contrast, today's industry produces drug-eluting stents, artificial hearts, automatic implantable cardioverter-defibrillators, advanced magnetic resonance imaging machines, DNA-based IVDs, and even clinical laboratories on a chip. 

The Changing Regulatory Environment

My role, and the role of my colleagues who are regulatory counsel to the medical device industry, has changed almost as much as the industry and its technology. In 1976, it was not uncommon to file a 510(k) premarket notification with FDA that would be a mere five pages long. It would simply describe the “substantial equivalence” of the device to one that was on the market earlier that year (i.e., before May 28, 1976). Today, we rarely work on a submission that is not supported by significant amounts of data. Biocompatibility data, electromagnetic compatibility data, detailed information on the device software, and clinical data are often included. Premarket approval (PMA) applications for “not substantially equivalent” Class III devices have also become much more challenging. It is not unusual for our work on a premarket approval project to span several years.

Not only has the medical device premarket clearance and approval process changed significantly over the years, but the FDA enforcement laws, policies, and procedures have also evolved. Today's enforcement environment and FDA's tools for enforcement are somewhat different than those the industry faced even 10 years ago.

The Evolution of the Premarket Clearance and Approval Process

Before 1976, the federal government did not require the premarket clearance or approval of medical devices. In 1970, a government panel called the Cooper Committee was asked to analyze whether the premarket clearance of medical devices was appropriate and necessary. That committee reviewed the literature and reports on medical technology and determined that medical devices had caused thousands of injuries over the preceding decade. Hundreds of injuries, and some deaths, were linked to such devices as defective heart valves, faulty pacemakers, and substandard intrauterine devices.

Congress, under the leadership of Congressman Paul Rogers, determined that the imposition of FDA device premarket clearance was necessary to protect the public health. It was also designed to provide for a regulated environment in which the industry and its technologies could grow and prosper. While the resulting processes have certainly not been perfect, the underlying wisdom of the Medical Device Amendments of 1976 and their regulatory structure have withstood the test of time.

On November 28, 1990, President George H. W. Bush signed new legislation titled the Safe Medical Devices Act of 1990 (SMDA). SMDA constituted the first major change to the Medical Device Amendments of 1976. However, the new law essentially maintained the premarket clearance and approval structure of the original law. Many in Congress at the time felt that granting premarket clearance based on substantial equivalence to devices on the market prior to May 28, 1976, was a flawed paradigm. Regardless, Congress adopted FDA's existing substantial equivalence policies in the new law.

Specifically, for the purpose of determining substantial equivalence, SMDA stated that FDA must determine that the new device has the same intended use as the predicate device. FDA, by order, must find that the device has either the same technological characteristics or has different technological characteristics. If the characteristics are different, the information submitted must demonstrate that the device is substantially equivalent to the predicate device. This information can include clinical data that demonstrate that it is as safe and effective as the legally marketed device and does not raise different questions about safety and efficacy. This is the same paradigm that FDA operates under today; this approach was not changed by the FDA Modernization Act of 1997 (FDAMA). 

As the 510(k) process was becoming more data driven, the PMA process also changed, becoming more extenuated and more difficult over the years. For example, in 1987, the average time for PMA reviews was approximately 337 days. In 1991, PMA times had risen to 633 days, and by 1994, times had grown to a staggering 840 days. Many in the device industry called the early 1990s the dark ages of FDA medical device regulation. Both the 510(k) and PMA processes were essentially paralyzed during the early days of the administration of Commissioner David Kessler, MD. Kessler's “zero risk” regulatory philosophy was quickly understood and adopted by many within CDRH. For example, Kessler believed that there were many devices on the market whose risks outweighed their benefits. This philosophy had a significant effect on the market clearance of new devices and led to the withdrawal of some devices, including silicone gel breast implants.

Kessler's concerns also led to the creation of the infamous Committee for Clinical Review. Leading the committee was Robert J. Temple, MD, a drug evaluation official from FDA's Center for Drug Evaluation and Research (CDER). The Temple Committee reviewed device clearances and approvals by CDRH and concluded that there were significant deficiencies in the policies and procedures for clinical trials. The committee recommended that CDRH require, whenever possible, randomized controlled trials (RCTs) for all Class III device approvals. The RCT, at the time, was the controlling CDER paradigm for all drug studies.

The resulting CDRH preference for RCTs and the clinical trial requirements adopted in the early 1990s are still the controlling policy today. The imposition of the RCT requirement has led to much consternation in the device industry. One reason for such consternation is that an RCT is often unethical or impractical for many medical device clinical studies. The industry continues to struggle with questions of appropriate clinical trial design for novel new medical technologies. While the agency has been directed to impose only the “least burdensome” requirements for device approval, that direction has barely dented the continuing CDRH preference for RCTs.

Medical Device Reengineering and FDAMA

Reengineering. Following the dark ages of the early 1990s, CDRH management recognized that many of its regulatory procedures needed improvement. Industry had made its dissatisfaction known to both FDA and Congress. It became clear that if CDRH did not reform itself, Congress might reform the center through legislation. This well-justified fear of congressionally mandated change led FDA to examine many of its processes. Among those examined was the 510(k) process, to determine whether there were more-efficient methods for clearing substantially equivalent devices to market. To this end, CDRH drafted what became known as the 510(k) Paradigm, which put in place the new special 510(k) submission as well as the abbreviated 510(k) option. 

The special 510(k) submission essentially allowed limited data to be submitted to the agency. It covered new devices that, though modified in a way that could significantly affect safety and efficacy, changed neither the intended use nor basic technology of the predicate device. CDRH determined that a summary of the company's design control activities related to such changes, plus a comparison with the cleared predicate device, might be adequate for clearance. With this process, device sponsors are not required to include the typical submission of detailed supporting data. The special 510(k) submission has become a significant factor in medical device clearance over the last few years. Indeed, the majority of special 510(k) submissions are cleared in less than 30 days, as opposed to the more-extended clearances of traditional 510(k)s. This idea not only made sense on paper, but when put into practice, it became a successful program for expediting clearances.

The same cannot be said for the abbreviated 510(k). Under this process, a declaration of conformity to individual consensus standards could theoretically be the primary basis for a device's clearance. The abbreviated 510(k) process, unlike the special 510(k) process, has not been a successful program. But it has not entirely been a failure, either. To this day, FDA continues to evaluate how to utilize the abbreviated 510(k) process more effectively.

FDA's reengineering efforts in the mid-1990s also led to a reevaluation of the product development protocol (PDP) process, originally a part of the Medical Device Amendments of 1976. Without going into the mind-numbing details of the PDP law, CDRH's reengineering teams believed that device sponsors could use the PDP process to get their Class III devices approved more quickly. The agency consequently reissued and updated its PDP guidance, hoping that the process would become another important route for approval of novel new devices. 

Once again, FDA's hopes for the PDP process were dashed. The industry does utilize the PDP path for a few products to a limited extent. However, the PDP has never become a significant alternative route to market for innovative technologies. 

Finally, the reengineering teams within CDRH also looked at ways to streamline the PMA supplement process. They initiated new procedures for Real-Time PMA Supplements and 120-day PMA supplement reviews for certain devices. These changes were a primary focus of FDAMA.

FDAMA. The foregoing CDRH reengineering efforts were not successful in heading off new legislation. On November 21, 1997, President Clinton signed the FDA Modernization Act of 1997 into law. Many of the reengineering efforts discussed above were codified in FDAMA. For example, Section 205 of the act included a provision for reducing the burden of PMA supplements. In addition, the 30-day/135-day PMA supplement paradigm now in place for certain modifications in manufacturing procedures or methods were enacted into law.

Similarly, Section 201 of FDAMA provided streamlined procedures for making certain changes to investigational devices. In addition, it made these devices more readily available when needed for humanitarian uses. Once again, Congress decided not to drastically change the substantial equivalence paradigm adopted for 510(k) clearances. The lawmakers did, however, add an important new provision allowing the “downclassification” of certain low-risk novel technologies for which no substantially equivalent product is available. 

Under what became known as the de novo downclassification process, a manufacturer can request de novo downclassification within 30 days of receiving a “not substantially equivalent” determination. In other words, the manufacturer can seek to have the device reclassified down to Class I or II. The potential benefits of this provision have not yet been entirely recognized, but the de novo process has become an important consideration for companies developing devices that otherwise would require PMA clearance.

Finally, Section 210 of FDAMA directed CDRH to consider the “least burdensome” means of evaluating device effectiveness or demonstrating substantial equivalence. Much ink has been spilled concerning this provision. In my experience, however, the congressional admonishment has had little or no effect on CDRH's practices in actually approving or clearing devices to market. The RCT, clearly not the least-burdensome method for evaluating a device in most cases, remains FDA's default clinical trial policy.

Medical Device User Fees

The Medical Device User Fee and Modernization Act of 2002 (MDUFMA) was enacted into law on October 26, 2002. For many years, FDA had argued that one way to clear up its review backlog was the institution of user fees. The medical device industry had resisted such fees for many years. Unlike the drug industry, which was made up of much larger firms and had been paying user fees for some time, the device industry consisted primarily of small companies. These companies could not afford significant payments to FDA to carry out governmental tasks that they already supported through the payment of taxes. Many said FDA did not need user fees and should instead internally streamline its processes for greater effect.

The “small company” segment of the medical device industry continued to resist the imposition of user fees up to the bitter end. These companies argued that the payments would have a disproportionate effect on innovation by smaller companies and even jeopardize their survival. Nevertheless, AdvaMed (Washington, DC) and many large device companies ultimately gave in to legislative and executive branch pressures to accept user fees. These companies hoped that the fees would speed new products to market, a valid but entirely elusive expectation.

I think it is safe to say that, generally, the user fee program has been a significant disappointment for the industry. FDA has held stakeholder meetings seeking to publicize the improvements resulting from user fees. For example, FDA has stated that during FY 2004, the agency hopes to spend user fees to add more than 100 staff members to CDRH. The agency has said that the fees will allow CDRH to meet the law's performance targets that become applicable in 2005. 

It remains to be seen whether either the performance targets or the increase in staff will truly materialize. To date, FDA has touted its hiring of 50 new scientific, medical, engineering, and other review staff, including 14 new statisticians, with user fee monies. However, Congress has concurrently underfunded CDRH. Most importantly, no drastic change in the times of clearances or approvals has been seen, to the dismay of industry. Has the device industry been bamboozled by congressional and FDA promises? This is a question that can only be answered by watching CDRH activities over the next few years.

FDA Medical Device Enforcement Policies and Procedures

The Federal Food, Drug, and Cosmetic Act (FD&C Act) has, for decades, allowed FDA to take action against adulterated and misbranded medical devices. Those actions have included medical device seizures, injunctive actions against companies, and criminal prosecutions of individuals and corporations.

FDA practices in the enforcement area, including the issuance of warning letters and litigation against device companies, have generally been appropriate and consistent. While there are some who would disagree with this assessment, it is clear that rarely does FDA bring a criminal prosecution or take draconian action against a medical device company without some clear evidence of wrongdoing. A review of a few of the most prominent FDA actions helps prove this point. On August 31, 1988, the government filed a 24-count information against Cordis Corp. charging it with violations of the FD&C Act. The charges related to the shipment of adulterated and misbranded pacemakers. In 1989, Cordis pled guilty to 25 criminal counts and agreed to pay close to $6 million in civil and criminal penalties. After a month-long trial, all individuals charged in the case were acquitted.

The Cordis prosecution was followed by the prosecution of C. R. Bard Inc. On October 14, 1993, Bard pled guilty to a 391-count information, again related to the shipment of adulterated and misbranded devices. Individuals were also prosecuted in that case.
In 2003, Guidant Corp. pled guilty to nine felony counts related to its failure to submit medical device reports to the agency in connection with its implantable aortic aneurysm grafts. The company ultimately agreed to pay $92.5 million in civil and criminal penalties for the settlement of that case. No charges against individuals have yet been brought.

While there have been numerous other device company criminal prosecutions and enforcement actions over the years, I mention these three actions for a reason. They were brought against large medical device manufacturers, and they demonstrate how rarely FDA and the Department of Justice (DOJ) have had to bring major cases against the medical device industry. Under the present FDA and DOJ regimes, it appears that this selective and discrete enforcement will continue. There will always be overzealous prosecutors who will want to change this long-accepted balanced approach, but overall, the FDA and DOJ enforcement policies have largely been moderate and reasonable.

In addition to the agency's authority to criminally prosecute medical device companies, SMDA also gave the agency authority to seek civil penalties. Under this act, anyone who violates a requirement of the FD&C Act relating to devices is liable for a civil penalty of up to $15,000 per violation, with a maximum of $1 million for all violations. Again, FDA has used its civil penalty authority discretely, and few companies have been subject to civil penalties since 1990. Some have argued that FDA should have been much more aggressive in both its prosecution of device companies and the application of civil penalties. Overall, however, I think that the agency's enforcement policies in the medical device area have been balanced and, in most cases, fair.

The Future

It is always difficult to predict the future of medical device regulation or technologies. But it is perfectly clear that with new technologies, both industry and FDA will face challenges. One major challenge relates to the quickly advancing field of combination products. Under Section 204 of MDUFMA, Congress mandated the formation of an Office of Combination Products (OCP). That office, now under the leadership of Mark Kramer, a former CDRH official, has been extremely active. It is seeking to streamline FDA's review of the rapidly evolving field of combinations of drugs, devices, biologics, and human tissues. FDA must be extremely creative in regulating this area so as not to stifle new developments, such as drug-delivery devices. The agency has been holding public workshops and meetings to address how it might better regulate these new technologies and foster better cooperative efforts between its own centers.

Another area of rapid development has been IVDs. CDRH created the new Office of In Vitro Diagnostic Device Evaluation and Safety (OIVD) to provide one organizational unit for cradle-to-grave regulation of IVDs. The office's duties encompass the premarket review responsibilities of the Office of Device Evaluation, the enforcement responsibilities of the CDRH Office of Compliance, and the postmarket surveillance responsibilities of the CDRH Office of Surveillance and Biometrics. 

The rapid advancement of home-use tests, genetic testing, pharmacogenomics, and home-brewed tests by clinical laboratories raise many significant issues for the IVD industry and OIVD. Whether former CDRH director David Feigal's vision of device regulation based on the “total product life cycle” will be an advance in regulatory philosophy is questionable. Nevertheless, suffice it to say at this point that OIVD has been an interesting experiment, and device regulation based on the total product life cycle may in fact be the basis for future regulatory advances.

Finally, one cannot write about developments in the field of medical device regulation without addressing quality system regulation (QSR) issues. One yet-untested CDRH initiative involves using accredited persons to conduct factory inspections of U.S. and European Union device manufacturers. MDUFMA authorized FDA to accredit third parties to perform QSR inspections of eligible manufacturers of Class II and Class III devices. This voluntary program gives certain manufacturers the option of requesting inspection by an accredited person rather than by FDA. However, the jury is still out as to whether the accredited persons program will be a significant benefit to the industry. At this point, FDA has accredited 14 entities.

FDA also has a program allowing accredited third parties to review 510(k) premarket notifications for certain devices. It is clear that these third-party product review and inspection programs are part of a more robust global harmonization agenda. Third-party reviews and the involvement of accredited bodies to oversee the device industry has been a well-accepted structure in the European Union and elsewhere. Ultimately, in a perfect world, the regulatory schemes for devices would be harmonized worldwide, with consistent review and inspection practices across national lines. That is a future that would benefit the industry, patients, and regulatory authorities. Hopefully, such a rational consolidated regulatory scheme will be an important future part of what has become a global device industry. 

Copyright ©2004 Medical Device & Diagnostic Industry

Industry Pioneer

Originally Published MDDI August 2004


Earl Bakken

Earl Bakken has lived many lives in the course of his eight decades. He is best known as the inventor of the first wearable, battery-operated external pacemaker and the founder of Medtronic Inc. But his contributions to medical technology go beyond tangible accomplishments or statistically measured achievements. Through his work as an inventor, entrepreneur, and philanthropist, he has brought hope to people who once had no hope of leading healthy, full lives.

The Minnesota native who founded Medtronic in his garage in 1949 changed healthcare not only with his scientific discoveries but also with his philosophy. His emphasis on the human side of healthcare is what distinguishes him from others in his field. He firmly believes that “there is a human side to healing that must be considered” when developing new medical devices or technology. Since 1959, patients have been invited to Minneapolis to share with thousands of Medtronic employees how the company's products have improved these patients' lives. 

In 1989, Bakken formally retired from Medtronic, but from his home in Hawaii he continues to push medical technology forward. Determined to make his retirement more than just an opportunity to swing in a hammock, he soon found himself drawn to the needs of his of community. The result is the North Hawai'i Community Hospital, a state-of-the-art facility on the island of Waimea, a self-described “cow town” on the Kohala Coast. 

The hospital adheres to Bakken's code of holistic medicine. It combines the latest technology in Western medicine with “high touch” techniques, such as massage, acupuncture, and herbal remedies, which may not necessarily have a mountain of statistical evidence proving their effectiveness. 

“I've devoted most of my life to the development of leading-edge electronic devices,” he writes in his autobiography, One Man's Full Life, “but I've learned that technology is most effective—in many cases is effective only—when combined with the awesome power of the human mind, spirit, and hand.” 

Copyright ©2004 Medical Device & Diagnostic Industry

Industry Pioneer

Originally Published MDDI August 2004

Industry Perspectives

George Miser

George Miser, a pioneer in the nonwoven tape industry, started his career as a chemist for Transparent Package Inc. after receiving a degree in chemistry in 1955. Getting the job was a more momentous feat than it seemed at the time. 

It was at Transparent Package Inc. or TEE PAK, as it was known, that he met his future wife, Patricia. But before they could ride off into the sunset, Miser was called to duty by Uncle Sam. In 1956, he was drafted by the Army Medical Corp. to serve overseas as a medic at a clinic in Germany. 

“We did night duty,” he recalls. “We treated stab wounds, auto accidents, we delivered babies. We worked at night and on weekends. After that I knew I wanted to get into healthcare.”

When his two-year stint in the army was finished, he returned to TEE PAK, where he became a supervisor in quality control. During this period, he was also active in the American Chemical Society and the Quality Control Society of America. 

In the late 1960s, Miser relocated to St. Paul, MN, to work for 3M Medical Specialties, attracted by the career opportunities available at the company. At 3M, he marketed Micropore Surgical Tape, Transpore Surgical Tape, and Tegaderm Surgical Dressing. It was a fortuitous career move. For the next quarter century, he focused on medical specialties and revolutionized the tape industry in the process.

The use of nonwoven tape had a profound effect on patients. Their skin didn't redden, unlike with zinc oxide tape.

“If it wasn't for 3M, I never would've ended up in marketing,” Miser said. “I would've ended up selling sausage casings.”

Miser saw the advantage of nonwoven tapes for use with ostomy and transdermal drug delivery products. In 1986, he was appointed international sales manager to bring this technology, and others like it, to the rest of the world. 

He retired from 3M in 1994. He lives in White Bear Lake, MN.

Copyright ©2004 Medical Device & Diagnostic Industry

Human Factors: Moving in the Right Direction

Michael E. Wiklund is vice president in charge of the Human Factors Research and Design Group at the American Institutes for Research (Concord, MA).

Are you familiar with the terms ergonomic and user-friendly? Do you, on a daily basis, think about whether the products you develop or use are intuitive? Chances are, you answered “yes” to both questions. This signals a substantial change in the visibility and effect of the multidisciplinary profession called human factors engineering. Twenty-five years ago, the profession had a relatively low profile in the medical domain. However, it had the potential to make significant contributions to the quality of medical devices requiring user interaction. Fortunately, it has a higher profile today, enabling it to have a positive influence on the product development process.

Fundamentally, human factors engineers focus on making products safe, effective, easy to use, and appealing. With roots dating back to the early 1900s, the profession gained ground during the mid-to-late twentieth century, as companies dealt with increasingly complex electromechanical and computer-based machinery.

Still, even in the late 1970s developers and consumers focused little attention on user interaction quality as compared with other engineering and design considerations, such as reliability and visual appeal. In short, most of the user-interface design work remained a seat-of-the-pants endeavor that drew on an engineer's common sense. This was particularly true in the medical domain, where there was a presumption that any medical device would be used by trained professionals. Everyone believed that their training would make up for any design shortcomings.

Paul Kirley, Stuart Karten, and Dennis Schroeder examine the Clarion Speech Processor by Advanced Bionics. The device works with a cochlear implant to help people hear.

Decades ago, if a medical device was difficult to learn to use, caregivers were more likely to blame themselves rather than the device, and they would find ways to cope. Flash forward to the present and you find the pattern has practically reversed itself.

Now, caregivers are well-acquainted with well-designed products matching their physical, intellectual, and emotional needs, partially because of products proffered by the consumer electronics and software industries. They are now more likely to call a product obstreperous than blame themselves for their ineptitude. This goes double for nurses, who normally take a can-do attitude toward arduous tasks, including some that intimidate physicians, but have little patience for ill-behaved equipment. In fact, they have been known to make emphatic statements, such as “I wouldn't want that thing on my unit,” after trying their best to use a product that seems designed by engineers for engineers.

Indeed, society as a whole and the medical community in particular have come to demand products that are engineered to be intuitive, feel right, and work nicely. And, it's about time—user-friendly medical devices are generally safer devices. Intuitive operation and lower error rates generally go hand in hand. 

Raised Expectations

Poorly labeled controls and maintenance tags that covered 
important gauges contributed to operational difficulties 
associated with the Three Mile Island accident.

Taking a broad view, design expectations started to shift in the early 1980s, accelerating through the 1990s. The shift occurred because of many factors, including the accident at Three Mile Island (TMI) in March 1979, which culminated in a partial core meltdown. Even though that statement seems farfetched, TMI really did influence consumer expectations about human factors engineering. It helped bring the discipline to the public's attention and created a demand for human factors engineers.

TMI accident investigators determined that simple things, like poorly labeled controls and maintenance tags covering important gauges, contributed to operational difficulties associated with the incident.1 In response to this finding, the U.S. Nuclear Regulatory Commission (NRC) directed all domestic nuclear power plant operators to conduct human factors reviews of their control rooms. These reviews identified many human factors engineering deficiencies. These led to control-room renovations that promised to improve safety through better operability. 

The NRC's safety initiative created a strong demand for human factors engineers that lasted about a decade. Later, when the power plant renovations were largely completed, the demand waned. This led some human factors engineers to migrate to the medical domain because it posed some of the same analytical and design challenges, particularly controlling complex systems through a relatively simple user interface. However, human factors engineers emigrated in greater numbers to the software industry. Companies like Microsoft and Oracle built large usability engineering groups to enhance their applications.

It was also about 25 years ago that the computer industry was rolling out a new kind of product called the personal computer. The first desktop machines employing the disk operating system (DOS) were suited for use by people who knew something about programming. Those with little or no prior computer experience found the machines quite daunting. In fact, their travails were the subject of numerous cartoons depicting people at their wit's end: images of clenched fists, red faces, and steam blowing out the ears.

People with little or no prior computer experience found that they could use the first Apple Macintosh computer easily.

But along came the Apple Macintosh computer—some consumers' first introduction to complex operations made simple and accessible, the automobile notwithstanding. Suddenly, novices could draw upon their basic intelligence and intuition to accomplish computer-based tasks, ranging from word processing to financial management to artwork production.
The race was on to make computers, software applications, and other aspects of high technology more user-friendly. Jump forward a quarter century and you find a marketplace in which the majority of manufacturers, including medical device manufacturers, make bold claims about their products' ease of use. In some cases, the claims have substance because they arise from a developer's investment in human factors engineering. It is a process punctuated by usability tests that prove, or disprove, that representative users can operate the targeted products.

Many medical devices seem to be getting easier to use and safer in the process. However, Julian Goldman, MD, a practicing anesthesiologist at Massachusetts General Hospital and professor at Harvard Medical School, is reserving a final judgment. The former chair of ASTM's Committee F29 on Anesthetic and Respiratory Equipment says that “things are getting better slowly—slowly—slowly.” Reflecting on his observations after almost two decades of clinical practice, he cites the lack of equipment integration as a major hurdle to greater progress. “Manufacturers continue to assess their devices in isolation,” he notes. “They are not addressing all of the environmental factors, including other equipment and different types of users, in their design development process.” 

Goldman adds that “some manufacturers may face economic limitations. Good human factors practice can be time-consuming and expensive. It may be hard to make an argument in favor of a large investment when a product has a short life cycle.…Manufacturers are going to be tempted to let consumers be their beta testers, knowing they [the manufacturers] can fix things for the second release. Of course, this is not the best approach from a patient safety perspective.”

While Goldman's comments ring true, there also are more medical company CEOs than ever before who are familiar with human factors engineering—a big change from years ago. And some have made usability an important part of their brands, recognizing that customers frequently cite usability as one of their highest design priorities. 

Some company leaders have learned that human factors engineering makes good economic sense, considering potential rewards like faster product development cycles and increased sales linked to customer preferences for a “user-friendly device.” Also, reduced demand on customer support and reduced liability exposure have good implications for a company. To capitalize on these rewards, firms have invested substantially in human factors. They have either set up internal specialty groups or retained human factors consultants to work with their engineering and marketing teams. For example, several large companies, including 

Abbott Laboratories, Alaris Medical Systems, Ethicon Endo Surgery, GE Medical Systems, Medtronic, and Siemens Medical Solutions, either employ or engage human factors specialists who focus on ensuring each product's safety and giving it a competitive edge. Many smaller companies do this as well, although they tend to hire consultants in lieu of being able to afford a full-time specialist. 

Still, some medical companies have yet to embrace human factors engineering as a key to customer satisfaction and quality care. Goldman attributes this condition to the fact that human factors investments pay off mostly in the long term. Thus, manufacturers may have an easier time marketing products that are user-friendly, but the benefits of higher workforce productivity and satisfaction, for example, accrue to their customers, such as hospitals.

To prevent needle sticks, a metal fitting extends over the end of the needle of the B. Braun Medical Inc. Introcan Safety Intraveneous Catheter.

Meanwhile, the complexity of technology seems to be on another upswing, largely due to the increasing power and decreasing cost of microprocessors. So the industry may be at a turning point. Companies must decide between spending more on human factors engineering to ensure the ease-of-use of devices, and limiting their spending, thereby triggering a decline in the caregivers' ability to cope with increasingly complex devices.

And speaking of device complexity, one has to wonder whether caregivers have already reached their limit in terms of computer-based interactions replacing good old knobs and dials. More medical devices than ever require users to navigate menus and to mouse-click. Meanwhile, recent studies indicate that a decreasing percentage of doctors know how to use a standard stethoscope effectively. These may be symptoms of the shift from a more hands-on style of equipment interaction and care to one that is mediated by computers, as well as the associated shift in medical education.2

Toward a User-Centered Design Process 

With due respect to those who oppose government regulation, FDA's updated quality system regulation that calls for the application of good human factors practice in medical device development looks like progress.3 For years, human factors engineers who struggled to have an effect in various medical companies hoped for such a regulation. Regulating the application of human factors in medical device design would give them ammunition to use in the battle for influence within their engineering departments.
Historically, other engineering considerations tended to trump human factors requirements when push came to shove within engineering teams. The regulatory imperative to apply good human factors brought a welcome change in the ability of human factors engineers to make a difference. They were able to push back on technical solutions that were not user-friendly and to encourage those that were. 

The Phillips HeartStart Home Defibrillator uses extensive voice prompts to talk people through the delivery of cardiopulmonary 

FDA's updated regulation, released in 1996, stresses the importance of identifying user needs as a precursor to user-interface design. The agency's Human Factors Program's Web site states, “Research has suggested that the frequency and consequences of medical device use errors far exceed those arising from device failures. Therefore, product developers must consider device use and use-related hazards to ensure that their devices will be safe.”4

FDA's prescribed development model encourages designers to build devices in response to demonstrable user needs rather than the desire to showcase new technology. In addition, the regulation calls for user feedback during the development process so that problems can be addressed before the product goes to market. This means that the initial customers and patients will no longer have to serve as the guinea pigs. 
Prior to regulatory change, manufacturers tended to address human factors requirements in an informal, ad hoc manner. In the late 1970s, most medical device manufacturers equated human factors engineering with market research. Developers considered obtaining design feedback from preferred customers and folks visiting trade show booths and hospitality suites to be a reasonable strategy for addressing user needs in the design process. As one might expect, this approach yielded unreliable user preference and performance data, if any, and burdensome designs.

For example, a new generation of microprocessor-based devices, such as patient monitors and infusion pumps, came to market in the late 1980s. They seemed better suited for use by engineers and programmers than by computer-phobic caregivers. Consequently, caregivers spent a lot of time adjusting to the new technology, which distracted them from direct patient care. You also saw caregivers developing coping strategies, such as learning by rote to perform the basic tasks, and ignoring the rest of an instrument's advanced capabilities.

Still, Goldman regards the kind of research conducted in hospitality suites as better than no user research. He strongly advocates a formal process for collecting design input from users at several stages of development. But, he adds, “Showing mock-ups in a hospitality suite provides some amount of feedback that can improve a product.”

Today, you see exemplary companies working harder and more effectively at user-interface design. They establish human factors programs that ensure the systematic definition of user requirements (which FDA calls design inputs), followed by the rigorous application of human factors in design and iterative usability testing. The resulting design and associated performance data (which FDA calls design outputs) typically set the stage for a smooth regulatory approval process and good customer acceptance of the final product.

In addition to FDA, the Association for the Advancement of Medical Instrumentation (AAMI) warrants recognition. They have produced a continuing series of standards for the application of human factors engineering principles to the design of medical devices. AAMI published its first human factors standard in 1988, which drew extensively from military standards on human factors. Thirteen years later, the standards organization released ANSI/AAMI HE74:2001, “Human Factors Design Process for Medical Devices.” The document, which has been endorsed by FDA, describes alternative ways for companies to conduct an effective human factors program.

Notable Products

An informal survey of several human factors professionals working in the medical industry identified the following products as landmarks in terms of their human factors engineering.

Pulse Oximeter. The Nellcor N-100 pulse oximeter, dating back to the early 1980s, is a landmark product. Providing a simple and intuitive indication of oxygen saturation level through multiple sensor channels, the product proved to be a boon to patient safety. It was and remains extremely popular among clinicians who became devotees due to its pleasing interaction style and clinical value. Specifically, the compact device provides a numeric value for oxygen saturation (i.e., 98%). The numeric reading is complemented by a vertical colored bar that pulses up and down with each heartbeat, emitting a tone keyed to the patient's oxygenation level (the higher the pitch, the higher the patient's oxygenation level).

Needle-Stick-Prevention Devices. Shifting the focus away from microprocessor-based devices and toward hardware, needle-stick-prevention devices presented the development community with an interesting human factors challenge in the late 1990s. The challenge was to keep caregivers from infecting themselves with blood-borne diseases such as HIV due to needle sticks. Numerous inventors tackled the challenge, producing a bumper crop of needle-stick-prevention needles. Some of the designs are better than others at enabling caregivers to give injections or place a catheter without extra fuss. B. Braun Medical Inc.'s Introcan Safety Intraveneous Catheter is one of the best, garnering a Medical Design Excellence Award in 2003. After a single use, a metal fitting extends over the end of the needle to prevent a needle stick or an unauthorized second use. Human factors aficionados and nonspecialists alike admire the device's elegance in terms of its protective and unobtrusive characteristics. In addition, caregivers are able to continue to use their conventional methods of introducing catheters.

Automatic External Defibrillator. The Philips HeartStart Home Defibrillator has garnered several design awards for its compact form and usability. The device is designed for laypersons to operate in the event of a cardiac emergency. It uses extensive voice prompts to talk people through the delivery of cardiopulmonary resuscitation (CPR) as well as the delivery of a cardioverting shock. Significantly, the developers conducted extensive user-needs analyses and usability testing during the course of developing the product's widely admired user interface.

Patient Monitor. Many within the anesthesia community would agree that Datex Medical Instrumentation challenged the major players in the mid-1990s by introducing the AS/3 monitor. It was arguably the first integrated patient monitor designed specifically for use in the OR, as opposed to critical-care units. New users found it remarkably easy to use and well suited to the needs of anesthesia providers. For example, they liked the fact that waveforms were accompanied by large numeric readouts. They also appreciated the integration of gas-concentration measurements (e.g., O2, N2O, and anesthetic agents) along with hemodynamic measurements (e.g., respiratory rate, heart rate, and blood pressure). Though technical in nature, these enhancements reflected a sensitivity to the specific information needs of anesthesia providers in the OR versus nurses in critical-care units. Since the AS/3's introduction, Datex (now GE Medical Systems) has gained a significant portion of the U.S. market for integrated patient monitors.

Infusion Devices. The Alaris Medical Systems Medley Medication Safety System allows users to run multiple intravenous infusions via individual pumps connected to a central programming module. The system's defining characteristic is its Guardrails Safety Software that helps to protect patients from medication errors. Hospitals use the software to establish drug dose limits that caregivers may not exceed, thereby eliminating the potential of an overdose due to a data-entry error. The use of software to help protect users from making lethal errors heralds a new era of devices with built-in safety systems that extend beyond switch guards and alarms.

The Baxter Healthcare/Bard Infuse OR, a syringe pump, has been a favorite product among clinicians for many years. Introduced in 1988, the product simplified the task of determining a safe medication dose through the use of interlocking magnetic panels that attach to the product like a refrigerator magnet. The panels, which are uniquely tailored to specific medications, communicate electronically with the pump's microprocessor to ensure appropriate dosing schemes, which are reflected by the panel labeling.

Glucose Meters. First-generation glucose meters were a breakthrough in terms of giving diabetics the ability to measure their blood glucose levels on a daily basis. However, many models were compromised by the large number of steps in the test process, the considerable dexterity requirements, and the need for careful timing. The latest models, such as Lifescan's One-Touch Ultra, are impressively compact and require a few simple steps. The One-Touch Ultra presents the user with a result within 5 seconds via a large, easy-to-read display.

Looking Ahead

What lies ahead regarding the application of human factors engineering in the design of medical devices? The fundamental challenge remains to make human factors engineering one of the central considerations when developing medical devices that require substantial user interaction. 

Even with the FDA quality system regulation in place, some companies still pay lip service to human factors engineering instead of making a meaningful investment in it. Fortunately, the marketplace is likely to correct this problem over time—possibly a long period of time—as customers come to reject hard-to-use offerings. In fact, some would argue that market demand, informed by user-interface-design excellence in other industries, is the real driving force toward better medical device user interfaces, as opposed to government regulation.

Aside from developing a design process that complies with FDA requirements, one of the biggest challenges for many device manufacturers will be smoothing the information management process. This is a challenge that extends beyond the design of a single device. The clinical environment is already awash in information generated by all sorts of devices, ranging from patient monitors, to infusion systems, to hospital beds, to wristbands that enable patient identification and tracking. Getting all of these disparate devices to work in harmony, particularly with regard to alarm messages, requires not only good design practice but also industry standardization so that they speak a common user-interaction language. The proprietary interests of various companies may create considerable hurdles to achieving this goal. In short, many manufacturers are committed to their brand of user interface and are loath to change. They want to avoid any added development expense and the loss of unique characteristics that serve to differentiate their products from others.

Goldman feels that devices need to get a lot smarter. He describes some of the alarms generated by devices such as patient monitors as totally unnecessary nuisances, noting that intelligent devices that adjust to the context of use would not be crying wolf. For example, Goldman says, “a patient monitor should communicate with anesthesia recordkeeping software to know when a patient is going on to cardiopulmonary bypass, and suppress the apnea alarm while activating other alarms in a context-sensitive manner.” As another example, Goldman presents the low-battery alarm. “It's fine for a device to alarm when there is 15 minutes of battery time remaining,” he says. “But after you acknowledge the first alarm, it should not alarm again every single minute.”

Goldman believes that the work that is under way in Massachusetts General Hospital's Operating Room of the Future will help pave the way toward the development and adoption of smarter medical technology that enhances information management. Currently, the R&D program is looking at several opportunities, including wireless communication between patient sensors and data collection devices.

Making medical devices easier to learn to use is another frontier. Today, an increasing number of devices have on-line help systems, tutorials, and simulators available that enable people to practice using them. However, caregivers still prefer hands-on training. Their preference can be attributed in part to tradition. The department in-service, complete with tasty refreshments, is a cultural norm that makes learning a shared experience rather than a solo adventure. But the limitations of in-service training—including the difficulty of reaching everyone working in a multishift, high-turnover institution—are well known. So human factors engineers and other specialists have their work cut out for them, striving to create effective and appealing educational resources. Advanced computer-based simulations that require substantial user interaction, making the experience feel quite hands-on, will clearly be in the mix.

One more frontier might be dubbed, “Back to the Future.” We may be experiencing the peak of physical estrangement from medical devices that feel more like computers than hands-on, physically interactive tools. 

Caregivers have always liked direct manipulation: the ability to turn a knob or press a button. This is because the mechanisms are immediately accessible and provide tactile feedback in addition to visual and audible feedback. By comparison, selecting an on-screen icon, scrolling through a list of options, clicking on an “up” arrow, then clicking on an “enter” button, takes more time and effort. So look for a greater number of dedicated knobs and dials to balance out screen-based interactions.

Human factors engineers will also play an important part in ensuring the interactive qualities of complex medical devices used in the home, such as dialysis machines. Here, the key will be to create user interfaces that do not call for advanced medical and technological know-how, but rather lead users effectively through operating procedures as well as help them overcome problems that arise. Meeting this challenge is especially important given the pattern of migration of medical devices from clinical environments into the home and workplace.


You could say that human factors engineering in the medical domain has come a long way, emerging from a dark age of design negligence. But to be fully applicable, the tag line would have to read, “You've come a long way, baby...and you've got a ways to go.” Generally, manufacturers still do not regard human factors engineering as a fundamental need, equivalent to mechanical engineering or software programming. When a widespread changeover finally occurs, it will assuredly improve patient safety and reduce healthcare costs. Importantly, the added investment will raise the quality of life-critical medical devices to the level of the better kitchen tools and digital music players!

Still, the medical device industry has some shining examples of good human factors engineering, ranging from pocket-sized testing devices to large diagnostic scanners. These are winning design awards and gaining marketshare. They create a desire among device users for more devices of similar and even greater quality. So, the human factors movement in the medical industry is certainly heading in the right direction.


1. “The Report of the President's Commission on the Accident at Three Mile Island,” in [on-line], (Dover, AR: 30 October 1979); Available from Internet:  
2. Steve Salvatore et al., “Study: Most New Doctors Can't Use Stethoscope,” in [on-line] (Atlanta: 2 September 1997); Available from Internet:
3. Code of Federal Regulations, 21 CFR 820.30.
4. “Why Is Human Factors Engineering Important for Medical Devices?” in CDRH Home Page [on-line] (Rockville, MD: FDA, Center for Devices and Radiological Health, 2003); Available from Internet:  

Copyright ©2004 Medical Device & Diagnostic Industry

Industry Pioneer

Originally Published MDDI August 2004

Industry Perspectives

Brian Baldwin

Brian Baldwin sure knows how to pick his friends. In 1958, he and college buddy William Cook founded MPL Inc., a company that made its mark developing proprietary technology for manufacturing hypodermic needles. However, that was only the beginning for both of them. 

In 1962, Cook left MPL to start the phenomenally successful Cook Group Inc., and Baldwin built up MPL and later founded Solopak Laboratories. Then, in 1975, Baldwin cofounded Baxa Corp. with Ronald Baxa. The company has transformed the medical field with its syringes for precision oral liquid administration, microprocessor-controlled peristaltic pumps for medication preparation, and computer-controlled mixing systems for IV solutions. 

Since its inception, Baxa has grown from a struggling startup with a 10-member staff to a worldwide corporation with 275 employees and $50 million in annual sales. The company's reach stretches from Denver to Denmark, with additional offices in the UK, Canada, and Germany. 

When Baxa was founded, oral liquid medications were measured with teaspoons or graduated cups, which resulted in inaccurate or even harmful doses. Baxa's Exacta-Med oral dispenser allowed medical personnel to administer doses much more precisely and soon became the industry standard. Products such as the Repeater pump, MicroMacro TPN compounder, and Rapid Fill automated syringe filler further established Baxa's reputation as an industry leader and innovator. 

Through the years, Baldwin has led the way in product design and development. Although he stepped down as Baxa's chairman of the board in 2003, he is still involved in the company as chairman emeritus and chief development architect. He continues to work closely with the company's product development teams.

Baldwin holds more than 30 U.S. and international patents and was one of the founders of the Colorado Medical Device Association. He is also a member of the World President's Organization and the Economic Club of Colorado.

Copyright ©2004 Medical Device & Diagnostic Industry

Industry Pioneer

Originally Published MDDI August 2004

Regulatory Affairs

Kshitij Mohan

Throughout his career, Kshitij Mohan, MD, has displayed his commitment to the medical device industry. His unflagging dedication to device companies and to the government make him a true pioneer in the industry.

Early in his career, Mohan served in various capacities in FDA, including heading up science and technology programs and all product evaluations and medical device approvals. During his service as director of the Office of Device Evaluation and acting deputy director of CDRH, all backlogs in all categories of medical device applications were cleared. Many of the policies he established were later codified into revisions of the medical device laws, including the often-cited “Mohan Memo” of 1986. That memo was the basis of subsequent legislation related to the premarket 510(k) notification program.

After his stint with FDA, Mohan went to Baxter International in 1988. There, he was a corporate vice president and was responsible for all corporate research and technical services. His responsibilities there included supervision of centers in the United States and in Europe. He was also the founder and leader of the Baxter Technical Council, and he served on the Baxter Operating Management Team.

Mohan joined Boston Scientific Corp. in 2000. There, he served as senior vice president and chief technology officer. He led research, development, clinical affairs, regulatory affairs, and government affairs units as well as corporate techonology centers. 

Mohan then became the chief regulatory and technology strategist at the Washington, DC, offices of King and Spalding, a national law firm. In this role, he worked with medical device, biologics, and pharmaceutical firms on product development and FDA approval strategies. He also assisted firms with clinical trials, quality systems and regulations, ventures and acquisitions of medical technology firms, and public policy issues.

Mohan then joined International Remote Imaging Systems Inc. as the chief executive officer. During his leadership of the company, it launched a major new product platform, established global distribution, and dramatically increased sales and the company's financial 

Currently, Mohan is CEO of Cytomedix Inc., a public biotech company. The company has a unique proprietary technology for treating nonhealing chronic wounds such as pressure ulcers, diabetic foot ulcers, and venous stasis.

In addition, he has been a member of numerous boards, including the boards of directors of AdvaMed, KeraVision Inc., and the National Academy of Sciences Roundtable on Drugs, Devices, and Biologics.

Copyright ©2004 Medical Device & Diagnostic Industry

1979–2004: Milestones in Medical Device Packaging

Originally Published MDDI August 2004


The past 25 years have been revolutionary for medical packaging, but it just may be that the best is yet to come.

Carl D. Marotta

Carl D. Marotta is the president 
of Tolas Health Care Packaging (Feasterville, PA).

Sometimes, synergy emerges in the most unexpected places and circumstances. Such a marriage—whether in science, industry, or art—usually results in significant advances in the state of the art. In the 1970s, a need for high-performance packaging, driven by a rapidly growing device market, presented the right conditions for just such synergy. Device manufacturers needed functional packaging for their sterile disposable (single-use) devices. They found themselves on a hunt for a product that did not widely exist—a package that could be sealed, sterilized, packed, and shipped, and then stored and opened easily and cleanly at the point of use.

A Little History

The state of the art in medical packaging prior to the early 1970s consisted primarily of film and paper bags, pouches, or boxes enclosed in an overwrap designed to protect the sterility of the package's contents. Traditional sterilization processes, such as steam or dry heat, were suited for only a few materials, such as vinyl-coated papers.

At the time, devices that required environmental barriers were packaged in aluminum-foil pouches, bottles, glass vials, or injection-molded barrier plastics. All of the available methods and materials fell short of the characteristics required for packaging sterile disposable devices. And, the term clean peelability was not in wide use. In some cases, an acceptable clean peel was determined by simply separating the top layer of a paper package and rating the degree of fiber tear.

The porosity was measured in Gurley seconds, and values of several hundred seconds were commonly deemed acceptable. However, with this porosity, a penalty was paid in the form of slow ethylene oxide (EtO) sterilization cycles. Packaging adhesives were limited, and those available gave manufacturers little control over the sealing process. The adhesives also were not user-friendly, and thus they were not high on customer satisfaction lists.

The Quest for Sterile Packaging

So, the search was on for tray lidding and pouch materials that would meet a growing list of exacting performance characteristics for sterile medical device packaging. Coincidentally, the key material that ultimately filled the bill was being introduced to market at the same time. DuPont's medical-grade Tyvek redefined packaging for the late 1970s. Not long after, Canon Communications launched MD&DI, and I began to write a column called “Packaging Forum.”

With the convergence of these events, a market hungry for a high-performance sterile packaging was satisfied. This synergy made medical-grade Tyvek the most important packaging material used in the healthcare industry then, now, and for the foreseeable future. The combined contributions of DuPont, packaging converters, coaters, and printers revolutionized medical device packaging in the 1980s. The sidebar, “The Role of Heat-Seal Coating in Seal-Peel Medical Packaging,” provides the functional and process criteria that illustrate how combinations of substrate, coatings, and process parameters work together.
The 1980s and 1990s were noted for the rapid growth of medical device packaging and the growth of the key players in this market. Maintaining its leadership position worldwide for its strategically important material, DuPont continued to lead the market by working closely with major medical device manufacturers and packaging converters. The market for sterile device packaging expanded, particularly in the development of film-based pouch and form-fill-seal (FFS) packaging. 

During these decades, heat-seal coatings were developed for specific applications. Other significant materials-related developments include:

• Sealant films (with uncoated Tyvek). This combination provides a cost-effective solution for pouches and FFS forming webs.
• Specialty coatings. These coatings are ideal for high-performance applications where quality is critical and the ability to seal to semirigid trays is essential.
• High-temperature coatings. These coatings are suitable for dry-heat sterilization and autoclaving.
• High-altitude lidding. This lidding can withstand the internal pressures generated at high altitudes.
• Water-repellent aqueous coatings. Rather than providing a moisture barrier, these coatings repel water.
• Zoned-coating lidstock. This lidstock is designed for die-cut or FFS packaging.
• Autoclavable packaging. Both lidding and pouches can now withstand the autoclaving process.

Trends in the 2000s 

The future for sterile device packaging holds both challenges and opportunities. First and foremost, packagers must listen to customer requests. Just when we think every material has been invented, a customer will require performance just beyond what's available.

Of the hundreds of inquiries received annually, each should be treated as important. Most requests can be resolved easily with currently available materials. It is the exceptional request that may point to future needs and could lead to the next packaging breakthrough. Packagers and package developers must focus on four primary areas:

• Functionality. Products of the future should enable customers to evaluate a range of coatings in terms of seal strength and other characteristics for an ideal solution.
• Lean manufacturing. With the increased adoption of six-sigma and other lean manufacturing methods, packagers should be able to eliminate waste and maximize efficiencies across their operations.
• Costs. It is unclear whether the industry can deliver packaging that is more cost-effective. It will be essential for packagers to look at alternative ingredient sources as well as source reduction.
• Prevalidated materials. Products are needed that will reduce sterilizer cycle time and provide higher temperature resistance. Ultimately, customers will demand higher quality at a lower cost.

The next decade will require improved materials that provide more functionality and more flexibility. They will need to meet international standards. Key advancements will likely include:

• Zone-applied coatings that virtually eliminate coating and product contact.
• Product lines that are fully ISO 2002 and ISO 11607 compliant.
• More designed structures, possibly all films or film-foil combinations.
• Multilayer thermoformable plastics.
• More grades of Tyvek with functional distinctions.
• Multifunctional pouches that offer removable sterilization vents with a wide assortment of layers.
• On-line printable, bar-coded film pouches in roll form. 
• Hybrid applications that combine device and drug packaging features.

It's not out of the question to see the development of a package based on nanotechnology, the science of ultrathin surfaces. Some large companies have identified nanotechnology as a key focus area. We all can dream about it, knowing that someone is giving it serious thought.   

Copyright ©2004 Medical Device & Diagnostic Industry

A Long Road: 25 Years of Clinical Research

Originally Published MDDI August 2004

Clinical Research

Part 812 in 1980 recognized the need for clinical research laws for medical devices. Many revisions later, the industry's clinical trials practices came into their own.

Nancy J. Stark

Nancy Stark is founder and president of Clinical Device Group Inc. (Chicago), an outsourcing firm for biological safety and clinical research.

I took my first job as a clinical research associate in May of 1980. I had never heard of clinical research, but I was a new graduate in biochemistry and I needed a job, so I agreed to work in the field. My boss warned me: “This will not be research as you know it in the laboratory.” 

In January of that year, something monumental had happened in the medical device industry. The Investigational Device Exemptions, 21 CFR Part 812, had been published. Clinical research in the device industry had come into its own.1 The regulation became effective on July 18, 1980; I remember the date well because our regulatory affairs manager brought in a birthday cake to celebrate. The cake said, “Happy Birthday Part 812.”

I immediately fell in love with my new career. Clinical research is the perfect marriage of scientific investigation and human interaction. The best clinical research associates have both solid technical skills and an understanding of human nature.


Drug manufacture had been regulated since the inception of the Food and Drugs Act of 1906, which prohibited interstate commerce of pharmaceuticals that were misbranded or adulterated.2 But the clinical investigation of drugs wasn't regulated until 1963. FDA didn't want to regulate clinical research; it was a political hot potato. But if you remember the 1960s like I do, it was the era of civil rights, and the nation had suffered through the Tuskegee project, thalidomide investigations, and plutonium injections. For an excellent historical review of human subjects research in the United States, see the Department of Energy's October 1995 report from the Advisory Committee on Human Radiation Experiments. The report is available on the Internet at

The sixties were an exciting, revolutionary, turbulent time of great social and technological change: assassination, unforgettable fashion, new musical styles, Camelot, civil rights, gay and women's liberation, a controversial and divisive war in Vietnam, the first manned landing on the moon, peace marches, flower power, great TV and film, and sexual freedom. America was in no mood for more unregulated human experimentation.

In this social climate, though, the regulations, 21 CFR Part 130—New Drugs, were published on January 8, 1963.3 But the medical device industry barely existed in 1963, and neither devices nor device clinical research were included.

Medical Device Amendments—1976

The Food, Drug, and Cosmetic Act wasn't amended until 1976 to include medical devices. I just so happen to have a copy of the 1976 law on my bookshelf. The whole of Chapter V, “Drugs and Devices,” is only 82 pages long, and the only mention of clinical research is under the description of Class III devices, the effectiveness of which “may be determined…on the basis of well-controlled investigations, including clinical investigations where appropriate, by experts qualified by training and experience to evaluate the effectiveness of the device….”4 The act spoke about clinical research as a sideshow or an afterthought, rather than a major step in the device development process.

Part 812—Investigational Device Exemptions

Table I. Part 812, then and now (click to enlarge).

It wasn't until 1980 that medical devices got their own regulations for clinical research: the Investigational Device Exemptions, or IDEs, were published on January 18 of that year. We were all confused about what exactly it was that investigational devices were exempt from (all the other device regulations). The basic format consisted of seven major subparts: A) General Provisions, B) Application and Administrative Action, C) Responsibilities of Sponsors, D) IRB Review and Approval, E) Responsibilities of Investigators, F) Informed Consent (this subpart is now Reserved), and G) Records and Reports. 

As with all regulations, the preamble is as important as the regulation itself. Even today, I refer clients who test consumer products on their employees to question 42 on page 3470: “One comment argued that ‘institution' should not include a manufacturer because a manufacturer should not be required to create an in-house IRB.” FDA disagreed, stating, “A manufacturer must be subject to the requirements of institutional review if an investigation is carried out on the manufacturer's premises, for example, using employees as subjects of the investigation. Employees are entitled to the same protections, including IRB review, as other subjects.”5 Table I compares the regulation from 1980 to 2004.6

Figure 1. This timeline shows the dates of revisions to 21 CFR Part 812; each date represents one or more amendments. The orange arrows indicate major related events: the medical device amendments, issuance of Part 812, the Safe Medical Device Amendments (SMDA), and the FDA Modernization Act (FDAMA) (click to enlarge).  

By my count, Part 812 has been amended on 22 separate occasions, as shown in the timeline in Figure 1. Each date may represent the revision of several paragraphs, and there have been more than 50 separate revisions to specific paragraphs. The revisions typically come in clusters. Here's a broad look at some of the most important ones.

May 30, 1980: Subpart F—Informed Consent Removed. The first major revision of Part 812 took place only five months after it was published. Subpart F—Informed Consent was removed and replaced with Part 50—Protection of Human Subjects. Publishing separate regulations for informed consent allowed FDA to have a common rule for devices, drugs, and other products regulated at the time. 

Many times, what is missing is more interesting than what is present. For example, in 1980, subjects were required to sign an informed-consent form, but not to date their signature. The requirement for signing and dating the consent form didn't appear until 1996.7 In another example, I was surprised to learn recently that Part 812 doesn't actually require an investigator to sign an informed-consent form, except in the case of obtaining oral consent from a subject.

January 27, 1981: 13 Revisions. In January 1981, FDA made 13 separate revisions to Part 812. Many of them had to do with institutional review boards (IRBs). 

812.42—FDA and IRB Approval Added. The relationship between significant- and non-significant-risk devices to full and abbreviated IDEs was (and remains) a source of continuing confusion. Significant-risk devices require full IDEs; non-significant-risk devices require only abbreviated IDEs. Section 812.42 clarified that both FDA and IRB approval were required for full IDE applications and supplements. We knew from other sources that non-significant-risk devices needed approval only from an IRB, which acts as an agent for FDA.

There are many other differences between full and abbreviated IDE requirements. For example, no protocol is required for abbreviated IDEs. Admittedly, it's impractical to get IRB approval without a protocol, but it would be helpful if the regulations were consistent with good clinical practices. 

812.65—Responsibilities and Procedures [of IRBs] Deleted. In another revision, Section 812.65 describing the responsibilities and procedures of IRBs was removed and replaced with Part 56—Institutional Review Boards. As with informed consent, publishing separate regulations for IRBs allowed FDA to have a common regulation for devices, drugs, and other products regulated at the time. 

812.66—Significant-Risk Determinations Added. Another major revision took place in 1981 with the addition of section 812.66. In this paragraph, FDA clarified that IRBs have responsibility for determining whether devices are significant risk. Other regulations tell us that the sponsor must make a contention (argument) to the IRB regarding the risk status of the device, but between the two of them, the IRB holds trump. 

April 26, 1985: 812.20—Application. Prior to 1985, the process for obtaining a full IDE for a significant-risk device was to submit an application to FDA, obtain its approval, and then submit the protocol to a local IRB for review (“…a sponsor shall not submit an investigation plan…to an IRB …before submitting an application to FDA…”). With the changes to the application process in 1985, FDA allowed more flexibility to the process. Now sponsors had a choice. They could submit a protocol to an IRB or to FDA first. If the IRB determined the device was significant risk, it was to notify the sponsor (through the investigator), who would then apply to FDA. This simple procedural change transferred much responsibility to the IRBs.

April 6, 1988: 812.20—Application [Environment Assessment]. Part 812.20 (9) was—happily—revised in 1988 to no longer require IDE applications to contain an environmental analysis report per Part 25. Now a sponsor could claim for categorical exclusion.
October 2, 1996: 812.47—Emergency Research and 812.38—Confidentiality. A significant provision was added to Part 812 in 1996, allowing for emergency research, i.e., allowing for research on persons who were medically unable to give informed consent. For the first time, there was a legal way to conduct research on trauma patients, comatose patients, Alzheimer's patients, car accident victims, or other patients who could not comprehend or give consent to participate in research. Before this time, such research could not be done legally in the United States. The provisions were intended to open up possibilities for important kinds of new product development. I remember they created such a sensation that newsletter editors would phone me, asking for interviews on the topic.

Concomitantly, Part 812.38 was modified to allow for public disclosure of information from IDE applications for emergency research. If you're interested in the information, you need only ask under the Freedom of Information Act. 

October 7, 1996: 812.1—Scope. The scope of Part 812 was changed in 1996 in a significant way: investigational devices became subject to Part 820.30—Design Controls. Never before had documentation of the design history or traceability of an investigational device been required, and the shift in policy served to finally bring the act of investigating a device into the product development process. Looked at from the manufacturing standpoint, a clinical study was now viewed as simply one more tool in the company's arsenal to verify and validate a design. In more than one corporate executive's eye, this change brought clinical research into the product development fold.

November 5, 1996: 812.140—Records. The records section was revised in 1996 with expanded definitions and requirements of investigators to keep records of each subject's case history. The definition of case history remains elusive: we know it includes case report forms, medical records, signed and dated consent forms, hospital charts, nurses' notes, and other documents; but what it includes and what it means are not the same thing. The issue matters because it is common for investigators to be cited for inadequate case histories.

March 14, 1997: 812.119—Disqualification of a Clinical Investigator. The clause regarding disqualification of clinical investigators was added in 1997. It placed a powerful tool at the sponsor's disposal for bringing recalcitrant investigators into compliance. For the first time, investigators who caused or attempted to cause false information to be submitted to FDA could be disqualified from participating in future clinical research. The disqualification list, affectionately known as “The Blacklist,” is published on the Internet. I have persuaded several investigators to cooperate with monitors by simply showing them the list.7

September 18, 1997: 812.36—Treatment Use of an Investigational Device. In 1997 came the addition of the treatment use provisions of Part 812. These provisions facilitated the availability of promising new devices to desperately ill patients as early as possible in the device development process and before general marketing. New devices for life-threatening conditions could be made available immediately after, or even during, clinical trials. 

February 2, 1998: 812.43, 812.110, 812.140—Financial Disclosure Provisions. The financial disclosure provisions were added to the regulations in 1998. These provisions tied Part 812 to Part 54 and required that trial sponsors disclose to FDA any equity interest in the device held by investigators, subinvestigators, spouses, or dependent children. At issue, of course, is that basic human tendency to bias our interpretation of events when there is a personal benefit at stake. 

November 23, 1998: 812.35—Supplemental Applications. Extensive additions were made to the clause addressing supplemental applications in 1998, delineating those types of changes to an investigational plan that required prior FDA approval, those types that could be made without prior approval but should be followed by a report within five days, and administrative changes that could be reported in an annual report. The provisions gave sponsors the flexibility to be managed “by exception,” i.e., to obtain prior approval or report promptly those changes to an investigational plan that had significant effects on patient safety or data integrity, but to refrain from cluttering up the reporting stream with minor events. 

Along the Way

Other major events have evolved more slowly, but should be mentioned here. Ethical mores have changed, too. The role of women of reproductive age, children, prisoners, and other vulnerable populations has taken a different place. Today, vulnerable populations are thought to have a right to participate in clinical research as much as others do. While the pregnancy of women must be established at the start of a study, pregnant women are not automatically excluded from clinical research. Prisoners have the right to participate in research on life-threatening illnesses, such as AIDS or hepatitis, for which there is no known treatment.

Regulators have joined together in global efforts to harmonize good clinical practices. ISO 14155, “Clinical Investigations of Medical Devices,” is an evolving global standard that sets the tone for conduct of clinical research in Europe. Data gathered under these standards are accepted by FDA.

The Global Harmonization Task Force, consisting of regulatory and industry representatives from the United States, Europe, and Japan, is establishing a new study group that will look at the requirements for clinical data to support medical device submissions. The goal is to harmonize the need for clinical research itself and to help ensure that all countries require the same quantity and quality of data.

A Slow Start

My corporate career lasted for about 10 years, first working for a big company, and then a small one.8 In 1990, I left corporate life to start Clinical Design Group. I thought the world needed a consultant in the clinical research field. But clinical research was so unimportant in the industry's mind in 1990 that I supported myself as a biocompatibility consultant. It would be 5 years before I won my first contract in clinical research. 

It wasn't only the industry that thought clinical research was unimportant, but the industry's most prominent trade journal as well. It shall remain unnamed, but I am compelled to say that it lost the first two clinical research manuscripts I sent them.

Today, clinical research is unimportant if you don't need to do it, and critically important if you do. Averaging between $100,000 and $1 million for a single trial, no one who needs clinical evidence to support an argument of safety, efficacy, or substantial equivalence can afford to take the subject lightly.


1. Federal Register, 45:13, 3732–3759, January 18, 1980. 
2. “Milestones in U.S. Food and Drug Law History,” [on-line] Available from the Internet:   
3. Federal Register, 179, January 8, 1963.
4. Food, Drug, and Cosmetic Act of October 1976, Section 513 (3)(A).
5. Federal Register 45:13, 3740, January 18, 1980. 
6. Code of Federal Regulations 50.27 (a). 
7. “Disqualified/Restricted/Assurances Lists for Clinical Investigators;” [on-line] Available from the Internet:
8. 3M Co. (St. Paul, MN), then Hollister Inc. (Libertyville, IL).   

Copyright ©2004 Medical Device & Diagnostic Industry

Industry Pioneer

Originally Published MDDI August 2004


James Benson

James Benson, who has served in a number of senior positions at FDA and AdvaMed, made communication between industry and government a mission of his 40-year career in public health. That, plus his role in shaping CDRH as we know it today, makes him an industry pioneer.

Benson began his career in public health in 1962 and joined FDA in 1971. The following year, he was named director of the Division of Training and Medical Applications of FDA's Bureau of Radiological Health. He became deputy director of the bureau in 1976 and then held the same position for CDRH when it was created in 1982. He went on to become FDA's deputy commissioner in 1988 and its acting commissioner in 1989 and 1990. He then returned to CDRH to serve as its director in 1991 and 1992. 

The radiological health bureau's culture provided both regulation and education, and Benson brought that ethos to CDRH upon its creation. “You can't just regulate. It's not that simple,” he says. “You need to use your head to figure out the right strategies needed to approach a problem.” He helped implement “criticism task forces,” which were the first significant forums for industry to provide input to CDRH about the regulatory process. “We decided that we were going to take criticism seriously, look at it, and see if anything needed to be done,” he says. These task forces identified necessary legislative and regulatory changes, that, among other things, provided the framework for 1997's FDA Modernization Act. Upon taking the helm of the new CDRH, Benson did much to restore the confidence of its reviewers after Congress had tightened conflict-of-interest scrutiny and the agency had become extremely cautious about granting approvals. 

He continued to foster industry-government cooperation after he left FDA and joined HIMA, now AdvaMed, as executive vice president for technical and regulatory affairs. Benson retired in 2002 but still consults, serves on the board of the FDA Alumni Association, and chairs a committee that is planning the 100-year anniversary celebration of the Food and Drug Act of 1906, which laid the groundwork for FDA's founding.

Copyright ©2004 Medical Device & Diagnostic Industry