Human factors engineering has slowly but surely become an ingrained part of how FDA regulates the medical device industry, according to Peter "Pete" Carstensen, but to keep that a reality he says the agency needs to hire more people and reestablish its former human factors-focused team.

If anybody would know, it's Carstensen. He has spent the past 34 years at the device center working almost exclusively on matters of usability engineering and how it impacts a device's safety and effectiveness. He joined FDA just as it was beginning to consider how it might regulate medical devices, and he retired from the agency in July to become a human factors consultant for Wiklund Research & Design.

Shortly before he left FDA, the three members of CDRH's human factors team were transferred from the Office of Communication, Education and Radiation Programs to the Office of Device Evaluation, where they were spread among several review branches.

The move to ODE was the right one, said Carstensen, because it allows reviewers to more closely consider usability issues when deciding whether a product should be cleared or approved. But he thinks the program would work better if a human factors team were centralized within ODE rather than being split up among branches.

"Everything I have learned over the years of people trying to build human factors programs in other organizations is ... as soon as the group gets split up and distributed within the organization, it is the end of the program." Carstensen says he told that to ODE management after the move. "That is just sort of the history of the human factors activities in other organizations. I am afraid that is going to happen."

Carstensen recently responded to questions from "The Gray Sheet" addressing his time at FDA and the evolution of human factors efforts at the device center:

"The Gray Sheet": What were your responsibilities when you originally joined FDA in 1974?

Pete Carstensen: I joined FDA's newly established Bureau of Medical Devices in June 1974 working for Bob Cangelosi, who was appointed head of the Bureau's standards office. Bob and I had come to FDA after spending 10 years working as aerospace systems engineers, where human factors engineering is an important engineering activity. FDA established the Bureau in anticipation of Congress soon passing the Medical Device Amendments. That didn't happen until 1976. Immediately after joining FDA, I was appointed acting branch chief of the mechanical products branch, one of the three medical device standards branches.

Early on, we identified a number of general standards to cover all devices, including one covering human factors engineering, one covering environmental factors such as temperature, humidity, mechanical shock and vibration, and a third covering electromagnetic compatibility.

TGS: Are there important parallels between the aerospace field and device systems engineering that informed your efforts at FDA?

Carstensen: The first aerospace project I worked on was the Apollo Lunar Module - the space vehicle used to descend to the lunar surface and transport the two astronauts back to the orbiting command module. The astronauts' safety and the mission's reliability were imperative. The prestige of the nation demanded nothing less. FDA's mandate from Congress was similar in that patient safety, like astronaut safety, was paramount. In both domains, the application of human factors engineering is essential. How else could you be sure that human beings could operate the equipment safely?

TGS: My sense is that human factors engineering really took hold within the device center in the 1980's with the center's response to high-profile anesthesia machine safety issues involving Puritan Bennett products. Tell me a little more about that.

Carstensen: Yes, you're correct that the anesthesia machine incidents and congressional hearings in September 1984 supercharged our efforts to build a human factors program. But we had first started writing a human factors engineering standard for medical devices soon after I joined FDA in 1974, when I started participating in an [American National Standards Institute] committee that was struggling to finish a performance and safety standard for anesthesia gas machines.

The ANSI committee had been meeting over several years and had almost completed their standard by the time I joined them. I saw that over 80% of the text in the draft standard pertained to human factors engineering and efforts to mitigate dangerous use error. To my astonishment, the committee members had never heard of human factors and were unaware of the body of knowledge and military human factors engineering standards they could draw upon. I convinced them to adopt much of ¹ MIL STD 1472, a comprehensive set of human factors engineering requirements and guidelines for the design of equipment. In 1978, I was also successful in persuading the committee chairman to approach AAMI [the Association for the Advancement of Medical Instrumentation] to establish a human factors engineering committee in their organization.

TGS: When did a human factors team get formed in the device center?

Carstensen: That came about as a direct result of the 1984 congressional hearing on anesthesia gas machine failures. I accompanied John Villforth, the director of the then newly established Center for Devices and Radiological Health, when he testified at that hearing (² Also see "Former CDRH Director Put Problem Solving Before Data Monitoring" - Medtech Insight, 10 July, 2006., p. 7). He promised to build a robust human factors engineering program because the FDA's investigation into the anesthesia gas machines deaths made us aware that at least 2,000 anesthesia patients died each year as a result of human error and the design of the machines very much contributed to these errors.

Had the company that caused the incidents leading to the congressional hearing done a design review and conducted a risk analysis, they would most certainly have discovered the design flaw that brought ruin to their reputation and resulted in them leaving the business. It all could have been avoided. FDA's human factors team was formed soon after the hearings - I think in 1985 or '86. Shortly afterward, I was appointed team leader and remained so until the team was transferred to ODE and dispersed in November 2007.

TGS: Did the anesthesia issue move human factors device design forward in other ways?

Carstensen: The establishment of the Anesthesia Patient Safety Foundation in 1984 focused a great deal of attention on medical errors and the need to apply human factors engineering in the design of medical devices. That resulted partly from the congressional hearings on anesthesia gas machines deaths. But it would have happened even without the hearings. The APSF is one of the great success stories in medicine. Largely through the APSF's efforts, anesthesia mortality was reduced by a factor of 30. That's a 97% reduction, a truly incredible achievement.

TGS: What does it mean for industry that human factors expertise is now housed in ODE?

Carstensen: Moving the human factors program to ODE, which is much more efficient for the human factors staff to work closely with the ODE reviewers, has now entrenched the program for sure and greatly increased the number of reviewer's requests for human factors consultations. That means industry can expect much more engagement with the agency on the pre-market and post-market because ODE technical expertise is so much relied on to support the post-market activities. That will proportionally increase as the human factors staff grows in number.

TGS: I have heard you say that FDA was its own "most formidable and entrenched obstacle" to building an effective human factors engineering program. What do you mean by that?

Carstensen: FDA's human factors engineering program was enthusiastically supported by the CDRH Director John Villforth and his successor, Dr. Bruce Burlington. From the time of Dr. Burlington's departure [in 1999] until the program was transferred to ODE late last year, the program languished somewhat as the center focused on other priorities, just when industry was starting to focus more attention on human factors. Our budgets were limited as was our visibility within the center. This was a frustrating time. There was so much work to do in terms of outreach within FDA and the device industry and in providing input to device reviews and enforcement actions.

The program was certainly re-energized when it moved into ODE, where it enjoys strong support. The gap I left when I retired is being filled with another human factors engineer, and the office is looking to hire at least one additional engineer. The present center director, Dan Schultz, and the director of the Office of Device Evaluation, Donna-Bea Tillman, are especially strong and enthusiastic advocates for the program.

TGS: What still should change about the way FDA handles human factors?

Carstensen: First, FDA urgently needs to hire additional human factors people. With the recent addition of a junior staffer from the Office of Surveillance and Biometrics, the total number of human factors personnel stands at three. It should be at least double that number given the workload and its role in improving patient safety. Second, the human factors team should be reconstituted and remain where the program now resides, within ODE. I think it would be a big mistake to allow the program to remain fragmented as new people are added. It risks losing consistency in how the staff handles human factors issues. In my opinion, FDA doesn't need to spend its limited resources on human factors research as some suggest. It just needs to focus on implementing the solid program it already has in place.

TGS: Would you say human factors design is as entrenched in industry as it is within CDRH?

Carstensen: The industry is coming on board and increasingly investing in building their human factors capabilities. Many companies' human factors programs far exceed FDA's expectations, by my estimation. But unfortunately, there are still lots of companies that have a long way to go. I'd like to believe the industry is more motivated by competition rather than fear of FDA, but if it takes fear, so be it. If the motivation is competition to build a better product to increase market share, then I believe the companies will do a much better job and will be handsomely rewarded. Success breeds success, and everyone wins except the loser who tries to cut corners. The marketplace will remove the loser through the process of natural selection.

TGS: Who, or what segments of industry, has most successfully adopted human factors engineering?

Carstensen: From what I observed, the earliest adopters were the monitoring companies - pulse oximeters, cardiac monitors - and anesthesia machine and blood glucose monitor manufacturers, who responded very positively to an FDA study of their user interfaces. More recently we have seen big improvements in automated external defibrillators, dialysis machines and infusion pumps. It seems evident that once one or more companies in any given medical device segment get on board, their competitors soon follow.

TGS: Infusion pumps have been the subject of multiple enforcement actions due in part to human factors issues. Where do things stand with these troublesome products?

Carstensen: Infusion pumps have been the poster child for poor human factors engineering for years. But that's changing. In the past couple of years we've seen some new pumps with excellent user interface designs. It's clear that industry has stepped up to the plate and applied good human factors engineering in the newer designs. I suspect that was driven in part by competition, with the wiser companies identifying opportunities to increase sales through improved human factors.

One recent incident involving an infusion pump...is significant. The pump employed a keypad to input infusion values, but the keys were designed in such a way that a key touch would sometimes result in the key being activated twice - so an 8 would be seen by the machine as 88 (³ Also see "Cardinal Health Is Second Infusion Pump Maker To Be Hit By An FDA Seizure" - Medtech Insight, 4 September, 2006., p. 3). The operator instructions said to always verify the reading before executing the command. However, busy people often failed to visually confirm what they had entered, and several reports came to FDA notifying the agency of serious patient injury. The FDA didn't accept the company's conclusion that it was the users' fault for failing to comply with the instructions, and the company was made to redesign the keypad to eliminate the problem. That wouldn't have happened 10 years ago. I call that real progress at FDA.

TGS: You are continuing work as a consultant. What keeps you engaged in human factors?

Carstensen: It's a lot of fun and very satisfying to know that your work will improve the lives of so many people. A few years ago, a headhunter tried to entice me to pursue a well paying job with his client, a medical device company. The position was in regulatory affairs. I was blown away by the money being offered but not the least bit interested because I so much enjoyed my work at FDA. I recall telling him how flattered I was by his interest in me and how astounded I was by the generous amount of money offered, but I wouldn't give up my FDA career for all the money in the world. Where else can you have found a job like mine, that's so much fun and so satisfying? I'm a very lucky dude indeed, and very much beholden to FDA for making it possible.

- David Filmore (d.filmore@elsevier.com)

Factoring In Human Factors: An Interview With Peter Carstensen

More from Archive

More from Medtech Insight