Q&A

It is hard to come up with anyone who has had more influence on the acceptance of human factors engineering by the medical device industry than Peter Carstensen, who recently retired as CDRH's human factors team leader. He not only helped CDRH's reviewers assess human factors issues for products seeking approval, but he constantly preached to industry about the value of human factors engineering, and how it correlates with patient safety.

Carstensen joined FDA in 1974 and rose to the rank of senior systems engineer. He drove the development of a number of human factors standards and guidance documents. He has served for years as convener of the International Electrotechnical Commission's human factors engineering committee. Its most recent work is ISO/IEC 62366:2008, “Medical Devices – Application of Usability Engineering to Medical Devices.” He has also served on the Association for the Advancement of Medical Instrumentation's human factors engineering committee that is presently developing an encyclopedic set of user interface design guidelines for medical devices (ANSI/AAMI HE75). In 2004, he was named one of MD&DI's 100 Notable People in the Medical Device Industry.

Carstensen will be very active in retirement, though. He will continue his work with the IEC and AAMI committees, he gave the keynote speech at the Human

Factors and Ergonomic Society's annual meeting in September, and he has joined Wiklund Research & Design (Concord, MA) as a senior analyst. In a consulting role, he will still be able to trumpet the value of human factors engineering to the medical device development process, just in a different way. He spoke to MD&DI Editor-in-Chief Erik Swain in August.

Q: You recently concluded a more than three-decade tenure with FDA. What were the most significant human-factors-related advancements you saw during that time?

A: The AAMI human factors engineering (HFE) standards HE48 and particularly HE74, the design process standard. AAMI agreed to expedite the development of HE74, “Human Factors Design Process for Medical Devices,” at FDA's request. We had just finished a survey of the industry where we learned that at least half, if not the great majority, of device manufacturers didn't have a clue what they needed to do to address the needs of their users. We urgently needed to educate them, and AAMI seemed to the best route with their HFE design process standard and their educational seminars. We at FDA learned a lot from our interaction with the AAMI committee and, dare I say it, they learned some things from us. In particular, about the role of human factors (HF) in meeting FDA's Quality Systems Regulation (QSR), and our viewpoint on what a manufacturer needed to do. The committee did an outstanding job of tailoring the document to meet the needs of the readers – namely the design engineers who were going have to take the lead in designing and implementing an HFE program for their companies. One of the committee members, Mike Wiklund, conducted a survey of the industry to determine what the engineers most wanted to see in the document. In other words, the committee practiced what it was preaching. First determine your users' needs! By the way, that standard served as the basis for IEC 60601-1-6, “Medical Electrical Equipment -- Part 1-6: General Requirements for Basic Safety and Essential Performance -- Collateral Standard: Usability.” That was the predecessor of ISO/IEC 62366.

Later, my colleague, Ron Kaye joined us and one of his first tasks was to write comprehensive and understandable guidance for industry on managing use error risk. This guidance describes how a manufacturer should go about identifying, prioritizing and assessing use error risk and how they should address it. The guidance is compatible with HE74 but is more specific to FDA's needs. It's something every manufacturer should be aware of and use to help them meet FDA's requirements. As I recall, both HE74 and the FDA guidance were published within a year of each other. They both were and still are very important tools in helping advancing our HF program.

Another important step was a guidance issued in July 2000, “Medical Device Use - Safety: Incorporating Human Factors Engineering into Risk Management.”

Last but not least, is the move to ODE (the Office of Device Evaluation). The entire HF group was transferred to ODE late last year. This puts us right in there with the people doing the pre-market reviews and allows us to better and more efficiently interact to provide HF consults. The ODE office director, Donna-Bea Tillman, is very supportive of the program. In fact, I think I'd describe her support as enthusiastic and committed. The only problem is her incredibly heavy workload doesn't permit her to devote as much time to helping advance the program as we'd like. ODE's first priority is to move the freight and meet MDUFMA's review times. They have an incredible workload made tougher by the demands placed on them to provide their technical assistance to the postmarket program and with little if any prospects for an increase in budget to hire more people. Nevertheless, the ODE director has made a commitment to replace me and hire at least two more if possible.

Q: What are some of the most common mistakes device firms make regarding human factors and usability design?

A: Besides doing nearly nothing, I'd have to say doing a very superficial job which fails to addresses FDA's main concern, namely use error that threatens patient safety. Too often they submit reports that describe usability tests that mainly address aspects other than risk – for example, user satisfaction, or speed and accuracy in completing tasks that have little impact on safety. When they do manage to touch on risk, they often fail to prioritize the risks. Every risk seems to be equally weighted. But if you look at the ISO 14971 or any other risk management scheme, risk needs to be quantified, or at least ranked in importance. That's often a tough job, but that's no excuse for throwing up your hands and making no effort in that regard. FDA published guidance on what we expect manufacturers to do to address use error. It's not like it's a big mystery. We make it pretty clear what we expect and provide the name of a contact person, Ron Kaye, who, as busy as he is, will be more than happy to clarify any points in that guidance.

Q: Why should a medical device company have a human factors program in place?

A: Because it's the law. It's the only way to satisfy the requirement in the design controls provisions of the QSR to address the needs of the users. A device that doesn't seduce the operator into making a mistake that ends up killing or seriously injuring the patient is a basic need of the user. If you don't think so, talk to a clinical person who made such a mistake. It's devastating to them and the patient and those patients' loved ones.

Besides, they should do it for their self interest. Establishing a HF program will result in better products, increased sales and customer loyalty, and increased bottom-line profit.

Q: How should a medical device company structure its human factors program?

A: There's no easy one-size-fits-all answer to that question. But the good news is, there is some good advice in the AAMI HE74 design process standard. Also, there are a good number of HFE consultants out there who can guide a company as it develops a HF program that meets its needs and situation, and do some or all the work for them. If companies take advantage of the expertise available in the marketplace, they will not only find it easy to comply with FDA requirements and EU requirements, they will also help their bottom-line profits. HFE pays off by a factor of more than 3:1 and often as much as 10:1. Every dollar spent wisely yields a $3-10 increase in bottom-line profit. Do a search at www.devicelink.com/mddi and you'll find articles testifying to this pleasant fact. It's what's commonly called a no-brainer. I just don't understand why companies don't get it. But I'm confident in time they will, or they will be eliminated by their competitors who do.

Q: How should human factors work be integrated with design controls?

A: First of all, HF should be fully integrated into the overall design process. It shouldn't be an afterthought that gets addressed late in the design cycle. HF people should be engaged at the very earliest stages of the design process. They need to be engaged in defining the users and their needs and specifying the design inputs. That may involve assessing existing devices and seeing how well they work in the real world to learn where things can be improved. It involves doing a preliminary task analysis and redoing it as the design progresses. And it certainly involves identifying risks and assessing them and finding ways to mitigate or eliminate them, starting at the earliest stages and continuously refining that risk assessment as the design evolves. There is nothing more expensive than to learn near the end that you have a problem. It's expensive but not as expensive as to learn about the problem after the product has been introduced into the market. Recalls are very, very expensive.

Q: While progress is being made, the medical device industry lags behind industries such as aerospace and automotive in its use of human factors and ergonomics research. At what point do you think the device industry will get to that level?

A: I don't think the medical device industry will ever perform at the level of the aerospace industry, and I don't think it should shoot for that high standard. There's a compelling reason aerospace spends the capital it does on HF. They are in the business of building very costly vehicles. They can't afford to lose too many, and besides, the whole world is watching when it happens.

I disagree that the medical device industry is behind the auto industry. More often these days, the auto industry simply puts marketing considerations before safety-critical user needs. That is not tolerated in the medical device industry.

Autos are pretty simple things compared to some medical devices. Almost anyone can drive a car, even the most complex ones. But if you look at sophisticated medical equipment, the user interfaces are very complex and challenging. The auto industry has a much easier job than companies putting together a robotic surgical system, for example.

Q: In recent years, CDRH has gotten more aggressive about sanctioning companies whose products suffer from poor or nonexistent human factors work. What prompted that stance? Has it gotten more medical device firms on the human factors bandwagon?

A: It's gotten more aggressive because our compliance people and senior CDRH managers have finally seen the light. It's a scandal that it didn't happen a lot sooner. [Former CDRH director] John Villforth testified in a Congressional Oversight Committee hearing in 1984, telling then Rep. Al Gore that medical device use error was a major concern that the center intended to address in a major way. That's when our CDRH HFE program really started to move forward. Villforth was really sold on the value of HF. He put a lot of emphasis on educating industry about use problems.

Now, it was a slow process getting other senior personnel on board. They all bought in to the concept, but they had a lot to do; it was a small piece in their world. We continued to have to fight to get the concept within our regulatory authority. The rule had been that if a company wanted to address a danger by warning about it in a manual, it could. We had to fight to eliminate that, saying it was not enough. Eventually [senior managers at CDRH] just came around. It just seemed to happen. There was no sort of watershed and no external political pressure. You may remember the Alaris key-bounce problem [input keys on an infusion pump had the potential to stick, raising the risk of overinfusion.] Ten years before, we would not have been able to get them to address that problem, because they mentioned it in the manual. So we've made some progress in that regard.

Sometime after Villforth retired and Bruce Burlington left, the program started to lose momentum. We went from seven staff devoted to HF to three in about five years. With my retirement, CDRH now has two people working their tails off trying to keep the program afloat. I have great faith that they will prevail against all odds – and I sincerely mean that. These two guys are absolutely dedicated to the HF program and they have strong, committed support from the ODE office director and the center director. The satisfaction these guys get from knowing what they do actually saves a lot of lives keeps them going. It certainly did for me, and it's what motivates me to continue to do so into retirement.

Q: What helped the device industry get on the human factors bandwagon?

A: A lot of things came in to play. Part of it was that word got out that FDA was paying more attention to HF issues, and you could get burned if you ignored them. A lot of it had to do with the various standards that came out. Those took a number of years to develop. Industry is not stupid. They see what's coming down the track. And they saw the horrible trouble that some companies were getting into as a cause of human factors problems. Industry came around faster than FDA management, to tell you the truth.

Q: What kind of impact do you expect ISO/IEC 62366:2008 to have on the medical device industry?

A: There is no doubt in my mind that it will have a very positive effect. Like it or not, it's something no manufacturer can afford to ignore. The product liability risk that they would assume by ignoring it is overwhelming. Also, I understand the EU will recognize the standard, as they have already recognized its predecessor, IEC 60601-1-6.

IEC 62366 has nearly all of ANSI/AAMI HE74 included as an informative annex. The only thing missing is an annex in HE74 that describes the relationship between HE74 and the FDA Quality Systems Regulation. I believe IEC 62366 will soon replace ANSI/AAMI HE74 as an FDA-recognized standard.

Q: What kind of impact do you expect ANSI/AAMI HE 75 to have on the medical device industry?

A: HE75 is a very comprehensive handbook describing almost everything a designer needs to know. It's a one-stop shopping text with most all the information a designer would need to design a good user interface and validate it. But it still requires intelligent interpretation. It's like someone could write a detailed text on how to perform brain surgery, but careful study and practice will be needed to pull it off. HE75 is a very good start but it's not a substitute for expertise in the field. There is an even more comprehensive text based on HE75 which should be published before the end of the year. It's around 700 pages and will sell for the remarkably cheap price of around $125. Given the way the dollar has been wasting away, it's a hell of a bargain.

Q: These days, medical devices often have to be integrated with IT systems. How should human factors research for those systems be approached?

A: It is not magic. The proven HE research and evaluation tools commonly used to optimize the user interface design of anything can be applied to medical devices integrated with IT systems. Sure, the task of integrating medical devices with IT systems will be technically challenging, but there are a lot of smart people out there working on the problems. I see great opportunity here with the integration of medical devices with IT systems to really advance medical device technology and greatly improve efficiency and patient safety. It's exciting times and HF will play a major role in the process.

Copyright ©2008 Medical Device & Diagnostic Industry