Click here to get 10x Brochure

How To Use Human Factors To Mitigate User Risk

MaryBeth Privitera is impressive.

She co-chairs the AAMI human engineering committee, is the Human Factors and Research Principal for HS Design, and teaches biomedical engineering at University of Cincinnati.

Naturally, she gave an excellent talk at 10x in October, here for your education.

Managing the Big Risk – The User
presented October 11, 2018 by MaryBeth Privitera, PhD FIDSA



PDF Click here for a copy of the slides.
PDF Click here to download the transcript.

MaryBeth Privitera: I like to take pictures of things. I observe people for a living, and I observe the surroundings I live in.

These are just some of the funny pictures I take along my travels.
Funny Pictures on MaryBeth Privitera Travels
Things are odd, but you have to look at them. We’re so busy going through life that, sometimes, you need to stop and just take a look around.

What I’m here to talk to you is about managing the big risk. And that is the user.

I work at HS Design. It is a certified design firm. I teach at the university. I like both worlds. Both worlds give me insights into a whole host of things, and given that I like to observe the world around me, that’s pretty prime.

About HS Design

HS Design is 40 years old. We do full service; we have engineers, designers that work there. I just wanted to give you the context of my world.

Overview

I’m here to talk about:

    • the role of the user in risk management;
    • how to uncover the users’ abilities, their actions and attitudes in the design process; and then,
    • the application of human factors.

One of my other jobs is I co-chair the AAMI human engineering committee. Human factors and usability are near and dear to my heart. I’m an industrial designer by training, even though I teach biomedical engineering.

Role of the User in Risk Management

First and foremost, if you don’t have any use risk, you don’t have any problems.

The Lowly Scalpel

Think about a plain scalpel. Would a scalpel get approved today with our current methods of risk analysis?

No, it would never get approved because you could never get rid of all the associated risks.

Users are unpredictable, and that is the point. You don’t know how a user will behave, but you can predict it.

When we do our risk and our use-risk analyses, we consider all the various ways a user could behave, but we’re very creative as individuals.

The best example of our creativity is if I were to open up everyone’s cell phone: the apps you selected, the organization … we like to customize; we like to do things on our own. That is true in every person’s being in their way of life.

This is a really good question – If a user performs the task incorrectly or fails to perform a task, could it cause serious harm?

That’s really what the agency is looking for – Could it cause harm?

Definition of Harm – CDRH vs CDER

When I look at the discussions between CDRH and CDER – for those that are in the pharma versus the device world – their definition of harms varies.

So the definition of harm according to CDER is, “Yes. You know what, a little bit is okay.”

And that’s why the scalpel stays on the market – there’s no possible way I can do surgery without doing some level of harm.

Whereas the pharma world says, “Nope. No harm. No harm at all.”

So when we get into the combination device, it gets increasingly complex, which means that you must pay attention to our use-base risk.

Know Your Users: Apply Human Factors

So what does that mean? The first and fundamental thing about assessing the user-end risk is that I have to know my user in and out.
Know Your Users - Apply Human Factors
If I don’t know my user, I can’t actually do an adequate risk analysis. It’s just fundamental, and that means everyone.

So good design is going to start with knowledge of the user – What are their capabilities, their education? What’s their training? What’s their bias? How long have they been there? Is there a difference between somebody that has been working in the hospital for six months versus 10 years?

Yes, of course, there is. What are they, and how does that information impact the design? How can I bring in some of the disciplinary bias?

So let me give you an example of what I mean by disciplinary bias. And we all have them.

If you were to spend a day with an anesthesiologist, you’d notice they have an entire box that they will mix and, then, they will administer their medications. If I give them an infusion pump that automates it, they don’t like it. They like to mix it. That’s a fundamental tenet of who they are.

So they (disciplinary bias) exist in everything, and it’s about control because we like control.

Hence, that’s why all of our cell phones are different. So you get to the element of creativity; you get to the element of how we go ahead and execute our life, how we execute our daily life and our daily practice. And it’s true for all of our users.

Why Apply Human Factors?

I have these two pictures from my academic world, and I think they are two really telling pictures.
Why Apply Human Factors?
And the reason is, someday, God forbid, we may encounter the fact that we could be that person or a loved one could be the person in that bed, in the ICU over there.

And if you look at the number of screens that are available and the amount of information that’s coming to that critical-care physician team – which is the picture on the left that is a group of critical care physicians, residents, nurses, pharmacists on rounds – the amount of information going to these people is overwhelming. and they’re going to make the decisions that impact their life.

So the way we need to think about human factors is really just making it easier for them to make the right decisions; harder for them to make the wrong decisions; and making sure that, when we consider the patient at the end of the day, they’re doing the right work for their patients.

I’d like for them to not really look at a screen but, actually, look at the patient. Those are my thoughts.

It’s true throughout the hospital. This is an example of the amount of screens that you would find in anesthesia, and you can find complexities everywhere.
All Devices For Anesthesia
If you look at the practice of medicine, it’s abundant. Even when we bring devices to the home care, they’re also complex. We’re not following the IKEA model of instructions for use. You’d find them to be pretty elaborate.

But this is just an example of some of the world in which we have developed.

One of the challenges that we have as device developers is, we look at our own device in isolation. We don’t look at our devices as part of an overall system, which is where we need to take a look at it from – a very specific “Here’s my user, but here’s the context with which it’s used in.”

So when we design in context, it becomes a little bit easier, a little bit better for us to consider how that user is going to make that clinical decision.

And we know, from a business perspective, if it doesn’t lead to clinical decision making, we don’t actually have a product to sell. So we have to be involved in that clinical decision and on the onset.

Uncover User Abilities, Actions & Attitudes

Someone’s going to prescribe our device; someone’s going to use our device; someone may be the recipient of that device. So how do we uncover our user’s abilities, attitudes, actions?

When we talk about considering capabilities and limitations, we’re talking really about –

    • how big they are,
    • their age,
    • their dexterity,
    • training and education,
    • their experience,
    • and their culture.

When I talk to physicians, they could be trained in a specific organization, and that is why they do what they do. It’s based on that training. They’ve always done it that way. If it’s not broke, don’t fix it.

And then, with more and more devices coming into the home healthcare, age becomes something we must consider. It could be a child that uses the device.

It could be something as simple as “how does the color go?” Color and design go beyond brand guidelines and preferences. We have to consider – are they going to be colorblind? How do we distinguish elements for one thing or another?

And that means we have to get the user feedback and asking them questions and how do they respond to the designs we put forward with them.

Physical and Sensory Characteristics

So physical and sensory characteristics include vision, hearing, manual dexterity, strength, and reach.

If you’ve ever designed a hand tool, one of the things with hand tools is “how do I actually reach all of the controls I need to reach and in what order?”

I had the fortunate experience of working for Ethicon. It was Ethicon Endo-Surgery back in the early 1990s. We did pistol grips, and it was pointed out that pistol grips for laparoscopic surgery were really ergonomically incorrect.

So, ergonomically incorrect meaning, I am like this when I do my cases.
Image Showing the Ergonomically Incorrect Pistol Grip
And the edict was, “Why don’t you just lower the table then, doctor?”

Well, they’re used to it. Now they’re accustomed to holding it up.

If I went into laparoscopy today, I can demonstrate very easily that, biomechanically, this is the worst position that I want my physician to stay in for 45 minutes. But it doesn’t matter because that’s how they’re trained.

And if I showed them something that wasn’t a pistol grip, they probably would balk at it, and I would negatively change it.

So there’s very much to learning manual dexterity and strength.

When I go like this, the reason that’s so bad is, every time I have a joint deviation, I have negated my ability to provide the strength in pulling that lever.

So how does that translate into the real world? Let’s just say I’m back at Ethicon, and I’m designing a linear stapler. And in order to get the linear stapler to fire two lines of staples and cut in the middle, I need to produce 75 pounds of force coming on this lever.

So, all the engineers in the room can think about the math. It’s inputs to outputs, when I think about it from an engineering perspective.

Well, I’m not physically able to do that if I’m a 50th percentile female. The reality is most people going to medical school are becoming increasingly women. So it was designed for men, but now it’s for women.

So that’s when gender size, all of the biomechanics come into play to impact the overall usability of a design. So it’s important to pay attention to that.

I find that element of design to be the easiest. There’s a great standard. It’s called AAMI HE75. It has all of that in there. Those are the easy wins.

What’s more difficult are knowledge-based tasks, cognitive task, where we are having to remember something.

And that gets to the legibility and the discrimination – How do I hear an alarm? Can I hear an alarm over the top of what else is going on in the room? And I just demonstrated for you all of the complexities of the devices that are in the room. So how does one alarm sound versus another alarm? Can I provide more cues to those alarms? Can there be a visual with that auditory signal?

Perceptual and cognitive ability

And that gets to this – perceptual and cognitive ability – our ability to detect and identify, recognize sensory input. Do I know that I have done something? What’s my confirmation?

Ever had one of those things where you hit the computer, you hit the button, and then you wait and you don’t see anything? And you’re like, “Did I do it?”

Well, what we do now is we do it again, right?

We are working, right now, on a device where we’re just able to change training materials. It is a great example. It’s 2006 technology. And when the user hits the button, it starts to blink.

What do we do when we see something blinking? We don’t like that. It’s something that says I need to do something so it will stop blinking. So I hit it again.

And whenever you hit it again, it stops the device from working. So you get a use error because it’s blinking.

So those cues we use to navigate our way through life are super important. And they tell us I’ve done something. They confirm back to us. It’s important to pay attention to it.

Expectancies

These are the established habits; the conflict that habits lead to use error.

So if I know I have a habit of doing something, then I know that I should maintain it.

Imagine this world, you have free rein to design a car in any way you want to. It’s going to have wheels and you’re going to have to go.

Well, imagine what would happen if I put the gas pedal on the left and the brake on the right.

I have a use expectation. Globally, I get into a car. Gas is on the right; brake is on the left. It’s just something that we’re going to do. You’re not going to change up because you know that is an expectancy.

Well, as physicians, as nurses, as parents, we all have an expectancy. You have an expectancy for what you expect out of the conference, out of your life. And when things don’t meet our expectations, we go, “Huh. What’s that all about?”

And then we either like it or we don’t like it. But when we think about it from a medical device and agency perspective, are we going to cause harm? Is there going to be an error that happens as a result of it?

And that’s because we have mental models of how our world is supposed to be. Those mental models are abstract concepts. And it’s through an assessment of how users go through their lives that you can figure out what their mental model is.

And that’s important in design because if I can get with them to conceptualize the operation of a new device that may have not been on the market, then I can walk through the move of what their expectation is.

So gone are the days we could sit there as engineers, as designers, and think that we are the user of the device. “Well, I can do it, therefore they can do it.”

Well, all of my patients are dead because they were either dead to start off with or they were made of plastic. I will never do surgery. So I need to get THEIR perspective because it is completely different than mine, which gets to culture and diversity.

Every one of the products that we have to work on is also global. Gone are the days when we only designed for the United States. We might say that we’re designing for the US, but the intention is for global market.

So that gets into even more of “what are the mental models when we take something and put it into a diverse world?”

Applying HF to Mitigate Use Risk

So the great thing is, when it comes to mitigating use risk in terms of human factors, there’s some fantastic guidance out there. I’ll just point you to a couple of them.

You don’t have a choice in this. I’m sorry about that, but you don’t. And that’s our regulatory agency.

This is from the FDA.
Applying HF to Mitigate Use Risk
It came out in 2016. It was first out in draft in 2012, so they didn’t draft for a long time. This is probably leading in the United States that they are expecting a Human Factors dossier.

In this document, which is IEC-62366, they are written in concert with one another. So they will be alike between the two of them. If you meet this, you may not meet this.

The difference between these two documents state that, in the United States, you have to use United States citizens. That is the largest difference between the two documents.

Similarities between the documents:

Both documents state you have to have user testing throughout your design process.

They also state that, at the end of it, you have to validate your users through a summit of usability test. And they give you the instructions for what they expect in terms of a dossier that gets submitted. And they give you the report format.

When you don’t follow the report format, they don’t like it. And they’ll tell you they don’t like it, and they’ll ship it back to you.

This does not give you a report format to follow. This will tell you that you need 15 users, but it is not as strong of language as this. That makes sense because this came out after 62366 did.

And then historically, this document here, which is from MHRA in the UK, talks specifically about the usability in combination products.

And the difference between this document and these other two is a greater emphasis on instructions for use.

And that means that part of the user interface, as defined by the regulatory agencies, are the overall design, the graphic user input, the instructions that are given to the user, everything that would be available to them upon market release.

So if training is part of your risk mitigation, then training is part of your user interface. And all of that has to be validated.

The human engineering committee helps support this work here. There are a bunch of technical information reports on Human Factors engineering that are available to you, and that gives you a good methodology and best practices. So I highly recommend checking out that website to understand the background.

And if you are not that familiar with standards organizations and how they work, the committee that I’m co-chair of, the Human Factors engineering committee here, puts the US position here.

FDA is part of that committee. So FDA serves with us on that committee to make sure that everything is communicated back and forth.

That doesn’t mean that we always agree with the FDA. In fact, sometimes we don’t, and we will argue in committee meetings with the FDA because, at that point in time, you’re at a level playing field. It’s the industry and FDA working in collaboration versus one organization coming up to the FDA.

So, a little bit of background there, you don’t have a choice to not do human factors. But it’s important to think about it in terms of “who my user is?” “what are those attributes, those physical attributes, the characteristic sensory?” “how do they think in the context of use?”

Case Example – Vagus Nerve Stimulator

And so I’m going to share with you a case example of some of our work which is in consumer health.
Vagus Nerve Stimulator - Map of Dose Refill Scenario
As a consumer health device, it’s a vagus nerve stimulator. I heard, yesterday, one of my students passed out. That’s a vagus nerve reaction. Your vagus nerve has all kinds of really great things. It’s also the nerve that runs all along your spine. It comes up to your neck.

And this particular one, this vagus nerve stimulator, it’s just new out on the market. The flow, the business model, the user experience, strategy is that the physician writes a prescription; the pharmacy provides it, then the patient uses the device; they need a refill.

It is an electronic device, so there are no drugs that go into it. It just sends an electrical stimulation to that vagus nerve to relieve headaches.
Possible Use Scenarios of Vagus Nerve Stimulator
What we did in this particular design exercise, we started off with working with different approaches to the whole product embodiment and the workflow of what that user interface is going to look like, defining all the possibilities of how it could be configured, identifying the risk areas, the digital and physical use interactions.

And then we started looking at how am I going to hold it, “what’s that user interface going to be?”
Ergonomics Forms of Vagus Nerve Stimulator
So this is really a great simple little merge of both physical buttons and digital user interface combined together.

So what are all of the variations? Obviously, you can get to quite a bit of permutation. What is the overall screen? Where are the buttons located? Looking at combinations of both touchscreen etc.

So, from here, we built preliminary phone models, and the users were able to take these study models and to match up what kind of graphic display and what kind of form is going to work for them.

We believe in a participatory design process where we’re inviting the users to give us direct feedback within our design process so that we’re not designing in solo. We’re actually arming the users to talk with us back in a way that we can communicate back and forth.

And that’s something that, academically, my students struggle with. They’ll go ahead and vote on designs that don’t exist or have a conversation about designs.

And we’ve all done that; where we’ve said, “Hey Doc or Hey So-and-So, wouldn’t it be great if we had a design to do x, y, and z,” and you have a conversation. You think you’re understanding each other but, in reality, you have no idea.

So, if I had a car that made me breakfast in the morning, that could help me put on my makeup, you have no idea what that really looks like in my mind. Nor would I have an idea of what it looks like in your mind.

You could go ahead and brainstorm on it. But at the end of the day, we may not even agree. We may not even have the same idea in our minds.

And so, we have to arm our users with the ability to communicate to us how they’re going to perceive a particular device.

Case Example – Vagus Nerve Stimulator – Usability Testing and User Insight

From there, we did more formal usability testing.

So this gets back into jumping over from just doing the design to “how do I get the users to really walk through that use scenario, walk through and identify the overall challenges they may have?”

And what we discovered here was pretty interesting. We discovered that users may be sensitive to bright screen.

So where we thought, “Put a big giant screen on here because I’m old and I can’t see. That would help me out,” actually was counter-intuitive because if you have a headache, you don’t want to see a bright screen.

And depending upon the use environment, if you’ve been out in the sunshine, when you look at a bright screen, sometimes you can or you may or may not see. So context is important.

And so, we were able to pivot the design direction based on the use feedback to give much more of something that’s more realistic and appropriate for this particular market, which is not to have a screen at all.

Anyways, it’s embedded underneath of it using lights. And you can see where it just became a much cleaner, simpler design than the preliminary designs we had.
Preliminary Designs of Vagus Nerve Stimulator
So, at the end of the day, this consumer product is out on the market. They’ll end up using all the charges. This just tells them the number of chargies they have, therapies they have remaining. And then this gives that nerve signal to it.
Vagus Nerve Stimulator - Consumer Health Device
It’s very successful. And they were able to design the packaging as well to just round out the whole use experience so that we were protecting from inadvertent damage on the electrodes.

Conclusion

So, in conclusion, when you involve human factors methods in and throughout your design process, not only are you meeting your agency requirements, but it also satisfies your customer

In consumer products, we want them to have a positive user experience. Well, why not have that in a medical device?

And it doesn’t matter whether it’s a surgical tool or it’s a consumer product. At the end of the day, the goal is to improve life.

I think the FDA wants us to stay consistent. They want us to reduce risks and improve the lives of all the users. So it just fundamentally makes sense. It’s not that hard to integrate.

Joe Hage: I think my lesson learned is, if the speaker is a teacher, you’re probably pretty safe with their presentation. So, thank you.

MaryBeth Privitera: I don’t know [laughs]. Because if you give them a marker, they may go on for a really long time.

Joe Hage: No. Not here. I would totally–

MaryBeth Privitera: Yes, you would gong them. Gong.

Am I getting gonged? I think I’m getting gonged.

Done.

Joe Hage: What percent of medical devices, either companies or products, would you say were designed without a solid understanding of human factor?

MaryBeth Privitera: The entire world of catheter interventions.

This is the craziest thing in the world. It’s like asking you, Joe, to please take your garden hose and water the flowers next to your swing set while standing on your front porch.

Let me give you a few tunnels to do it. So you figure out your tunnels. You have a map of the tunnels, but you have your garden hose, right? Because it’s a skinny little catheter, I can only push it, pull it, twist it. That’s all that I have.

I want you to pick a flower out of your flower bed that’s in your garden and, then, bring it on back around for me.

Joe Hage: Let me bring you back from the brink here.

[Laughter]

How about among products that don’t have a 1975 predicate?

MaryBeth Privitera: Okay. Newly designed?

Joe Hage: Something after the regulations.

MaryBeth Privitera: Ah… Oh golly.

Joe Hage: 10%, a quarter, half, a third.

MaryBeth Privitera: I would say we’re not too terrible. That have been designed, I would say at least half of them.

And it really depends on the area that you’re looking at. Because if you’re looking at devices that are in the hospital versus home healthcare, home healthcare is getting a little bit more attention.

Joe Hage: MB Privitera, for you, HS Design. Thank you.

MaryBeth Privitera: You’re welcome.