How Human Factors Can Save Lives | Dr. Joseph Cafazzo on Fixing Healthcare Systems
Dr. Joseph Cafazzo, a biomedical engineer at the University of Toronto, argues that healthcare safety improves through system-level changes like standardization and automation, not culture talk or checklists. Canadian hospitals face a 1-in-18 chance of adverse events partly because each uses different electronic records and medication formats. He points to home hemodialysis and closed-loop insulin pumps as innovations that work because they remove human error from the equation.
Chapters
Just go on to any social media platform, Instagram, Facebook, X, formerly Twitter, and you'll find, and you can find somebody boasting that they've been up for 24 hours. And it's a badge of honor. It's a humblebrag, right? And we're supposed to be science-based, right? And I think we understand the effects of fatigue. On decision-making, on performance, and so on. And other industries do that. I'm sure people have experienced this when they're traveling by plane. All of a sudden there's a delay because the crew has timed out, right? There was delay upon delay, and they have to swap out the crew because they're beyond their hours of service. Happens in the nuclear industry. Like, if you are on a 12-hour shift, which is the maximum you're allowed to, If you accidentally leave the plant and forgot something inside, you can't get back into the plant.
Well, Dr. Cavazzo, thank you so much for coming on.
Pleasure. Thanks for having me.
Do you mind giving a brief introduction about yourself and your expertise? Sure.
I'm a biomedical engineer by training, and I lead biomedical engineering at the University Health Network, and I'm a professor of engineering and computer science at the University of Toronto. Unlike a lot of engineers though, I do have appointments at Rotman Business School. And in fact, my home department is actually the Institute of Health Policy Management and Evaluation, where there's health informatics and quality and safety. And that's really the space that I work in.
And you gave this amazing TED Talk about human factors and system designs, and that's what really sparked my interest in bringing you on and also just having you teach me. Through the master's program at Queen's. I wanted to tell you this for a long time. You said a line that resonated with me so much. I always repeat it. So I'm plagiarizing you a lot that you said in any other industry, if you said we can't change anything because of culture, they would laugh at you outta the room. And I think about that all the time.
Yeah.
'Cause why do you think culture plays such a role and as a hindrance to changing anything in healthcare?
The engineer in me says that it's too qualitative of a concept. It's not something that is substantive. Like, you know, if when you're talking about designing safe systems, you need kind of hard data, things that are measurable, but culture is something that is, you know, for one thing, it's in the eye of the beholder. Um, and you'll never get an agreement on what is the appropriate culture, right? And may perhaps in broad strokes, but culture is not something that can be measured. It's not, not something that can be enforced to a certain extent. And it, it will not, it's not maintainable over time. And so if, if you're trying to improve patient safety, I, I kind of lose my mind when people say, It's cultural. We need to change the culture of the people who are at the front lines. I mean, that to me is an impossible task.
Can you imagine saying that to the aviation industry?
The nuclear industry, any, you know, safety-critical industry, not possible. And I mean, healthcare suffers from this. Because of a lack of regulation. And I know everybody's talking about deregulation, this deregulation, that. And yes, there is regulation in healthcare, but not at the point of care. As one of my former CEOs at UHN once said, we are a farmer's market. Like everybody does things, has their own little cart, and they're doing the things their own way. And you know, you go in a hospital from one floor to the next, to the next, and they, you know, the nurses document differently. They're perhaps trained differently. They're under different leadership and leadership styles. All the outpatient clinics operate somewhat independently. Some of them are booking patients through the EHR. Some of them are booking them on Excel spreadsheets. And at the end of the day, That level of heterogeneity harms patients. And you do not see that in any other industry, any other safety-critical industry is the, that heterogeneity. Now we have strict regulations around pharmaceuticals and medical devices and some critical processes, but not the point of care. That the point of care and, and the professions are all self-regulated. And it's something I'm going to plug this. This is a report that I published about 5 years ago with one of my master's students. And we compared healthcare to all these other safety critical industries. Well, namely aviation and nuclear industries. And the differences were stark. I mean, there was lots of comparables, especially in quality management, but there's no explicit safety management. Uh, in, in hospital care or clinic care. And, and that is what's really harming us because we have no consistency, uh, and no standards, um, on how care's gonna be delivered.
Correct me if I'm wrong, that this, all this heterogeneity in the hospitals, whether they're documenting on paper versus on EMRs, there's no standardization like in any other critical, safe industry.
Mm-hmm. If you look at how the professions and in, and, you know, in hospitals in Canada are also, I mean, they're publicly administered, they're funded from the provincial government, but otherwise they have their own boards and they operate independently. So, you know, here I am on University Avenue in downtown Toronto. And there are 3 hospitals within 100 yards of each other with 3 separate leaderships, right? And, and, uh, boards. And, um, up until recently we had different EHR systems, right? Um, and can you imagine there was a time with the TASSIN hospitals, right? The Toronto Academic, uh, hospitals, they were, you know, with the 7 of them. There was a time when all 7 of them had different EHR systems. Like you couldn't plan that. You couldn't even plan that. And so imagine being a resident and getting rotated throughout the city, right, for your training and having to use a different EHR system each time. Now, I mean, things are better now, but barely, but barely.
How does that affect patients?
Well, look, self-regulation is that one of the things that really bothers me about, about the, especially physician culture, right? And again, if there was better regulation, this wouldn't be allowed. But the culture that it's almost a badge of honor to be worked to exhaustion, right? And I point this out all the time. Just go onto any social media platform, Instagram, Facebook. X, formerly Twitter, and you'll find, and you can search this, you can find somebody boasting that they've been up for 24 hours and it, it's a badge of honor, right? They, they're, it's, it's a humblebrag, right? And, um, I, look, we're supposed to be science-based, right? And, and, and I think we know, understand the effects of fatigue. On decision-making, on performance and so on. And other industries do that. And I'm sure people have experienced this when they're traveling by plane. All of a sudden there's a delay because the crew has timed out, right? There was delay upon delay and they have to swap out the crew because they're beyond their hours of service, right? Happens in the nuclear industry. Like if you are on a 12-hour shift, which is the maximum you're allowed to, If you accidentally leave the plant and forgot something inside, you can't get back into the plant. Your card won't allow you back into the plant. They have rules about you can't do safety critical tasks between the hours of 2:00 AM and 6:00 AM. Why? Because your circadian rhythm is such is that you are a poor performer during those hours. We have no such rules in healthcare. All right. I mean, things are improving. I know that some of the, some of these residents associations have gotten to the point where they've kind of rebelled and said, you know, we're only going to work, I think in Quebec, for instance, it's only 16 hours. And so there's rules around that. But, you know, again, it's there, there in the States, I've heard that They've gone up to as much as 28 hours. And look, you would have to have major reform because you would basically, if you wanted to reduce that to half, you would need to double the residence enrollment. And also these residencies would, would be prolonged because they need patient hours, right? In order to get the experience they can. So I understand it's not a simple change, but this is where if For lack of a better word, the culture of medicine is such that, you know, it, it, you have to pass through this gauntlet of overwork and exhaustion and probably a lot of poor performance and mistakes in order to get through the training of being a physician. And, and, you know, it is better once you become staff, but they do also work long hours and as do nursing.
I love that you said the Like airplane analogy, cuz let's pretend they didn't have limits and you got onto a plane and the pilot was like, I've been up for 28 hours. Are you okay if we go to Cancun? You, you would not get on that plane. But if the surgeon's like, I've been up for 28 hours, I've trained for this, you'd probably be like, are you sure?
Yeah.
But most people just go ahead and follow cuz that's just what we do.
You're absolutely right, Ryan. And, and I know that there's been In the media over the last few months, there's been a lot of high-profile aviation accidents, but it is a total like recency bias reporting every incident now. But if you were to compare the fatality rate within industries, it's no comparison. I think, you know, before the accident in Washington, there hadn't been a major aviation accident in the US for the last 15 years.
How does that compare to healthcare though?
In Canada, it's about a 1 in 18 chance that you're gonna have an adverse medical event in a hospital setting. It's about 1 in 18.
And correct me if I'm wrong, is that underreported, overreported, or about on par where it should be for the statistics?
If you're saying about reporting from a A media perspective, it's terribly underreported. We don't talk about it the same way. And, you know, the common statistics that we've been quoting over the last 20-plus years is that a 747 goes down every day from medical error, right? Or preventable adverse events, as they call it, right? That's been happening for over 20 years. So can you imagine, right? But that never gets reported because it's, you You know, it happens so incrementally and so decentralized, right? And we've just been accepting it. And, you know, we haven't improved all that much over the last 20, 25 years since the first reporting of these statistics. As with respect to your point about how adverse events are reported, you're probably right that there is an underreporting because let's face it, there's a lot of patients who come to hospitals to pass. Uh, and you know, whether they passed as they should have or prematurely because of adverse event, that's always subject to interpretation.
What's the main bulk of all those errors? So 1 in 18 people will suffer a medical event or error. What's the main culprits? Is it human problems, like the human made the mistake, or is it more of a system-focused lens?
Well, I mean, from a human factors, uh, perspective, It's never the human. I mean, that unless it's, you know, unless it's, it's very nefarious and deliberate, right? Like, and, and, and quite frankly, that is extremely rare, right? Extremely rare. And, and we should just put that aside for a moment. We are here to design a system to deliver care. And one of the variables is the human performance, right? The, the training of that person, their expertise. The fatigue that they have, and they are they are but an element within that system. So to me, it's all systems problems that you have to solve when you are building a medical device, and you know in your testing you realize patients are having or patients or providers depending on the device are having difficulty using your particular technology. It's up to you to minimize those errors, right? You cannot tell the FDA, for instance. That you're going to mitigate that risk through training. They will not accept that. You need to exhaust all design possibilities to minimize the error, right? You're responsible for making sure that your device is used properly, right? You can't leave it up to the the rest of the system, meaning the individual at the point of care, be it the patient themselves or the provider, to. make do and be able to interpret your screens and, and the loading of cartridges in, in your, in your device, et cetera, to, to make sure that it's safe to use. It, it's up to you. It's, it's a system design problem. So getting back to, you know, the, the types of adverse events there are, I mean, the common ones are medication-related, underdosing, overdosing. They are transfer of accountability, like who's in charge, failure to rescue. There are all solutions to this, but we're, we, we try to look at, you know, the nurse at the point of care as the solution. And you probably remember from your, your course is I spent quite a bit of time on the hierarchy of effectiveness, right? And at the top, the, the, the, the interventions that are most effective are standardization and forcing functions and automation and all those things. Down at the bottom that are less effective because they are dependent on humans and unfortunately humans are fallible are things like checklists and more training and policies, right? Those are unreliable. And included in that is culture change.
Right. I've never not been able to look at medication labels again since that course, because once you see it, you're just like, these are generally designed poorly. You can't read the labels. There's a lot of extraneous information on it. The information you do want to read and understand is kind of very small. Yeah. It's not readable.
And there, there's the lack of standardization. Like every label is a diff— is, is slightly different. What? What is stopping them from, you know, creating a label where the medication name is exactly in the same position, the dosing is exactly in the same position on the label, et cetera? Like, what's stopping us from doing that? I mean, you don't, you know, pilots don't go into the cockpit and find that the, you know, the instrument panel is arranged differently each time. They don't. And in fact, I heard this recently in the media is that, you know, despite the fact that there are dozens and dozens of companies and lots of different airports, it's incredibly consistent, the experience in airports. And I also, from a business perspective, a lot of airlines have a lot of difficulty differentiating their product, their service to the public. And that's why they're incredibly price sensitive because Really, what difference does it make which airline you're flying? One of 2 aircraft, an Airbus or a Boeing, the experience is pretty much the same, right? And that's because of regulation. It's true that, you know, the airport experience is basically the same, but that's not the case in healthcare.
So this is a genuine question. How often Do you think in healthcare we focus on the end user, whether it's the clinician or the patient actually using the devices?
Well, I mean, look, I think everybody has good intention and I think we are geared toward the patient themselves, right? And I think You know, the professions and, and just to be a little bit more positive now, I do believe that people go into this, into healthcare for all the right reasons. And the messaging that I've heard in senior leadership roles in hospitals is that, you know, the patient is the most important aspect of what we do, right? Their experience and caring for those patients. So I believe that to be true. It's just, I think it's a matter of debate how we achieve that. And I've seen, like, I'm now 30 years into my career and I unfortunately are seeing the frayed edges of the healthcare system just because of the demands are so great and we are having trouble maintaining the quality standards. So, you know, it pains me to see when we go into our emergency department and see, look down on the floor and see the markings where, you know, the stretcher is supposed to be, right? In the hallways. Like we've, we've made that spot in the hallway now a permanent location by painting, you know, markings on there as to where that, that stretcher is supposed to be. And so, you know, I would not have seen that in the '90s, but now, you know, it's becoming commonplace. And, and, you know, we can also talk about the fact that in recent years how access to primary care has plummeted. And so we do have access problems now, the worst I've ever seen in my career.
So from what I'm hearing, it's, There's a lot of system factors behind this that we can try to fix and resolve to overall help patient care. And yeah, because I've, I'll need to be quoted on this, but I think it's like by in 10 years, 5 million Canadians won't have access to family healthcare, which is horrible for outcomes. I'm interested, what got you into this field? Like what sparked all this passion and your impetus to fix it all?
Yeah. In my youth, I did, you know, like a lot of kids, I liked technology. I liked video games and computers and that kind of thing. And so I had an aptitude at math and was more technically inclined. So I went to engineering school. At the time, I didn't know anything about healthcare, and I met one prof who was a biomedical engineer, and I had no idea what that was. And I realized, wow, engineers do have a role in healthcare. And that kind of surprised me. And one of the things, and again, I was in electrical engineering in the '80s when he modeled the human heart and then down to the cellular level as an electric system, you know, and using all these concepts, engineering concepts of electrical engineering for the human body. And I was just kind of blown away. And, uh, you know, by my 4th year, I realized that engineers can contribute to healthcare. I had gone through some stuff, uh, late in my high school years. My mom had passed away from, from cancer. And, uh, I kind of forgot about computers in, in that sense and, and looking at applications. And, you know, I ended up doing biomedical engineering. I mean, university is supposed to expose you to new ideas and, and I couldn't have picked a better school than the University of Toronto to expose me to things I would never have conceived of, right? That engineers could contribute to caring for patients. And again, the personal experiences of, and the trauma of losing a parent pretty young in life. And that kind of set me on my way.
It is quite evident that you've had a lot of personal experiences that I've kind of shown you down this path and you've done incredible work. Have you had any like patient perspectives and situations that helped you try to rectify the problems they're facing?
Yeah.
Help design?
Yeah. I mean, later in my career, in fact, when I was kind of, um, I was already a biomedical engineer at master's level, but I was kind of interested in doing a PhD and, um, I came across this one researcher at Toronto General, Chris Chan. He's a nephrologist. And, you know, this was about 1999, 2000, and he was doing something that I thought was pretty crazy. He was doing something called home hemodialysis. And so for, for those people who don't know, people with end-stage renal disease, with renal failure, have to go on to some sort of renal replacement therapy. Sometimes peritoneal dialysis, or if it's more severe, then they have to go on hemodialysis. And, you know, it, and, and again, here's the other personal connection. My father had just so happened to be going into end-stage renal disease and had been put on hemodialysis. So I knew what the experience was like. So you go into the hospital 3 times a week and you basically have your entire blood volume recycled and filtered. From its toxins and so on using an artificial kidney. The whole process is probably beginning to end 6 or 7 hours, right? 4 hours on the machine and then travel and all that stuff. And you do that 3 times a week to, to get you to about 15% of your renal function, right? So that's kind of like just subsistence, right? Like just, just getting enough to live. And so that's the plight of someone who's in end-stage renal disease. Until maybe if you're eligible to get transplant in a kidney transplant. So what amazed me was that Chris Chan was identifying patients to go on home hemodialysis. And the difference is that there's really no difference in the therapy. It's still an artificial kidney, a dialysis machine, but patients go home with them and their routine is that they dialyze overnight. For a much longer period of time. And they do it nocturnally and they do it almost every single night. And instead of 15%, they're up to 50, 55, 60% of their renal function. And the health benefits of that were quite profound. And I ended up, to answer your first question, I ended up meeting some very young women in their childbearing years. That were miraculously being able to get pregnant and bring a child to term. And that was quite surprising to me because I, I knew full well that most people who are in it, most women who are under end-stage renal disease cannot bring a fetus to term. Um, and, but these women were motivated, right? So young women. For whatever reason, they have end-stage renal disease, whether it's, um, whether it's genetic, whether there's injury, what have you. But they, again, they're, they, they are in renal failure and they have to go on dialysis. And some of them are not eligible for transplant for whatever reason, or don't want transplant because their transplant has other, uh, issues such as immunosuppression. And one person in particular, Yvonne, Yvonne Maffei, Uh, I met and, uh, she was one of the first women to, uh, have a child under home hemodialysis. And she was trained over a 6-week period. Most nurses take a couple years to get their training, uh, in renal, um, you know, renal nursing or, or, and, but, but she was trained over 6 weeks. She learned how to operate the machine. She learned how to self-cannulate. Right. And I don't know if you, I don't know if you've seen the cannula on a dialysis machine. It's not a little needle, right? It's a big needle. Like it's like 12-gauge, right? And so these self-cannula, and it's not just one, it's 2 needles, right? One for taking the blood out of the system and one for returning it. And I was blown away by this. That I couldn't believe. And you know, again, my father is going through this. Look, he was in his 70s by by this time, and I would never have suggested it to him. I I mentioned to him that I was working on this kind of thing, and but I would never have suggested it to him because there's you've got to be a really determined special person. And you know, after Yvonne, there were dozens and dozens of women. And, and patients otherwise who started going on hemodialysis. And now it's anywhere from 25 to 40% of patients are on, on home hemodialysis now. Uh, now 25 years later, Chris Chan was a pioneer in those days, and I was just happy to contribute and ended up doing my PhD thesis in that area.
So what I hear you saying is that just because of innovation, Pushing the system designs, thinking about a patient-centered lens, you're able to design and push the boundaries for people to stay at home, get the treatment they need for half the cost and for a better outcome.
Absolutely. And, and I mean, in human factors, we, we're really preoccupied on our fallibilities, right? But we should also be looking at what humans are capable of doing, and they're, and they're capable of doing incredible things if you give them half the chance, if you create a system that is amenable to doing hemodialysis. And these machines, you know, originally these machines were never built for the home. They weren't built for home hemodialysis at all, or for patient operations. So, and you know, over the last 20 years that has changed. There's more and more devices that have been designed for that purpose.
And I just think about all the patients that I saw as a paramedic with hemodialysis at home and, Just seeing how they operate it. And like you said, it connects to Yvonne that even though she has no medical training, but she's well-versed in her disease process, she was able to operate it just so intuitively.
Yeah.
Because that's how well they were designed.
Yeah. And I mean, I have the story of, and I looked this up, is there was this fellow in China who, um, his dialysis clinic was, was moving or something. Mm-hmm. And, um, And so he would've had to travel quite a ways to get dialysis treatment. So he said, you know, screw it, I'm going to build my own dialysis machine. And he built his own dialysis machine. And this is what I mean is that, you know, there's this level of patient agency when you have a serious chronic condition. It's amazing what patients are capable of doing. And that has been very inspiring to me. And I figured, you know, if we can meet them halfway on something as invasive as dialysis, hemodialysis, we could be doing other things with relating to, you know, other chronic conditions.
Do you have any more examples? Because there's obviously a large list of chronic diseases that are greatly impacting the Canadian healthcare system.
Well, I mean, the other one is I'm, I'm kind of amazed by the type 1 diabetes community and especially when they took They, they, they were increasingly frustrated with the pace of technology, the usability of technology, right? What was, you know, used to be called the artificial pancreas and, or, but, but just basically it's closed-loop insulin delivery, right? You've got, you've got a, you've got a pump, you've got an insulin pump, you've got a continuous blood glucose monitor. There are established algorithms, some of them very simple, others more machine learning based. Why is it that these vendors have not produced a system that, you know, you set it and forget it? The ability to have a closed-loop system. And it was that patient community, a lot of parents of children with type 1 diabetes who really pushed the regulators and pushed, you know, JDRF because, you know, JDRF had a lot to do with this. The juv— what used to be called the Juvenile Diabetes Research Foundation, they've called— they now have changed their names to, uh, I think it's called Breakthrough T1. I, I don't know why they changed their name. I, I think I still refer to them as JDRF, but JDRF for many decades would only fund a cure. And the artificial pancreas and closed-loop insulin delivery wasn't considered a, a cure, so they didn't put any money towards it. But, you know, they got some leadership, uh, about 10 or 15 years ago that decided, no, you know, this is low-hanging fruit. This is— This is technology that could be realized. And, and, you know, again, a combination of the grassroots, a combination of organizations like JDRF, you know, push the companies like, you know, Medtronic and Abbott and Dexcom into creating these technologies. And now, you know, they're becoming fairly readily available.
It's interesting you said that. And this is sheer anecdotes, but it's kind of echoed among the prehospital world that even when I started 12 years ago, we would go to patients having hypoglycemic episodes. We'd give them dextrose, wake them up, and a lot of times leave them at home. But that trend has gone down because of technology like continuous glucose monitors. Yeah. And they're better able to manage that.
And I think one of the personal experiences I had is one of our software developers Unfortunately had his son pass at Queen's. I mean, this is a kid who had been managing his diabetes for probably 12 or 13 years. You know, he's now in second year at Queen's. He's moved to a rooming house and doesn't have too many friends. They don't know his situation and they find him one morning dead. I mean, it's such a tragedy. And I think I hear that story repeated and I hear it from the JDRF leadership. And that's why this push towards CGM and closed-loop systems, right? Unnecessary deaths. And I think the other rationale is that the status quo was, and you could put all kinds of regulation and concerns about the technology, but the status quo was so poor, right? The idea that you were doing point measurements with, you know, capillary blood sugar measurements. I mean, it became obvious as to what needed to happen.
