The following blog post is a transcript of a speech Jeff Berkley, PhD (CEO, Chairman & Founder, Mimic Technologies) gave during this year’s American Urological Association Annual Meeting in Orlando, Florida.
“I want to start by making a case for the need for better training. Some of you may be familiar with this report, it’s called ‘To Err is Human,’ produced in 1999. This report estimates that we are not only wasting somewhere between $17-29 billion a year on preventable medical errors, but also somewhere between 44,000 and 98,000 Americans are dying each year because of preventable medical errors.
Now these are big numbers, so it’s kind of hard to get your head around them. This is the equivalent of five jumbo jets crashing every single week. If the general population knew about these numbers, then the public might be a little bit more concerned about surgery and the current level of training undertaken by surgeons.
Now, part of the problem is, how do we know when a surgeon is ready to perform surgery? How do we classify an expert? How do we even classify someone as being proficient?
I thought this study was somewhat illuminating when considering the need for better evaluation. 150 self-proclaimed ‘laparoscopic experts’ were recruited and then tested on a basic surgical skill. They were asked to tie an intracorporeal knot in a box trainer. It was found that 82% of the experts could not tie a knot within 10 minutes. What kind of expert is someone if they can’t do one of the most fundamental laparoscopic tasks? This may be a case where some surgeons think they are the ‘best of all time’ when they are simply ‘legends in their own mind.’ We do have methodologies now (especially with simulation) that allow us to assess a surgeon’s ability to execute basic surgical skills. We can do this now before a surgeon in training operates on a patient.
So you might expect that simulation should be playing a pretty big role in training. Indeed in the 90s and early 2000s, a plethora of surgery simulation companies were created. Here is a list of simulation platforms that were created that have now gone extinct. I think that one of the biggest reasons for this is that when you set engineers to a task and you don’t have the iteration with the surgeons, you don’t always end up with something that fits the need. Rather than engineering putting an end product in front of one or two surgeons asking, ‘Does this look good?’ this should be an iterative process with a lot of surgeons defining and then assessing the simulator throughout development. That’s how you end up with something that actually fits the need, rather than just an engineer deciding that this is something really cool.
So, we really do have a big challenge when it comes to creating simulators with real training value. This slide shows a sign saying ‘Hikers and bikers move to the side of the road when a vehicle approaches.’ This is clearly dangerous instruction due to the crocodiles sitting by the side of the road. This analogy is made to point out that if we instruct somebody poorly because we build a bad simulator, we run a real danger of training somebody to do something incorrectly.
I can tell you that it is extremely difficult to build simulators from an engineering standpoint, especially when it comes to procedure specific modules. Working with soft tissue is one of the hardest things for an engineer to model, period, and we have to do it in real time. We have to update our model at least 30 times a second for smooth animation and up to 1,000 times a second to have haptics, or the sense of touch, involved.
So what we can do right now is model simple bodies within our environment, sort of like what you’re seeing on the screen now for small bowel anastomosis. When we want to bring in all of the organs of the pelvic or abdominal cavity, that’s when we start really running into problems.
I’ll make the argument that we aren’t even really quite there with some of the basic task modeling. I think that we’re getting closer with suturing, but what about dissection? Can we do this realistically?
A lot of people say, ‘Why can’t we leverage more from the gaming industry?’ And the reality is that it is a very different set of problems. There’s not a lot of deformation in video games. There is certainly not a dynamically changing environment where you’re cutting into tissue and you need to realistically display that tissue graphically. Also, there’s a little bit of disparity when it comes to the money. Our industry, as a whole, including manikins, classes, everything together, is valued at just under $800 million. Digital surgery simulation is only worth about $100 million. This compares to $3.2 billion for flight simulation and $68 billion for gaming.
Just to put this in perspective, here is Grand Theft Auto. In its first three hours, it made more money than the entire surgery simulation industry combined. So this gives you an idea of our society’s priorities when it comes to spending on entertainment versus medical education.
So where are we right now? For robotics in general, we start with the physical models. Most of what you see in today’s da Vinci simulators is actually replicas of physical models. This gives us the ability of training in VR while experiencing similar training with the real robot. This way a trainee knows for certain if simulation training translates to better robot control. A lot of the new standards that are coming out, which will soon be used for credentialing and privileging, start with a dry lab component, or a physical model.
You train in simulation because it’s more affordable. Keep in mind that a da Vinci surgery can cost an extra $3,000 per case mostly because of the robot disposables. The instruments are expensive, the OR access is expensive. So you are better off training in simulation and then testing with the physical models.
We also can’t forget the rest of the surgical team. We now have the ability to network somebody working patient side with the surgeon that is sitting at the console. This allows us to practice communication, practice the handoff of sutures, assist with applying clips, etc. Remember, all it takes is one person within the surgical team to make a mistake that can cause catastrophic results during a procedure, so we do need to think about the entire team.
Of course the topic that everyone wants to get to is procedure specific surgery. This is very difficult to simulate. If we are going to build anatomically correct models of patients in VR, it could take us a year or longer just to generate one accurate virtual patient. At today’s level of technology, realism would be questionable in most cases. However, what we can do right now is take an intermediate step and use augmented reality. This utilizes the real thing… images of real surgery into which we can inject simulation content relatively easily. So what you’re seeing on screen is a Partial Nephrectomy that was conducted by Dr. Gill. We have made the video interactive. You reach out and touch anatomy to identify it, you have to predict your tissue retractions, and you have to predict where to cut. The trainee is now actually a part of the surgical process.
I can watch a lot of videos of Tiger Woods playing golf, but unless I actually swing the clubs, I’m just not going to get much better. And that’s what we’re trying to do through augmented reality. We want surgeons to swing the clubs, not just watch the video.
And for those of you who want to know when patient specific simulation is coming, it isn’t too far off. A lot of what we’re talking about today, especially along the lines of imaging, will play a key role in our ability to create accurate patient specific simulators. Patient specific pre-planning and rehearsal is already playing a very key role with image-guided robotic systems like the Mako and the CyberKnife system. So I don’t think it’s going to be long before simulation will train you specifically for a particular patient. Also, more of the guidance tools now only available in training simulators will cross over to surgery itself. The line between what is training and what is surgery will certainly start to blur.”