Parvati Dev, CEO of SimTabs: When to Use Virtual Simulation in Healthcare Training

Van Ton-Quinlivan: Welcome to WorkforceRx with Futuro Health, where future-focused leaders in education, workforce development and healthcare explore new innovations and approaches.  I’m your host, Van Ton-Quinlivan, CEO of Futuro Health.

The use of simulation in healthcare training used to be confined to actors posing as patients and the use of medical mannequins. But the options have grown far beyond that to include 3D virtual reality, augmented reality, and game-based learning, amongst many other approaches.

Today, we’re going to get an overview of the space with one of its leading experts, Dr. Parvati Dev, CEO of SimTabs. Dr. Dev has four decades of experience developing tech solutions for life sciences education in industry and academia, including groundbreaking work at Stanford University. Today, her Health TeamSpaces provide opportunities for nurses, doctors, pharmacists, and therapists to learn together.

Thanks so much for joining us today, Parvati.

Parvati Dev: Thank you. It’s a pleasure. Looking forward to this.

Van: Well, how did you get interested in simulation and e-learning? And do give us some highlights of your work along the way.

Parvati: Sure. I’ve always been interested in mathematical modeling, which is a way of representing how things work. As a teenager, I read an article that completely affected me. It was an article about bee’s eyes and how the bee saw through multiple lenses and maybe we could make new cameras that had completely different ways of looking at things. That concept — first of all, that you could understand how the natural world worked, beyond just the fact that it worked, and that you could relate how you made things work in your own world to the natural world — was a driving force for me in many, many different ways all through my life.

Simulation is that you’re representing some situation, some object, a hospital, a patient in a way that you’re making it work in the computer and then you allow people to make changes, affect it in some ways and see how the “system” reacts…what happens to the patient, what happens in the hospital? And then maybe you learn from that, you try again something different. So, that concept of trying to figure out how something works represented is what’s driven me all the way.

Then I went into 3D medical imaging. I spent eight years developing 3D from CT scans. It was the very early days of that. I think we were real innovators bringing out the first commercially available medical 3D, which is now available in every scanner that you can think of. I took those ideas and brought them to Stanford University, realizing that it was a way of seeing the world that students might really be able to use. We developed a lot of educational programs, both with 3D and without, with simulation and without, and many of those were used around the world.

But I realized that developing things for 400 medical students…it was really hard to take it out and apply it elsewhere, which is why I then, after eighteen years at Stanford, started a company and I’ve done that for the last ten years. That was a wonderful experience to actually take something and make it commercially available to people you never met, you weren’t going to see, and they were out there using it. That was both delightful and sometimes profitable. Business is a hard business!

We’ve made a number of products along the way, and at this point, I feel that 3D simulation, VR…these are such familiar concepts that to some extent, I can back off and watch what’s happening in the world.

Van: Well, you’re such a pioneer in this area. I’m curious, Parvati, there was a moment in time I would imagine when the tactile experience of maybe being in the room, or touching the equipment, or touching the cadavers…there was a sensory experience to how you learned and retained the material. Was there a technology that enabled the 3D experience to be similar or even better than the in-person experience? When did that happen?

Parvati: You raise a really interesting point. 3D in the computer is not the full 3D experience in the real world. At one point, I really worried about that and I got deeply into the world of haptics, how you can feel virtually. I did that through working with laparoscopic and endoscopic tools. So, you would have a real tool in your hand. You would see the 3D anatomy patient, the inside of the patient. As you moved your real tool, a virtual tool interacted with the inside of the patient and there was a set of motors inside your real tool that resisted based on what your virtual tool was touching. So, you got haptic feedback from a virtual system. It works. It actually is pretty good. But boy, is it expensive to do all that!

So, I feel that one has to be very thoughtful about what aspects of 3D and realism you need for the task at hand. Teaching, practicing, experiencing…they don’t need every aspect of realism to be available.

Van:  That’s helpful to think about the investment in that way. So, is there a rule of thumb of when you would apply the 3D and the haptics for learning sake?

Parvati: Actually, my rule of thumb is first, don’t do it. Don’t do it unless you know why you’re doing it. If a web page suffices, OK. If a 3D sketch is good, OK. But there are some times when you need it. Is it something the student will have a hard time understanding — maybe take ten times longer if they didn’t have 3D? Would it be fifty times more difficult if you didn’t give them the sensation? Then use haptics. Haptics does not have to be computer-based. A lot of medical learning happens on chickens, on oranges. You practice putting in a needle, you practice injection on a piece of meat. You learn how the needle goes through the layers of the piece of meat. You don’t have to have a virtual haptic system. You learn the feel of tissues by using real items that maybe were going to be thrown away by the butcher and that you use.

There are companies using computer-based haptics. One of the most useful applications is in teaching minimally invasive surgery where you have tools interacting with anatomy. Haptics in the tool is actually one of the more doable kinds of haptics. As soon as you move to finger-based haptics — touching and feeling — now you need to put an exoskeleton on your hand…something that gives a little sensation to your fingers.  That technology is really not well here. So, haptics is something that I would hesitate to use except in very restricted situations.

Van: That’s very helpful because this field of technology is so exciting, I’m sure, to everyone as they look from the outside. So, it’s helpful to hear your expertise in explaining what’s the basic case for using simulation.

I was wondering if you could share even more stories about when using this type of technology is helpful specifically for healthcare training.

Parvati: First of all, I’d like to begin with a little story.  I was speaking with the head of surgery at Stanford about why he was so excited about simulation. He said, “Well, think of it this way. No airplane pilot would fly a jet without first going through the simulation hundreds of hours. Yet, we allow medical people to jump right in and work on a patient.”  At that time, he said what is happening is the equivalent of one jumbo jet worth of people crashing and dying in preventable medical deaths every day. That was a very graphic description of what simulation might be able to prevent.

As far as actually observing the most useful simulation situations, there are two kinds of simulations that are extremely useful. One is whole person, whole environment scenario simulation. These are very difficult to set up in real life.  You have to take a space, bring in people, set up equipment, teach them the story of what’s happening.  Everybody plays a part. These are done all the time, but they are expensive to do. They take a lot of effort and they’re not always replicable time after time after time.  How many times are you going to let a person practice that one scenario with all these other people and equipment and space? That kind of situation, if you can set up a 3D, maybe even a VR simulation, it is a very effective use of resources once you’ve created the resource. That’s one whole area where I think simulation can be very useful.

The other is a very focused, task-based approach. Now, think about learning minimally invasive surgery with tools. You’re learning a task, you’re learning to go in, cut, suture and so on.  Let’s think of it a little bit and go beyond that.  There’s a resident learning surgery. They’re helping, they’re watching and they’re actually doing all of a series of steps to, let’s say, resect a piece of bowel. First of all, a learner is not going to be given a lot of opportunity to do this many times. But secondly, if you really study that surgical procedure, you will find even amongst experienced people that there are certain steps where the mistakes happen, where omissions happen. Maybe that’s the step at which they needed to check whether there was an anomaly in that person’s body and the nerve was in front of rather than behind a structure. So, certain critical tasks where failure can happen.

That kind of critical task is also something that is very amenable to training through simulation, particularly if you do the sequence before and the sequence after — just that segment of that procedure — and let them arrive at that task and make sure, first of all, they practice the basic tasks correctly, the variations correctly. They don’t need to do the whole surgical procedure again and again and again.

So, two very different aspects of use cases of simulation.

Van: Oh, that’s very helpful to understand.  On the latter of the two, you said “task-based approach.” Do they tend to be high stakes tasks? And when does your organization get called in to do that, to create that simulation?

Parvati: They often are very high stakes because an error in surgery means fixing it later, which is difficult.  Either you fix it during that same surgery or you have to go back in. Bad. In that sense, these are, particularly in surgical cases, these are high stakes situations.  How often is training done with simulation? At this point, that kind of high stakes training is not happening as often as it probably should.

What I’ve noticed as far as deciding what simulations to develop… unfortunately, I believe it is not really very need driven. It tends to be what can be done, what a person thinks should be done or what is some obvious thing to do. There is not enough analysis of the details of the curriculum where simulation would be particularly appropriate. Or of post-facto analyses of problems and errors in the hospital where going back in with simulation would be good.

Now having said this, let me just say that when there is good collaboration between the simulation center and the various services, they do come in with the full mannequin-based simulation system and run scenarios to train on difficult or erroneous situations. That is happening.  But what I’m talking about, this more task-based focus — let’s go in and fix very specific problems — I don’t believe it’s happening yet.

Van: I see.  On the first example you mentioned, which is the whole person environmental scenario, I remember back when I was working in the energy sector, there was an emergency planning mock session where all the hospitals, the electric company, the gas company, all the first responders did a simulation where if a plane crashed at San Francisco Airport, for example, how would the coordination of response occur?  It occurred to me that something like that is so complicated.  Would that be an example of what you mean by whole person, whole environment scenario?

Parvati: Running mass casualty scenarios of different kinds is a requirement for hospitals nowadays. The Joint Commission requires, I believe, two such major scenarios to be run every year involving most of the different divisions of the hospital, the first responders, the media and outside agencies. So, yes, that is a very good example of a large-scale scenario that is required today.

Van: Well, as you were laying out these possibilities, I thought, oh yes, that would have benefited from a true simulation, right? Because it’s hard to know what your role as an individual would be in this mass casualty event unless you get a chance to practice. Who do you talk to? What’s the chain of command? What systems do you use? Where should you physically appear?

Parvati: Scenarios that are being done nowadays are very much labor intensive. They involve real people, real red material simulating blood, real gurneys, real ambulances. It is a very large effort.  There have been efforts to put this into the computer, into 3D. And in these cases, typically everything is simulated except the one person who plays a certain role, which might be the first responder’s role, the nurse’s role, the hospital incident commander’s role.  But it often is a single-person simulation.

Now, there have been efforts to make multi-person simulations and we actually did one of those ourselves for University of Memphis where we simulated an earthquake. There were lots of casualties, and the task was for a team of caregivers to fan out, go through the earthquake scenario in 3D and find victims, triage them, tag them so that somebody else could come and decide whether to put them in an ambulance and take them away or let them stay where they were or ask them to walk to a site for help.

That was a very useful scenario for training and they actually did a study comparing the computer-based scenario with a real-life scenario and with kind of a text-based practice. The computer-based scenario actually came somewhat close to the real-life scenario in what people learned, how they did things, and particularly how they learned to triage and write the triage tag. The large-scale joint commission scenarios are even bigger and I don’t think I have seen a full computer simulation of that, though I think there are a few places, maybe in South Carolina, where they are making a very good effort to do this.

Van: Parvati, I just came from a meeting of the state council that is focused on the shortages in the healthcare workforce. As you know, it goes from physicians and nurses all the way to the allied health occupations. One of the gating factors is always the clinical training, which is the moments where students can spend time practicing skill sets on live patients. I was wondering, to what extent can clinical training be replaced or augmented by simulation?

Parvati: That is actually a very important question these days for many reasons. Getting enough people to get enough clinical training means finding enough clinical slots for people to train in, and they just aren’t available. The National Council of Boards of Nursing did actually look into whether simulation could be used as a replacement for some of clinical training, and they concluded that with the right simulations, you should be able to replace up to 50% of clinical time. That was specifically done for nursing, but probably would apply to many of the allied health jobs.

It is not actually happening at that level yet.  Some states are allowing up to 25% time to be replaced, some not at all, and some up to 50%. The question that arises is, is there enough clinical training content in simulation to actually teach, say, up to 50% of the 600 hours of clinical training that a nurse requires? Is it all doable with mannequin-based team training, which is how a lot of simulation-based training is done today? Is there enough other type of training content in 3D, in VR, in role-playing that can be used? And one probably needs a lot more needs-based development. So, you specify the need based on the curriculum and develop the simulations for that.

I think you really could use simulation to do a huge part of the training. You still, of course, need the clinical time, but you could arrive at your clinical time a whole lot more prepared and able to get more value out of your clinical time. So, yes, I believe it can and should be done.  Are we there yet? I don’t think so, except with very expensive, large-scale mannequin-based training.

Van: Parvati, if you were to look at the horizon and your wish in terms of a breakthrough technology, breakthrough tool that would enable that world of better clinical training via simulation, what would be that breakthrough technology for you?

Parvati: It would actually be what I call an ecosystem.  We need a Lego-style ability to put together simulations. We need bits and pieces of already working simulations that we can put together to create the simulation we choose, and then, of course, modify pieces to teach exactly what we want to teach.  But right now, so much of simulation development is, if not from scratch, it is very customized and each company or group that’s developing the simulation uses their own capability.

In the simulation world, we do use some very good tools and there are software capabilities from companies like Unity that make it much easier for us to get going. But we still do way too much custom development on top of that. So, particularly in medical simulation, if we could build on each other’s simulations so that we could more rapidly create the desired exercise, that’s what I would consider to be a desirable technology.

Could I add something to that just on the side?  Ever since multimedia started, we’ve talked about sharing and reuse and stuff like that, and it’s really not worked well because for many reasons, it doesn’t fit the right thing. That’s the same thing that’s going to happen with simulation, unless we go to this really, truly Lego-like approach, which is not here.

Van: I see. So, this will be akin to the early days of web pages, where you had to program everything from scratch, and now, you could just ask ChatGPT to generate the code for you?

Parvati: It is actually, yes. In the early days, we all made frames and positioned things and created each page and took a lot of effort to create them.  And now, you can — if not ask ChatGPT itself — there are enough tools that one could pretty much throw together a web page and then put in the content you want. We are not there with simulation, but that’s where we want to go.  We want to have enough bits and pieces that we can construct the simulation we desire. For example, I should be able to have a full-scale patient room and all the equipment all working in the simulation. I should be able to say, “Okay, today we are going to practice fever that is not coming down.” So, there is one thing that we are changing in that whole rest of the simulation. One symptom that we are changing, one functionality that we are changing.  And you would decide to do it today and have it ready tomorrow.  That’s how it should be.

Van: That would be a very powerful learning tool, if we could get to that state.  Well, tell us, Parvati, about the Health TeamSpaces that you’ve created and what is that bringing to the field?

Parvati: Health TeamSpaces is both a product and a platform that we developed, and it’s an outgrowth of a lot of earlier work.  Originally, we had CliniSpace and now we have Health TeamSpaces. It is an effort by us to create somewhat of a Lego-like simulation world for ourselves. We create the entire environment — the rooms, the equipment, the patients, characters — and we also made it multiplayer, which means a team could log in as different characters and see each other and work together in that space

In that sense, it is very different from what a lot of people are doing, which are single-user simulations. We also created an authoring tool that allows you to change the patient’s condition and have it be dynamic. For instance, if you gave medication or fluids, it would appropriately change the blood pressure, bring the fever down, and so on.  So, it is our own Lego set, you could say, to create simulations. In fact, we have used that to create simulations for others.

It turns out that multi-user-based simulation is actually difficult for faculty to figure out how to use, to keep their minds on so many different roles and the teaching around that. So, even though we have Health TeamSpaces as a powerful multi-user clinic and hospital environment, we find that we are asked much more often to create single-user, single-learner environments.

One of the most powerful such simulation products that we have created is in the area of anesthesiology. We were asked to create five scenarios where an anesthesiologist is called in — in the middle of an ongoing operation or some other situation — and as they work with a patient, different things go wrong which they have to perceive, identify, explain, and act on.

You could think of these as branching scenarios where the person comes in in a certain situation, and as they work, the program throws in problems which have to be dealt with. If the person is successfully dealing with the situation, the patient gets better and maybe another problem appears. If the learner is not successfully dealing with the situation, the scenarios will actually progress to a worse and worse case.

In our case, we were requested to allow the patients to die, which is a tough thing for learners to handle.  There’s a lot of discussion around that topic. These five anesthesiology cases were requested originally as recertification cases, so very high-stakes simulations for practicing anesthesiologists, that means very senior people.  The way it is actually being used, is very interesting. They are allowing the anesthesiologists to do these cases again and again and again until they successfully get through the case, and then giving them the credits to recertify.  But it turns out that that whole process of doing the branching scenarios in multiple different directions is both a very good learning situation and actually is a good test to see whether they’re able to actually manage very difficult cases.

Van: That sounds like a fantastic learning environment.  My father is an anesthesiologist, so I’m fascinated as I hear how this training is occurring.

Parvati: The scenarios we developed are probably the most complex out there currently in the medical simulation learning world. It took determination on the part of the society to make it happen. It took money from them.  Frankly, it required a small company to ally with a very large company, also in healthcare, so that the society had the confidence that the products would not disappear. And it took a team of anesthesiologists to design these scenarios in a way that they tested all the aspects that they wanted to test.

Van: And you’re absolutely right that some situations probably don’t arise on a daily basis or maybe even an annual basis such that the anesthesiologist can practice. So, how powerful it is for them to get exposed to those situations in simulation.

Parvati: You’re right. There are so many situations which are very low frequency, very high acuity, which are not going to be available for practice in real life and which are just perfect for doing in simulation.

Van: Well said. Well said. My final question, Parvati, is I wonder how the world of AI and machine learning is going to affect the technologies and tools that you have at your disposal?

Parvati: AI is going to be wonderful. It’s going to be able to help in many ways. Intelligent tutors built into these simulations observing every action and offering either immediate or maybe delayed suggestions for learning are a direction that is very desirable. The possibility for scoring performance with all the data one’s collecting in the simulations and identifying those critical points where the person seems to have difficulty, again, is an excellent AI application to the data being collected. The possibility of working with cohorts with the data from a whole cohort doing a simulation and using that to analyze and decide what is necessary for the next generation of simulations.

The data analysis from these cohorts going through simulations tells you a great deal about deep problems sometimes in how people are performing those procedures and where they are having unnecessary problems.  The possibility of taking this information and not just building another simulation, but going back to the hospitals and saying, “Look, your anesthesiologists, your nurses seem to have real difficulty in this area. Have you looked into why it’s happening in your hospital?” These are possibilities using big data machine learning AI that could be so useful in more efficiently improving what is going on in the hospitals.

Now, let me step to another side. There is a whole lot of excitement in AI about the new generative AI tools, which is just one part of the larger world of AI. I think there’s a real interesting possibility for generative AI to generate scenarios from a body of scenarios. If we had the underlying infrastructure such that a scenario generated through this AI tool could then be implemented into a simulation we are then closing the development loop and then I think we have something really good.  At the same time, of course, you have to evaluate all the scenarios generated by that AI for accuracy and usefulness.

Van: Wow, you have got us thinking about the future of learning.  Thank you so much for being a pioneer in this area and sharing what you know with us.  Any final words?

Parvati: I think the possibility of computer-based simulation, VR, computer-based learning, AI, intelligent tutors…these are so amazing. At present, I think we are excitedly building whatever we can build. We need the space to step back, understand where these could be really useful, perhaps essential, and work on those problems and get those learnings built.

Van: And that was your point when we first started: what is the best application of these technologies?  Well, I so enjoyed having you with us.  Thank you so much, Dr. Parvati Dev, for being with us today.

Parvati:  Great. Thank you so much. That was such fun.

Van: I’m Van Ton-Quinlivan with Futuro Health. Thanks for checking out this episode of WorkforceRx.  I hope you will join us again as we continue to explore how to create a future-focused workforce in America.