[Blows dust off mostly dormant website.] This post is the written version of a presentation delivered at the Australian and New Zealand College of Anaesthetists Annual Scientific Meeting in 2023.

My job here is to talk about the very small task of creating a world you hope to see. It is a pretty circumscribed world as far as worlds go. It tackles the relatively small planet of changing one little element of care for younger people. And the change involves the attempt to explore newer media as part of what we do in anaesthesia.

But we should start with a question I guess because it’s also my job to engage the people out there.

So the question is – is this new world of technology-enabled working utopia? Or a broken dystopia that doesn’t even need a flickering fluoro light to produce a cold pang of terror?

OK, that’s a little dramatic, so let’s go back to the basics.

It’s a simple job

This is not the first occasion or first time that I have noted that my job as a paediatric anaesthetist is to get paid to provide many controlled substances to small children. Sometimes those children have a little bit on their minds given the incredibly strange plan for their day involves starving for a bit and then meeting a bunch of people in a hospital to have *a thing* done. Or those children are actually up for some kind of adventure. Because some kids just are.

So on a good day my job involves creating new worlds with children. Maybe we’ll be exploring underwater. Perhaps that funny sensation is just a really obvious thing because that’s what happens when you reach out to tickle a star – it sparkles in your arm. Or perhaps today they’ll have the job of showing the big person in their life how to take a ride into space.

So when we first started thinking about Virtual Reality (VR) as an option for providing distraction for children and young people it seemed like a an amazing new option to change how we go about things. Not just something that could provide distraction here and there and for something a bit more than just the cannulation at the kick off for which there are some pretty good solutions out there. We wondered if this was a technology that we could use to support longer procedures that currently have a general anaesthetic as part of the mix.

We wanted more than that though. Could this be something in the technology mix that we could use to enhance the therapeutic link between the medical types and the patients? Perhaps it could even be technology that could see kids designing their own experiences for distraction to teach us what they need? Or maybe it could reframe how they participate in looking after the next child by leaving something behind in the virtual world to be found later?

The possibilities seemed to explode our options to something like … well, the scarcely comprehensible galaxy captured in an open access image with a highly impressive purple blob in the middle of it.

Which is why we got to work changing what we do.

Utopia is Right Here

Really it was all pretty simple. We purchased headsets, came up with a way to share it to another screen, negotiated with ethics to try and turn it into research work for around a year, and built up a library of available things to use.

And the result when we introduced it to patients was pretty much a fantasia of exploding fireworks. As long as we kept our focus on the right scaffolding around the experience we were able to undertake procedures that were typically a routine trip to the land of nod with nothing more than local anaesthesia and VR.

More importantly it really delivered on that goal of enhancing the engagement between patient and clinician, all while empowering the young people who were trying it out. And avoiding a bit of fasting.

Line insertions, lumbar punctures, bone marrow aspirates and really pretty intense amputation dressings all supported with VR.

Which is awesome because I’m done here I guess.

Except for no. Not done.

If I was to give some sense of where we are now on our mission to provide new options, redefine the therapeutic compact and make picking up this alternative device as easy as switching on the volatile agents, I would use this picture…

It’s not a regular thing. It’s very ad hoc. And we keep dealing with technical issues while Meta mess with their systems.

Which means maybe it’s time to evaluate how our utopia turned into the sort of place that is only utopia for scavenging rodents.

Does the evidence back it up?

Perhaps with the benefit of a bit more time and with more out there it’s time to evaluate whether the evidence still says this gear is good. So to provide a quick way in let’s consider this systematic review and meta-analysis published in 2022 by Tas et al.

It operates as a sort of update to an earlier systematic review by Eijlers et al (which was actually the same team!) that looked at what was out there up to 2018. This one extends the search to 2020 with its eye firmly fixed on including studies that investigated the effect of VR on pain and/or anxiety in patients 21 years old or less who were undergoing medical procedures. Since VR can mean an awful lot of things they defined it as ‘a technology that allows for the creation of immersive and interactive environments and is projected in front of the user’s eyes via a head-mounted display’. They were so keen on ‘interactive’ they actually excluded anything they felt was not interactive enough, such as any sort of 360 degree video. Which is a really interesting choice.

After their search they were left with 27 studies. This was thanks to the addition of 13 publications beyond the Eijlers version, which also had 13 papers included. Now if you’ve immediately jumped onto the maths there, this is because one of the publications actually included results from two studies. So they treated the data as two separate studies.

And that makes 27. Consider the 13+ as a research baker’s dozen maybe.

And the end results? Well the overall weighted standardised mean difference for VR distraction was -0.67 for patient-reported pain, -0.74 for patient-reported anxiety and -0.58 for anxiety when used to familiarise a patient with their procedure before the event. To be clear, for this analysis a value in the negative range means that VR resulted in reduced pain or anxiety compared to ‘Care as Usual’.

So this evidence suggests that VR as a tool is super effective and excellent. Which brings us back to why are we failing?

Maybe a closer look at the same reviews give us a different kind of hint on that front.

Under the Surface

Another look at the systematic reviews might just hint at the fact that getting things up and going with VR isn’t quite so straightforward.

Let’s start with the numbers that relate to studies. For this it is important to know that VR is not new technology. It’s been around for decades although it definitely took a leap in the consumer space over the last decade since Palmer Luckey first got going with what would become Oculus in 2009 in his garage at the age of 16.

(As a sidenote if someone scripted that a kid would unlock consumer VR better than anyone had thought feasible, then sold the company to a mega corporation which would eventually push him out amidst controversy over his support for a multi-impeached President, but would rise again like a phoenix by developing a start up focussing on autonomous weapons systems to give the US and its allies a military edge I am fairly certain they would be drug tested. Then maybe offered a Netflix series, who knows?)

Anyway, in this whole systematic review they used the fishing net search terms of “VR”, “children” or “adolescents”. This led them to 1824 papers. Which they whittled down to … 26 papers that included that bonus study to make 27 studies. The earliest paper seems to be from 1998. 22 years to produce 26 papers worth including seems like not that many.

Dive a little deeper again into the sea of numbers and across these studies, with patients ranging from the ages of 4-21 the total number of study subjects included was 1695 and the largest single study had 201 subjects in its RCT.

If we were looking at any other change in care, how confident would we be if the breadth of the literature seemed that skinny? Even the authors say ‘there’s just not many studies’.

This is absolutely not a criticism of those undertaking the review, or the committed researchers who have published their work in the area. It just makes me think that exploring this area is tough.

For those of us seeking to interpret that literature the challenges don’t end with the numbers. Tas et al note themselves there are limitations that are hard to overcome in what is available to them. For starters, ‘care as usual’ which is the most consistent comparator, is often very poorly defined or it varies greatly from study to study. To back it up, when the reviewers compared the size of effect across studies within their areas (pain reduction or anxiety reduction or anxiety drop with familiarisation), they noted that the heterogeneity was around 60-70%. That’s a lot of variability when you’re trying to consider outcomes with only a few studies to choose from.

The issues are well summarised in a response from Brigden et al who note that across the 26 publications, there are 7 different types of procedures included as settings for VR distraction and 17 different arrangements of VR hardware and software (and I’d note a bunch of those hardware options don’t exist any more).

So if really clever and committed people are struggling to either build this up as a big enough phenomenon to study, and comparing apples with apples is made impossible by there being actually 7500 types of apples in the world to compare (OK there is not 7500 VR arrangements to try out but let’s go with it) then what are the barriers to making this standard practice?

What does it take to actually do the VR bit?

Now there are different systems out there and indeed at least one of those in the systematic review which does well on RCTs is a pretty straightforward ‘set it up and press play and then click it into the headset’ system called Smileyscope.

What we’ve been trying to produce though is something with flexibility and length that makes the social interaction with the clinician a key part of the experience. So we have been delivering experiences using the Oculus Quest. To really make it work it’s vital to have the inside headset experience casting to a separate screen. This is important for practical reasons in case they need to be guided through menus or the like.

But the equally crucial bit is to give the clinician the opportunity to be part of that immersive interaction. There is solid evidence that children and young people are keen on synchronous sharing of their experience. There is an underlying desire for lots of users to talk about what they are seeing and experiencing. This provides a tremendous opportunity for the clinician to be part of the experience, but also to offer up prompts and direct the focus of the person having the procedure to enhance the distraction. This scaffolding makes the most of the experience and ensures that you as the clinician have contact with what’s going on. That contact is particularly important to make sure that when you need to provide anchoring to the actual procedure (‘you’re going to feel something cold on your back which is just from cleaning your back’) you are an expected part of the scenario.

All of that preamble might have already got you thinking ‘there’s a bit to this, hey?’

And this is where the concept of friction comes in.

Friction in a user experience is a shorthand way of expressing all those little barriers or hurdles to clear to actually get to the main event. To get to that point of having the headset on and enhancing that experience we’re still at the point where we need to put in effort to make it seem effortless for the star of the show – the patient.

To get to that point we’ve had to match the procedure and the patient, then consider which experience in our library can work in with the procedure itself. We have to check in with the young person and those around them to see if they’re keen. We then need to select the location, set up the guardian boundary within which it can work, consider if the experience itself will work for the procedure (e.g. you can’t easily lie down in a lot of experiences built in the commercial space and you might not be able to use your hands freely) and ensure that the headset and the casting screen are working and on the same WiFi network.

After all that we have to get the proceduralist ready for to play their part while also being ready with the back-up option if it doesn’t work.

That’s a lot of friction.

And if you take a step back and think about how frictionless the experience is of walking into the room, turning on a rotameter and cranking up the volatile agent, the relative ease of that option compared to everything I just described might just be the story of why the full potential of VR just doesn’t seem to be hitting.

The key thing about the ‘standard experience’ is it is sort of invisible. That’s what needs to happen for a change in practice to become embedded. There’s literature describing this under the umbrella of ‘normalisation process theory’. This basically says that for a change to become embedded practice it needs to disappear from view. The goal is to make it so seamless that the challenges of choosing that path just don’t register.

This can be achieved by delivering a few relatively simple elements:
1. Coherence – there has to be a clear and defined reason for the clinician to want to do it. It can’t be too big and vague. The goal has to be a very simple ‘using VR can offer a better experience for yourself and the patient’.

2. Cognitive engagement – the clinician has to think it is worthwhile. That could probably be best achieved by having a few champions showing the way and acting as trusted messengers.

3. Collective action – this is the work done to make it easy and remove all those barriers.

4. Reflexive monitoring – we have to keep sharing the impact of the change.

It seems like a lot. But having experienced what it is like to have the VR option work out for a child I just can’t let go of the idea that it’s up to us to solve that friction problem and make it all disappear.

What do we need?

Well actually probably I need someone to speak up and tell me how they did it better. We could use tips.

But I could summarise all of what we’ve learnt by saying that the most important part of trying to make the most of new technology is the same as the rest of healthcare.

It’s about people.

In those initial phases we actually need to have people available to solve the friction problems and be the people who make it make sense to the clinicians who just need a great option to do a great job.

And that’s absolutely been one of our challenges. In a strained health system it’s easy to keep choosing what is, right now, the easy choice. An anaesthetic.

This lack of people isn’t an excuse ultimately because we deliver things, including change, without really having enough in the way of resources all the time. So that’s one we just have to solve better.

But there’s another area where we need people.

We need people to design better experiences.

Specifically we need clinicians who understand the work to be involved. And we need to make the most of the expertise of the children and young people who, let’s be honest, spend plenty of their time teaching their parents how to make technology work.

Making the Most of Learning from Kids

That is one bit of research we’ve managed to do.

In between waves of COVID we partnered up with a team from the School of Design from The University of Sydney to run workshops with kids where we just had them explore a VR world and show us how they might build their own experiences in that world.

Using Tinytown VR, which offers up pre-built assets and worlds, 15 children from the ages of 8-15 spent about 30 minutes after a little guidance building whatever they liked. While the work is still being written up it is easy enough to say there were definite themes emerging from the real time recordings of their builds and the explanations offered up afterwards.

We met ‘explorers’ who spent their time seeing what was out there before building what seemed like fairly unstructured and disconnected elements but which rapidly became part of a narrative they created linking all of that randomness.

We met artisans who carefully worked through every asset before creating their very special thing.

We found planners who went straight into a structured plan and delivered it before further exploration.

What was profound though was some of those earliest theories we had – that young people would be interested in the ongoing sharing of their designs and draw immense value from the concept of ‘building for others’ – were very much borne out.

The other key group to bring into the design process are those involved in the procedures. The commercial folks out there just don’t need to answer the same questions that we face on a daily basis at the hospital. They don’t need to consider accessibility in the same fashion, or limitations that might be placed on your movement or use of controls.

They don’t have any interest in trying to figure out how to facilitate a procedure that requires you to sit up and curl over a little bit to make a lumbar puncture easier.

An hour in a workshop with clinicians however has led us to a raft of ideas involving peering beyond a keyhole to the world beyond, or stories in the forest that let you curl up and peer through the undergrowth or roll onto your back to follow a ladybird into the sky when you just need someone to lie flat.

So better VR for clinical use will only truly be possible if we get the right people involved.

Or of course AI.

Words You Have to Mention

That’s right. This is a thing that is about technology so the rule now is you have to mention Artificial Intelligence. I’m sorry. I didn’t make the rules. I’m just living by them.

Let’s say that we do learn all the right things about how kids might help design their own experiences (or how clinicians might actively design in real time) we still have to figure out how to give them the tools to build what they have in their heads.

And genuinely, AI might be able to help with that. A system that can be trained by the user to respond to ‘I want to make this….’ by giving them what they will need specifically in their hands for their style could be the key to dynamic design that removes the barriers to getting on with free design.

As it is we’ve already used a language learning system to test out whether that kind of system can help write VR experiences with a little bit of prompted set-up from the design team and a couple of examples we’ve worked up already. In a handful of minutes we had the script for a VR experience called ‘Magical Tree of Life’ where the user can nourish and grow a tree in surprising ways within a namespace, complete with user instructions and key goals within the game along the way.

Of course any system that relies on a learning system like that will still need those who understand the space to keep an eye on things.

Also known as people.

So my whole topic was about introducing new technology into the real world of patient care.

But it turns out making the most of technology is actually about people.

And to fully realise the potential of VR we still have a way to go to solve the friction problem.

If we do that in a way that empowers both our patients and the people who want to look after them we can genuinely offer something entirely different to enhance parts of our practice and the patient experience.

Which means that on a really good day at work I won’t tell stories to kids about flying into space.

We’ll build better stories together.

The References and Assorted Miscellanea

A good deal of this content is describing our practical experience or referring to some of our research work that isn’t quite published yet.

That systematic review is a solid effort though and you can find it here:

Tas FQ, van Eijk CAM, Staals LM, Legerstee JS, Dierckx B. Virtual reality in paediatrics, effects on pain and anxiety: A systematic review and meta-analysis update. Pediatr Anesth. 2022;32:1292-1304. doi: 10.1111/pan.14546

And here’s the earlier review:

Eijlers R, Utens EMWJ, Steals LM, de Nijs PFA, Bergman’s JM, Wijnen RMH, Hillegers MHJ, Dierckx B, Legerstee JS. Systematic Review and Meta-analyis of Virtual Reality in Pediatrics: Effects on Pain and Anxiety. Anesth Analg. 2019;129:1344-53.

That letter response to the Tas review is this one:

Brigden ZM, Zarook E, Patil V. Virtually pain free: a comment on the use of virtual reality in pediatrics. Pediatr Anesth. 2023;33:176-7. doi: 10.1111/pan.14591

Were you looking to shake things up in your neck of the woods and think normalisation process theory might be just what you need? Well you could choose to have a look at this paper:

Murray E, Treweek S, Pope C, et al. Normalisation process theory: a framework for developing, evaluating and implementing complex interventions. BMC Medicine. 2010;8:63.

That work mentioned about seeing how kids design isn’t published yet. However we do have some earlier gathering the insights of many on how to do that whole design thing and what people want from VR in the procedural setting:

Ahmadpour N, Weatherall AD, Menezes M, Yoo S, Hong H, Wong G. Synthesizing Multiple Stakeholder Perspectives on Using Virtual Reality to Improve the Periprocedural Experience in Children and Adolescents: Survey Study. J Med Internet Res 2020;22(7):e19752. doi:10.2196/19752.

The images in this are all Creative Commons and found via Unsplash. Image credits to Jeremy Bishop, Conor Samuel, David Close, Alexander Andrews and Markus Spiske.

Did you get this far? Then maybe you should just spend a few minutes enjoying this Hot Wheels adventure via The Kid Should See This.