Cost-utility analysis of surgical fixation with Kirschner wire versus casting after fracture of the distal radius

Listen to Andrew Duckworth and Matt Costa discuss the paper 'Cost-utility analysis of surgical fixation with Kirschner wire versus casting after fracture of the distal radius' published in the November 2022 issue of The Bone & Joint Journal.

Click here to read the paper.

Find out as soon as the next episode is live by following us on Twitter, Instagram, LinkedIn, Tik Tok or Facebook!

Transcript

Copy Transcript Copied

[00:00:00] Welcome everyone to one of our BJJ podcasts for the month of November. I'm Andrew Duckworth and a warm welcome back to you all from our team here at The Bone and Joint Journal. As always, we'd like to thank all of you for your continued comments and support as well as the big thanks to our many authors and colleagues who've taken part.

We hope that you're continuing to enjoy our podcasts and all of our knowledge translation work delivered from your team here at the Journal, our podcasts continue to focus on papers published each month here at the BJJ, as well as our accompanying special edition podcast series. So today I have the pleasure to welcome back my editorial board colleague here at the journal, Professor Matt Costa from Oxford to discuss their paper entitled, 'Cost-utility analysis of surgical fixation with Kirschner wire versus casting after fracture of the distal radius: a health economic evaluation of the DRAFFT2 Trial'.

Welcome back, Matt. It's great to have you with as always. Thanks Andrew. So Matt, the, the aim of, of this study in particular that the one that's published in the journal was to compare the cost effectiveness of fixation with K wires versus moulded cast for adults following a manipulation of a fracture of the distal radius in an operating theatre setting, which is important as we'll [00:01:00] come onto using data from the DRAFFT2 trial.

So Matt, maybe is a background and maybe taking a step back, you know, for our listeners, can you give us a brief overview of maybe DRAFFT1 briefly. A lot of them will know about it and then how you sort of led into to DRAFFT2. Sure. So DRAFFT1. Crikey that seems like a long time ago now.

It started out in sort of 13, 14 years ago. But it was a very, a simple idea to compare locking plate fixation versus wire fixation for patients with a dorsally displaced fracture in whom a closed reduction was achievable, so that's important. It was, if you couldn't reduce the fracture closed, then obviously it wouldn't be appropriate to wire it.

So it was for those patients who had a closed reduction that was successful, of course, which was most of them, and much of my surprise and the surprise of the Hand Society in particular, it turns out that if you can maintain the reduction, then how you hold it there doesn't really matter. So the wires gave very similar clinical outcomes to the plate fixation.

But because interestingly, all of the, [00:02:00] the cost, the resource used was linked to the use of the, the choice of implant, then the wires were much cheaper. And so it was cost effective to use wire fixation. And that led to a change in clinical practice in the UK quite a marked one actually quite interestingly, and also change in, in a NICE guidance.

So our guidelines on how to manage a broken wrist which basically said if you can achieve a closed reduction, then you can use wires to hold it there and reserve plates for those patients to require an open, open reduction. Yeah. So the the follow on from that was, well, if we can achieve a reduction, it doesn't matter which bit of metal we use to hold it there.

Do we need a bit metal at all? Or will a moulded close contact cast intervention provide similar outcomes? And that was the basis for DRAFFT2, which compared wire fixation for versus plaster cast, basically a moulded cast for patients who where a, a closed reduction could be achieved. That's brilliant. And like I say, I think you always emphasize this one.

I've heard you talk about it as well, but it is those fractures that you can reduce closed. And I think some people get a bit [00:03:00] wound up about these things and, and think that, you know, it's not suitable for all fractures. But that's not actually what, what it says, it's for those fractures that you can reduce closed and hold in that way.

Yeah. Yeah, absolutely. So I, I still use plenty of plates in my, my clinical practice as we all do, but I'm just a bit more selective and reserve those for where, you know, the bits of the lunate fossa fragments are facing the wrong way and a, a need to open up in front of the wrist and if I'm there anyway, to, to put a plate on that makes a lot of sense.

But yeah, so it was reducing the indications for plates. We're certainly not saying throw them out . Absolutely. Absolutely. And so if we, just briefly, before we go onto this, this paper, DRAFFT2 was recently published in the, in the BMJ. And what, what were the key clinical findings, I suppose, from that? Yeah, so I guess two really important bits of information.

The headlines if you like, were that if you can hold a reduction in the right place, it doesn't matter how you hold it there. So the patients who had wires versus casts ended up with functional or had the same very, very similar functional outcomes and quality of life outcomes within the first [00:04:00] year.

So at each time points in that first year of, of follow up. However, one in 8, of the patients in the plaster cast group lost their reduction in the first couple of weeks afterwards, and everyone clinically will recognize that phenomenon. When the swelling goes down, the cast gets a bit loose and you can lose position.

So, not a huge surprise, but really important to quantify what proportion of patients ended up with that problem. So one in eight required a, a surgical intervention and usually then a, a fixation of course for their fracture. The corolla- corollary of that, of course, is that seven, they didn't need a a fracture.

And from a patient's perspective, that's not too bad odds to avoid surgery if you, if you can, yeah. But this is where the cost effectiveness comes in cause the, that degree of nuance is, is important for the interpretation of the, of the resource use data as well as the clinical outcomes. No, absolutely. I think, I think that it is nice that we've covered that already.

Cause I think, like you say, when you're interpreting the, this paper, the cost effectiveness paper, knowing that one in eight number is actually quite important, isn't it? Cause [00:05:00] it has implications for, for, for how, how the costs and the complications, et cetera. So if you move on to this study, obviously as I said, there's an economic valuation based on data collecting from DRAFFT2, you know, DRAFFT2, a big multicentre UK trial.

So maybe just as a, going back again a step, just roughly the inclusion exclusion criteria for the trial. I know they're, they're very clear from the original paper. Just, just so that people know what sort of patients we're talking about here. Yeah. So we, we try and keep these as broad as possible. So the inclusion was basically if you were an adult, so anyone over the age of 16 with a dorsally spaced, distal radius fracture.

And your surgeon in conjunction with the patient obviously decided that manipulation was appropriate. So the fracture was displaced, then you were included in a trial. Patients who were excluded if a closed reduction couldn't be achieved. That was the key criteria. So all of the patients in this study were randomized after they did, had a successful closed reduction of that fracture.

But a few other exclusion criteria were used rarely. So open [00:06:00] fractures, where the soft tissue in job obviously dictates the management much more, patients who had extensions into the shaft, a a different pattern of injury. And the only other serious inclusion criteria, important inclusion criteria is that patients who couldn't provide follow up data Yeah, were excluded.

So patients with cognitive impairment dementia and so on, couldn't take part. But obviously surgical intervention in that group is relatively rare. Yeah, absolutely. Absolutely. So. And, and, and, and Matt, just before we move on, so, and as we've already mentioned, these are patients, like you say the, the fracture was reduced closed, and it was in theatre, and that's the point at which they were randomized.

That's correct isn't it. Absolutely. So in the, in this study we, we learned a bit from the first DRAFFT trial. So I think surgeons were understandably cautious about randomizing before they did attempted reduction. Yeah. So we wanted them to be able to be confident they achieve a reduction before randomizing to the two interventions.

And obviously if one of the interventions required a surgical implant that, wires in this, in this case, that had to be done in the operating [00:07:00] theatre environment. So all the patients were randomized in theatre once the reduction had been achieved. Yeah. Most of them under general anaesthetic, some under regional blocks and and so on, but all in an operating theatre environment.

Excellent. Yeah. Great. So if we just move on before we talk about the results, just briefly about the sort of important data that collected for this type of study. So how was sort of data on resource use collected and, and how did you sort of sort of assess and evaluate evaluation of costs was done?

Sure, yeah. So there's two key elements of the data that inform the cost effectiveness evaluation. Obviously the, the outcomes for the patient and also the resource. We, the resource use that we collect is as broad as possible, but the primary analyis is based on what NICE look for, which is the National Health Service and Personal Social Service Perspective.

That's what the economists call it. So what does that mean? It's basically what the taxpayers paying for. So everything in the NHS and everything that's paid for through social services in [00:08:00] the, in the community, we then add onto that what it costs the patients. So time off work and so on. So a broader societal perspective.

But the primary analysis is based on the, the NICE recommendation for the perspective that NICE recommends, which is about the NHS and personal social services. Yeah. In terms of how we collect the data, the within hospital data was collected by research nurses on the, in the wards and in theatres and so on.

So time in surgery. How many physiotherapist appointments they had, so on all of that was collected through the research nurses. And then the extra resource use was collected from the patients directly at three, six, and 12 months afterwards through a questionnaire where they say, have they required any extra home help?

Have they required any other support at home? Have they been off work? And if so, for how long? And all of those data. Yeah. Great. And as I know we've said before, Matt, and people who've been involved in trials, the health economic data collection is often the, the vast majority of a, of a, an out [00:09:00] of a booklet, isn't it really?

In many ways it is. It's, it's huge. So just for everyone, so they're aware that, you know, we micro cost all of the resources, so. Not only do we ask what medication you've had, but it's how many pills for how long, and then we cost each pill. So it's quite a big burden for the patients to fill us in. And that's important to understand because that affects obviously completion rates and how much missing data we have in there, which I guess we'll talk about Andrew.

Yeah, absolutely. Absolutely. And you've sort of mentioned it already in terms of the importance of sort the, the valuation of health outcomes and obviously we, we, we, people will have heard QUALYs and quality of life years again, but could you sort of very briefly before we move on, sort of give patient listeners a background to sort of the NICE effectiveness cost thresholds and how these were sort of employed within this study?

Yeah. So the, the primary outcome measure for the, for the clinical trial is actually patient reported wrist evaluation. So pain and function at the wrist, [00:10:00] but you can't use that score and compare it across healthcare systems for different conditions. So the way that NICE asked us to collect outcomes from clinical effectiveness, so patient centered outcomes is through quality adjusted life years.

And we do that to a general health related quality of life questionnaire. We used something called the EuroQol 5 Dimension which is the most commonly used, certainly in Europe and actually increasingly around the world to collect quality of life data. And it's basically five domains of health. So pain can you look after yourself?

Daily activity, living anxiety and depression and so on. So it's a very broad measurement and that allows NICE to compare and contrast cost effectiveness across different conditions. So you can cost up someone having a heart attack versus someone having a stroke versus someone having a broken wrist and compare the effectiveness of different interventions.

So that's why we, we use that particular outcome tool. Yeah, absolutely. And in terms of the thresholds that NICE often use that they're sort of 20, 20,000 pounds is often the, the [00:11:00] quoted for equality. Is that, is that right? Yeah, so the two bits of information, then you've got the resource use, so basically what it costs.

Yeah. And then how much benefits in terms of QUALYs, quality adjusted life days you have. And then you can calculate a ratio for each intervention of what it costs per QUALY gained from that intervention. And the, the new intervention, if you'd like. Usually more expensive, but not necessarily is compared to the standard of care.

And then the difference in the cost per QUALY gives you this ratio that allows you to judge cost effectiveness. So the economists talk about incremental cost effectiveness ratios, so probably won't go into any more detail in that. No, no. And then each country, each healthcare system has to decide what it's willing to pay for one quality adjusted life year improvement.

And that's where these thresholds come in. And NICE sets this at 20,000 pounds per QUALY, roughly. And that changes a little bit depending on the economy and so on. Within the study we did what we call sensitivity analysis, where we, the, the primary analysis was based on 20,000 pounds per QUALY. Is it cost [00:12:00] effectiveness at that threshold?

But we also did a, a, a subsequent analysis using 30,000 pounds. Or 15,000 pounds or higher or lower value. Yeah, and that's hopefully to make it a bit more applicable toward the healthcare systems that might pay a bit more, or indeed a bit less for an improvement in in quality adjusted life. Absolutely. Yeah. No, that's brilliant Matt.

I think that gives the, listeners a really good overview of how, you know, sort of breaking down those nuances and, and those, and those terms that we use. So moving on to the results of the paper just very briefly for our listeners. 500 patients randomly allocated in the trial. 255 received a cast, 245 K-wires.

After manipulation of their fracture, the mean age, you know, in the two groups of 59 years and 61 years respectively, which, and the majority of trial participants were female and had an extra articular fracture very consistent with standard sort of a distal radius fracture, epidemiological data. So first on Matt, we've already mentioned it briefly, but before we get onto the sort of the nuances, of the results, you mentioned the word like completion rate.

What was the completion rate of the health outcomes and resource use [00:13:00] data for this, for this study and, and how did you sort of deal maybe with missing data. Sure. So there's always, because of the detail we're asking patients to complete there, there's always some missing data. Now, quite remarkably 58%, so over half of all the patients give us a complete data set.

So every single question was answered, which is quite amazing considering how much detail we're asking here. So that's a testament to an awful lot of work from the trial team to, to chose patients shop and ask them to fill out these questionnaires. The way we deal with missing data though, is we, with the primary analysis is using imputation, which I won't go into detail on that mostly because I'll get myself in trouble with the economist.

But it basically you take repeated samples from the data you have got and run those samples so they're independent trials and you do that thousands of times and that gives you an overall estimate to how much effect missing data would have on the actual overall outcome. Yeah. I've not explained that terribly well, but it's probably, probably the best I can do that.

So that's, that's how we [00:14:00] deal with missing data. It's built into the primary analysis. And you can do that with clinical outcomes as well, using the same bootstrapping techniques. But actually it's most, very commonly it's the usual way of doing things for a health economic analysis certainly in the UK.

Yeah, no, absolutely. I like for many, this is, like you've already said, you know, almost 60% of patients having every field completed is, I know. It is remarkable really, in terms of the amount of data they've got to fill in. And that's actually, like you say, that's every time point, every data point. Whereas, you know, you might have some at three months, some at six months and, and very, a variable you know, return rates at that time, won't you?

Absolutely. So yeah, some patients. Three months, but not at six, and then reply again at 12. Yeah. Yeah. And so the data's monotonic, I think is the term that is the word become misused. But yeah, so we, if the any patient gives us any data, we always include it in the analysis basically. Absolutely. So, If we go into sort of the key findings of the results briefly, so what did you find, you know, the key things in terms of the healthcare and the personal service use rates and [00:15:00] costs?

Sure. Well, as you probably expect most of the resource use is driven by the initial hospital episode. So what it costs to actually have the, the surgery and then have the, the immediate care in the hospital. So whether it's day case or overnight stays. Follow up was also quite a, a strong resource driver.

So the amount of physiotherapy people had, the number of outpatient appointments they had and so on. And then a really important item was the secondary surgery. So we micro costed or anyone had to have a further operation further down the line. And what it basically chose is that it's, it's slightly less resource use, even with one in eight patients requiring the second operation.

Putting a cast on is slightly cheaper than it is to actually treat people with wires, but only a tiny amount really. On the quality of life side, there was no difference essentially at any time point. So really nothing at all there. So you've got the same QUALYs and a marginally cheaper intervention overall for the NHS and social care [00:16:00] services.

The differences out of the hospital in terms of social care use and time off work and the science perspective were, were minimal. There was very little difference there. So the, the small cost differences were driven by the, the actual hospital care. Mostly the, the first visit. So when they had the surgery.

When they had the surgery, absolutely. And it, and if you then take that forward in terms of the health utilities and cost effectiveness, what did that show? So the ratio really shows that. So at a willingness to pay of 20,000 pounds per quality adjusted life year. So the NICE threshold, it's unlikely that wire fixation is cost effective.

There's only a 24% probability that it would be cost effective. So that's the headline figure is that it's, it's unlikely that wires are a cost effective intervention for this particular problem, but there's some nuance associated with that. Yeah, absolutely. No, we'll come onto them. So I think we'll come to them now actually, cause I think that's a nice way to lead into it.

So, you know, if we sort of draw that all together, you [00:17:00] know, they, I think the strengths of the study are, you know, are clear, huge, large UK base, multi centre, trauma trial, excellent design and execution as always. And, you know, a comprehensive collection assessment of resource use and costs associated with it.

Very cl, very, very nicely done. And obviously very clearly robust analysis, which you can see from the study itself and the paper. And, and like you say, you know, with. Within this sort of within trial economic evaluation of draft two, it it indicates that, like you say, manipulation, KY fixation is know it's slightly more, more costly and doesn't really result in a significant increase in the QUALYs in comparison to a moulded cast

and therefore it's unlikely but to be of benefit. But, you know, I think like there's a few nuances to it. And, and what, what, what, I suppose what were, what those nuances you'd want to highlight to the listeners, Matt, in terms of some of the things we've mentioned already. So with any health economic evaluation, you, you write an evaluation analysis plan.

So the same way that with any trial, we write a statistical analysis plan to say upfront what we're gonna do with these data and how we're gonna [00:18:00] adjust it so everyone knows that what, what we're planned to do and make sure we stick with the plan. And we do exactly the same thing with any economic analysis.

We say what analysis we're gonna do, and then we work from, from that, that template. So essentially we're playing a very straight bat. So what's reported in the paper is the absolute, what we said we do, and these are the results. So it's not likely to be cost effective to use wires in this situation, but the absolute cost difference is, is marginal.

It's about 13 pounds, I think, for the patient. So if you take the patient perspective, I'm going to an operating theatre and I need to have something done, they will accept that if that's the recommendation. And they agree with that , if they've gotta go back again though, that's quite a big deal. So it might be marginally cheaper for the NHS, but if I'm a patient and I'm in an operating theatre environment and for the sake of 13 pounds, you can reduce my risk of coming back.

Well, frankly, I probably want the wires put in. Yeah. Yeah. And so for me, the outcome of this [00:19:00] was that I would probably, if I can, if I have to be in the operating theatre and I'm doing a closed reduction and I've achieved that, then I would, I'm gonna put the wires in to save one in eight of my patients having to come back again.

Yeah. So there's the headline results of the trial and then what that means in clinical practice. Yeah. The really important bit of information though, is if you can hold them with a cast, they will get the same functional outcome. Yeah. So then the question is, if we did the cast outside of an operating theatre environment, of course will be hugely cost effective.

So you've reduced the cost dramatically if you do it in the plaster room in outpatient department. But that wasn't the intervention in the study. So we're extrapolating from there and everyone, all the listeners have gotta decide whether they think that's a reasonable extrapolation of the, of the study, but it's certainly changed.

That's where my clinical practice has gone and my colleagues in Oxford, and I think lots of the other hand centres are moving that direction as well of attempting a closed reduction and close contact cast outside of the [00:20:00] operating theatre. And then if you're going to theatre, then it's probably better for the patient overall to have some sort of fixation.

Yeah. Of the risk. I think that's really interesting about, and a really nice way to put it. Like you say, it's sort of like you can always look at just the headline result of the trial, but actually there's how you adapt that to our day to day clinical practices. Do you have a feel of what the patients would want, you know, in terms of Yeah.

You know, because I mean, you know, I think you could get a variety of responses there in terms of somebody say, I, I want it reduced, but I don't want any wires in it or, or vice versa. I just want the definitive treatment now and I don't wanna have to come back . So, no, absolutely. It's really interesting and we, we've done, we've done a lot of patient experience.

Yeah. Work across loads of trials for quite a long time now. And one of the very consistent messages, whether it's a broken leg or a hip or a wrist, is that the only people that think surgery a good idea is surgeons. So patients generally wanna avoid going to an operating theatre. It was interesting some of my colleague Dan Perry's work with the children is that the parents on behalf of their children, that's an even

bigger [00:21:00] issue cause handing your child over in an anaesthetic room and then being escorted out of the room and leaving them there is a huge issue for the patient. So avoiding surgery is massive, so they will accept all kinds of things you would think, you know, empirically, they wouldn't want child to have a wonky looking wrist.

Yeah. But if it avoids an operation and there's a hope it will remodel the patient's happy to take part. So it's the same with the broken wrist that patients who, particularly the older patients who are the predominant, you know, patient group for distal radius fractures, were not afraid of surgery. So they said if it needs doing, it needs doing, that's fine.

But if we can avoid going to an operating theatre, then that's what we want. That's empirically what we want. And that was a really clear message. When we ask the patients about their interpretation of the trial, and so that appears in the patient facing materials quite, quite clearly that they, they want to avoid surgery if they can.

Yeah, absolutely. I think that's really interesting. And, and just a, a final few questions in terms of, you know, you know, we've talked about the strengths, you know, are there any sort of limitations to the study or anything you [00:22:00] would've done differently in, in hindsight maybe? Yeah, so we, we made a choice up front to do the randomization in theatre after the reduction was achieved.

And that committed us to being in the operating theatre. So I, I guess that's the key thing really. Now, the win from that was that we include the, the surgeons were already happy about the reduction before they were , so that's clear. The downside is, of course, that you drive up the cost by just being in an operating theatre.

Yeah. So we have ummed and arred about whether we need to repeat the study. Looking at manipulation and casting in, in the fracture clinic to see whether that's actually gives you equal result. The difficulty is that since we've started doing that in Oxford, and I know you guys have in in Edinburgh as well, and, and lot sense using finger traps and having a bit of patience and getting the plastic techs to put a cast on.

Most patients don't even require analgesia, nevermind nerve blocks or anaesthetics and things. So I'm struggling now a little bit to know whether ethically I would [00:23:00] be able to randomize patients to go into an operating theatre versus having that manipulation done in in the plaster room. And we're not talking about haematoma blocks in ED, which I think frankly is torture.

But it's still done a lot. I know, but it, it's really not a pleasant thing to witness, nevermind to do. This is a controlled environment with people who do it every day, and I think that's the way forward. So, I mean, we've got a protocol written for that trial actually to look at doing it in the operating theatre versus outside.

But whether we can ever do that, whether the community thinks that's acceptable, I'm not sure. We'll have to discuss that over the coming months when the trial's really settled down and people have digested it. Absolutely. I suppose that would be the, that would be the, the trilogy though, wouldn't it?

That'd be DRAFFT3 potentially. Well, don't worry. We've got DRAFFT3, that's another one. In fact, DRAFFT3 is just about to start. Excellent. This is for the patients who haven't haven't needed a manipulation so, if the patient's been undisplaced something with displaced fractures, we're gonna randomize to a, a cast versus a a removal splint.

[00:24:00] Yeah. And then discharge the patients in a removable splint to see, to check they're okay. So it's a noninferiority trial. Coming a hospital near you very soon. Absolutely. And that'll be fascinating cause if one of the real other things we've learned a lot from the patient feedback, and particularly during the pandemic, was they don't wanna be in the hospital unless they absolutely have to be.

So a pathway of care that allows them to manage themselves at home, remove their own splint. It's something they're very keen on, but only if it's safe. And that's what we need to ab. Absolutely. But I think you're right. I think actually, and I think empowering the patients to, to lead their care is, is not a bad thing in many ways.

I think some people do need our help and intervention, but actually I think we've assumed for a long time, they do always need to come back to us. And that , that's just not necessarily the case. Yeah. And it's, they're just inconvenient for them. And it's more for our benefit maybe than thiers sometimes, but no.

Well, I think that's great. I'm afraid that's all we have time for today, but so thank you so much for joining us always, and congratulations on another outstanding study that's again, further added to the literature in this area. And it was always great to have you with us. [00:25:00] Thanks Andrew. And to our listeners, we do hope you've enjoyed joining us and we encourage you to share your thoughts and comments through social media, like feel free to tweet or post about anything we've discussed here today.

And, and thanks again for joining us. Take care, everyone.