Economic outcomes associated with deep surgical site infection in patients with an open fracture of the lower limb

Listen to Mr Andrew Duckworth interviewing Professor Matt Costa about his paper "Economic outcomes associated with deep surgical site infection in patients with an open fracture of the lower limb" published in the November 2018 issue of The Bone and Joint Journal.

Click here to read the article

Transcript

[00:00:00] So welcome to our listeners, to our inaugural podcast from your team here at The Bone & Joint Journal. I'm Andrew Duckworth, and this is our first production so I thought I would just briefly explain what we hope these new podcasts will add to your reader experience. 

Our aim and hopes are that we will improve the accessibility and the visibility of the studies that we published for both you as the readers, but also for our authors as well. During the next 15 minutes or so when some of you I'm sure are busy commuting to and from work, or even on your morning run, we hope to discuss a range of aspects for your chosen study, emphasizing the important points of how the work has been designed as well as the key findings from the study and how these potentially fit into each of your busy day-to-day clinical practices. 

We also hope to give you a behind the scenes insight into how the authors have developed the study conceptually, delivered it, and also give them the opportunity to put forward the key findings that they think are important.

So today I have the pleasure of being joined by Professor Matt Costa, from Oxford to discuss their study, entitled Economic outcomes associated with deep [00:01:00] surgical site infection and patients with an open fracture of the lower limb which will be published in the upcoming edition of the BJJ. So a big welcome to Matt, and thank you for taking your time to be involved with our first BJJ Podcast.

Hey Andrew, its a pleasure.

Great. So obviously the management of open fractures has continued to move forward in the past few years, as you obviously note in your paper with the advent of major trauma networks and previously published guidelines. So to kick us off, could you give us a brief background to the paper and what made you look at this particular topic in relation to open fractures, which utilises data, obviously from your recently published WOLF trial.

Sure. So  the trial itself, WOLF was really looking at a particular type of wound dressing. So the intervention that has been tested were two times wound dressing, but with these large clinical trials and WOLF was really unique and it was the first major trauma trial to come through the UK major trauma network. So it really represented the whole national perspective on the management of particular injury. In this case, open [00:02:00] fractures. And the trial allowed us to collect research quality data by which I mean data with very complete followup, prospectively  collected within the context of the trial. So it gave us an insight into open fractures more generally beyond just the interventions actually in the study, in the trial  itself.

And just for the  listeners, so what sort of patients were included in the WOLF trial? Obviously there were lower limb fractures, but was it sort of a wide range? Just to give them a concept of who involved? 

Yeah. Well, the trial was really about the most serious open sort of fractures. So these were the ones where the surgical team and nearly all of these patients were treated by joint orthopaedic and plastic team as per sort of major trauma network guidelines. So these were the injuries where the surgeons felt that at the end of the first wound debridement, the wounds could not be closed primarily. So, the higher energy injuries with the biggest soft tissue damage. So the most severe injuries you can have really. 

Okay. And in terms of obviously this, [00:03:00] the main purpose of the paper that we're going to discuss is about deep infection. Just how was that defined? Obviously I think you used the CDC definition. And was that just done basically by the clinicians who reviewed the patients? Was it? 

So the infection data, and it is a really important point because one of the problems with talking about infection is indeed the definition. Bizarrely, most papers that report the confection following any surgical intervention don't actually report the definition they used, which is pretty, pretty worrying really. So, in this study all of the data regarding infection was collected independent of the *inaudible*. So the patients were reviewed after 30 days independently by research nurses trained to assess the wounds. We took photographs, made clinical assessments and reviewed the records of that stage. And then patient-reported symptoms at three, six, nine and 12 months as well. And if the patients reported any symptoms, so it could be related to a wound healing complication, [00:04:00] we went back to the clinical teams and the primary care teams to check what those symptoms really were and what the treatment was. And from that information, we extracted the CDC definition of infection for the patients.

The primary endpoint was 30 days cause that's what CDC said at the time of the study, but we also collected the infections up to one year as well, which is also according to the guidelines. 

So it seems like there were sort of multiple ways, you know, you're covering all the bases in terms of making sure that that definition for deep infection was pretty clear.

Yeah, we tried to. I mean, it was a major effort and we're really still breaking new ground certainly in the context of open fractures, trying to collect this data in a systematic way. And it's not straightforward. So we learned a lot along the way, but I think we probably did the best we could in terms of really working out which patients had infections.

Yeah. It's what is feasible, isn't it? Yeah, absolutely. And so obviously the primary outcome of the study was looking at costs wasn't it and in particular the cost associated with the deep infection. And you also mentioned [00:05:00] in the paper that there were two sets of long-term economic consequences which were measured within WOLF, sort of preference-based health-related quality of life, outcome, and overall economics cost. So can you just give the listeners just a brief description of that because the analysis is obviously quite complex in the papers. Isn't it? 

It is, it is. So there are different forms really with health economic analysis associated with clinical interventions. And the one that the national Institute of clinical excellence in the UK recommends is cost per QUALY. So essentially that means taking all of the costs associated with managing the injury. And this is not just what it costs the NHS in terms of the initial treatment in hospital that also includes NHS costs afterwards. So all the followup or the rehabilitation for the surgery, interventions, but also the cost to society more broadly, so the personal social service perspective. So that includes any benefits taking them home. And we also assess costs according to the patient's personal point of [00:06:00] view. So time off work and so on. 

So actually within the context of the trial, the great majority of the data by far the biggest data set was actually the resource use that allowed us to generate the costs associated with injuries in the study.

The second facet is then the quality of life and we used two health-related quality of life measures in this study. One's the EQ-5D and the other is the short form 12, which are just two ways of collecting similar but subtly different information about general quality of life. And by plotting quality of life changes over the year after the injury, we were able to work out the quality adjusted life year, the QALY .Then it's a simple matter of dividing the costs associated with injury by the quality of life of the patient over the year after the injury. And then you can compare the cost per QALY of those patients with infection to those patients without infection. So that's the gist of it. That's a whistle-stop tour of what we were trying to do. 

The actual way that the [00:07:00] analysis is done is, you know, as you say, reasonably complicated for the average orthopaedic surgeon, including myself. So, do you want me to talk a little bit about the actual regression modeling? 

I don't think so. I think you've done a great job there of explaining. I think it's just getting an idea of the two separate things in terms of, like you said, the QALY and then the  overall healthcare costs. I'm sure as you say it, anybody who's involved with clinical trials whenyou. Collect the data, all the data that, the thing that collects the most data is the health economic analysis often. Isn't it? Cause it's just the detail of which is collected is quite phenomenal really. 

Yeah, absolutely. That is by far the biggest part of any of the major trial datasets as you know well. And collating that data is not straightforward. So yeah, it's a testament to the clinical trials unit team really, the amount of effort that goes into following patients up and checking the goal, that information. 

Absolutely. And sort of just sort of moving on to the results of the study, so you had 460 patients that were there, included 35 who had a deep infection, which is about a rate of about 7 to 8%. And looking at it, the completeness of the data [00:08:00] is quite remarkable really. There were two patients that there wasnt complete data on is that, is that right? 

Yeah.  So complete data in terms of the infection we had very powerful follow-up. You're always inevitably going to miss some elements of the resource use data. So there was far more missing data in resource use. Nearly all the patients, in fact pretty much all the patients gave us some information about costs, but not everyone filled out every single detail and resource use. So for example, we asked patients not just that they have medication after their wounds, so painkillers or antibiotics, and so on. We asked them what dose and for how long, and that's a big burden for patients to give us all that information. So inevitably we lost some  detail. And therefore we used a technique called  multiple imputation to actually account for missing data within the analysis.  I think in the context of open fractures, the follow-up is, as you say, quite remarkable because the patients who have these severe injuries and not necessarily the patients who are going to...

Absolutely.

 Young men who [00:09:00] ride motorbikes, don't particularly fill out bits of papers as a  general rule.

Absolutely. Absolutely. I mean, that's the thing, and I think it's a good example of how using, you know, good complete trial data on a further analysis after you've done the initial trial can be very useful cant it because you have such a complete dataset from that original study. 

Well, I think that's right. I mean, you know, people find increasingly that the best quality data we're getting is actually from the big multi-centre trials, simply because we have the resources in the trial context to collect very complete data in a way that it's almost impossible in routine clinical practice. So, yeah, I think we're, you know, there's a new...there's a trend certainly towards using this data in different ways to look at some sort of observational studies, as well as actual trial questions. 

Yeah. So just looking at the, sort of the patient population, there are obviously slightly higher rates of infection seen in diabetes and patients with smoking and smokers as well wasn't it? Which, you know, again, you know, it's obviously something we know from the literature and intuitively, but it's actually good to see it in a big data set like that that's [00:10:00] shown there is a difference there. But looking at the health-related quality of life, can you sort of summarize that in terms of the costs and how it changes between those with deep and without deep infection.

Yeah. So, theres loads of information here, I mean, it's really fascinating stuff. I mean, the first thing to say, the really good news is that the rates of deep infection found within WOLF in the context of the UK Metro trauma network is probably the lowest ever reported for these most severe injuries. So 7.5% on average across the 460 patients in this study is testament to some fantastic work going on across the country in terms of trying to reduce the rates of infection. I think if I had to guess the biggest factor is the joint orthoplastic work in, so the collaborations between orthopaedic and plastic surgeons to really debride these wounds and plan the surgery from day one, I think is probably the biggest reason why.

 In terms of change in quality of life, well that's where it gets a bit more depressing really because we often [00:11:00] as surgeons think that at three months, if the wounds healed or the flap has taken, and there's a bit of callus on the x-ray well, you know, we're pretty happy, our job is done. This looks like it's going to heal. But what the patients were telling us that in terms of quality of life, even a year after the injury, there was still showing quality of life reduced by about half. So a 50% total loss of quality of life, which, just to put it into context is equivalent to having a stroke. So massive effects on patients and in the longer term as well. So that is a major worry and I think certainly I, as a clinician, and I think many people underestimated quite how debilitating and how much of a general effect on life, these terrible injuries have. So overall throughout the study patients recovered obviously after the initial treatment, but we're still reporting significant losses after the injury. 

Yeah. And in terms of the cost, the costs obviously with infection and we have data about infection after sort of heart arthroplasty, but it includes something you [00:12:00] think infection costs money, but it's quite stark the costs that you've presented there. Theres some very high figures.

 There are. Yeah. So the cost of just treating these injuries is, you know, around 15,000 pounds per patient and they go up considerably. But the really interesting thing I think about this data is that actually, I was quite surprised how little the costs in general went up associated with infection compared to what we know as you say from the arthroplasty literature. And I think we tried to explain why this was, and I think it is because if you, having infection per se is not the issue, having, you know that to some degree is inevitable after these injuries, no matter how good the care or having early successful treatment is the key. So if you have an infection and you have very early debridement, you have deep samples, antibiotics and then, you know, definitive treatment and you get a resolution of your infection. The costs associated in the longer term are then not particularly greater than those patients who never had an [00:13:00] infection in the first place. However, if you have chronic osteomyelitis, then you've got lifelong costs and all the costs associated with amputation, both from a patient perspective and also from a, you know, resource use perspective. So actually it's not infection that costs money its the long-term consequences, not treating it aggressively and fully at the beginning. And that's probably the key message from this paper. 

Yeah, no, I totally agree. I think it's all those extra bits. I think you put it quite well there is that you think once you've got the wound healed and the fracture healed, then you think, you know, the job is done, but there's so much more after that isn't there that happens to the patient and the healthcare costs associated with that.

There is and we wrote about this recently me and a colleague of mine, wrote about the patient experience from a qualitative point of view recently. Sorry very far away from this health economics paper Andrew, but the long-term emotional consequences, psychological consequences injuries are huge. And I think as surgeons we underestimate that. I [00:14:00] think we're increasingly aware of the problems, but actually how to manage them, how to support patients is. 

Yeah, absolutely. Yeah. And then in a busy practice, these things are hard to sort of take time don't they? That's the thing. 

Oh completely. Yeah. I mean, the content, you can't imagine the worst place for your psychological wellbeing and then a busy fracture clinic. 

So sort of moving on for a general overview of it. So obviously the stance of the study in terms of the size and the quality of the prospectively collected data are without question, you know, and the economic analysis is very good. So in a nutshell, what do you think are the clinical findings, but also what do you think, are there any limitations to this study that you'd like to talk about? 

So the key takeaway messages are that in the context of a fully functioning major trauma, national major trauma network you can reduce deep infection  rates down to 7 or 8%, even after the most serious open fractures, which is fantastic news. If patients have infection, the costs go up, [00:15:00] but if you treat those infections early and aggressively and you get resolution the costs in the longer term are manageable. You know, they're not particularly greater than those for patients without the injuries. However, if you have the long-term, the deep infection, osteomyelitis problems then that's when the costs really spiral out of control and also the consequences for the patients spiral out of control.

In terms of limitations then, because the infection rate was low, which was obviously fantastic news for patients you have fewer events within your analysis. So we actually got the 35 actual deep infections to work with in this dataset, which, you know, we produce it. We're less precise in our estimates of the costs associated with these injuries because the numbers are lower. Having said that deep infection after open fracture, there's not many studies that have this sort of level of followup. And so it would probably, it's certainly a useful contribution to the literature, even if it's not the final word.

No, no, I agree. [00:16:00] I think it's not really limited limitation, but I think what would be of interest is actually the longer term outcome, because obviously these patients are still, even a year there's quite a big impact there. You'd be interested in two to five years, if that impact is still there really.

Yeah, they're all in long-term followup at the moment. We're going to follow them until at least five years and then see what the data looks like. I don't have that data for you yet. 

No of course. Absolutely. I think it probably, I suspect, it might shock us, in terms of their impact on their quality of life potentially.

Yeah. It's a worry. I mean, if at one year, I'm pretty pleased myself if, as you say, the wounds healed, the bones healed, then thinking happy days, but actually the patients, that's not what they were reporting. So I think you're probably right. I think we probably underestimate the effects in the longer term.

Yeah. And just a sort of final question, you know, in terms of the research on open fractures, in general, moving forward, what do you feel are the areas where we should move into next? 

Well, I think particularly related to this particular  study is really about reducing the rates of [00:17:00] infection. And so I think theres work to do with implant design and surface typography and so on... coatings perhaps, looking at reducing infection rates, but I think probably the bigger win is in the reducing the variation in clinical practice, because although major trauma networks undoubtedly are going to step forward in the management of these serious injuries, there's still considerable variation around the country in terms of how quickly we're able to get defensive wound. What types of flaps are used and there's research to be done there.  And exactly when is the best time to get these injuries closed? I think there's still lots to be done that we could improve in open fractures. 

Yeah. I think like you say, the improvement that's been made already you feel with the right infrastructure it certainly is achievable isnt it? 

It is. It is. I mean, it's huge. And it was interesting when we presented this work abroad, people were kind of interested in the study and the infection rates and so on, but actually how the network functioned was the major source of [00:18:00] interest. And I think probably no great overestimates to say that it was jealousy from a lot of countries, you know, including very well developed health care economies in North America and Australasia, were actually quite jealous of the facilities we have and assistance we have in place now in the UK.

So, yeah, I know it's pretty good news, I think for UK, sort of trauma care in the NHS in particular at the time when it's not always great news, actually to show a success story through major trauma networks is pretty important for people. 

Absolutely. So, Matt, thank you so much for joining us for our inaugural BJJ podcast and congratulations on another excellent study.

Well, thanks very much, Andrew. It's great pleasure.