At St.Emlyn’s we recognise that there are a lot of journals out there. Far too many to keep a track of and far too many for one human to read and collate (thereby providing evidence of Weingart and Reid possessing super powers as they are the only people who seem to manage). Anyway, I digress. The point is that if you want to keep abreast of the literature you need to be pretty savage about filtering what you are going to read. For me this means screening papers in roughly this fashion.
- Is the title interesting (to me)?
- Is it relevant to my practice?
- What is the aim of the paper?
- Can the method answer the aim?
- Read paper
- Review paper
(Ed – should we do a post on how to keep up to date, or does everyone know how to do this already?)
This process serves me well and means that I focus my time on papers that are going to make a difference to my practice. However, there are a few exceptions to my rules on filtering papers. One of these is when there is possibility that a paper might come back to bite me even if the methods are not up to scratch. So this week I was running through the journals and came across an interesting pre-publication in Injury on the management of displaced ankle fractures.
Time to apply the filter…..
- Is the title interesting (to me)? – YES
- Is it relevant to my practice? – YES
- What is the aim of the paper? – To look at whether ankle fractures should have a pre manipulation X-ray – YES – apparently we should X-ray displaced fractures before manipulation – WOW – REALLY? DOGMALYSIS ALERT!!!!
- Can the method answer the aim? Nope..
- Read paper
- Review paper
So we get to stage 4, and we will explain why below…, so why read on you ask? Well I did because I am a little concerned that this attempt at dogmalysis might be used to change practice. So let’s get right back to the paper.
Just before we do though……let’s get a quick reminder of what we’re talking about (so long as there are skateboards this will be an issue).
[DDET What’s the issue here] For as long as I can remember the management of displaced ankle fractures involves the urgent reduction of the deformity to prevent or alleviate neurological or skin problems. I even remember this turning up in my med student finals so I’m fairly sure that this is embedded knowledge in the culture of EM. It’s not just EM either, I spent a year as an orthopod in my youth (long story I’m much better now) and vividly remember the ortho team getting jolly cross if they even saw a pre-reduction film of an ankle fracture dislocation. In this paper the dogma of immediate reduction is questioned. Do we really need to rush about reducing fractures in the ED and are there potential benefits to getting an X-ray in advance? Immediate manipulations might be riskier (in terms of sedation) and might be less likely to succeed. If that’s the case then we should stop and question our practice. Perhaps these are reasonable questions and so we should permit the authors to present their case. [/DDET]
[DDET Who and what was studied] This is a single centre study that retrospectively looked at 197 patients admitted to hospital with ankle fractures. 90 of these had manipulations, 31 of which were done prior to X-ray.[/DDET]
[DDET Is this a good study design?] There are significant issues here. The retrospective nature of the study will inevitably lead to bias. Let’s have a quick think about some of the more obvious ones….
- Documentation issues are always an issue in retrospective studies. We get a hint of this in the paper as the reason for manipuation in 25 of the 31 pre-Xray patients was not documented.
- It is highly likely that the patients who received manipulation pre Xray would be different from those who were manipulated post X-ray. An ankle pointing backwards is a arguably much more likely to be manipulated urgently than one with a small degree of talar shift. In other words the greater the deformity the more likely ED docs would proceed to immediate reduction.
- There are no descriptors in the paper as to whether the type of injury seen in the manipulated and non manipulated group is different. This is such a key issues I am amazed that it is not available.
- As we know almost nothing about the patients in each group we can make no meaningful conclusions about whether pre or post X-ray manipulation is an independent risk factor to the need for remanipulation.
[DDET …but they have results and statistically significant findings…..!!] Don’t be fooled folks. No statistical analysis can save poor study design, but since they are there let’s take a closer look.
- The remanipulation rate in the pre Xray group is 10/31 (32%)
- The remanipulation rate in the post Xray group is 7/52 (12%)
- This is described as statistically significant with a p value of 0.039
They have also re-analysed the data by removing the manipulations performed by the ortho team. This makes little difference to the results as only 9 of the manipulations were performed by the ortho team. There are a few other concerns. For example there are 32 patients in the graph of patients having manipulations prior to X-ray, but only 31 in the analysis. This does not add up and an additional patient not requiring manipluation would move the p value to 0.47. I would also argue that in my experience the ortho team is less likely to remanipulate their own attempts vs. the attempts of others….. it’s a circular and poor measure.
Remanipulation itself is an interesting outcome. It is important to patients, but it is not everything. From a patient perspective we should really have something clinically relevant and firmly patient centred. There are no functional outcome scores here and that means that I am unsure what this means for patients. There is simply little or no evidence here to help us.
[DDET The authors conclude that we should take pre-reduction X-rays though]
Well that’s interesting. The evidence presented here is poor. The research design is likely to be grossly biased and the findings themselves do not always add up and lack the the level of detail needed for reasonable critical appraisal.
It is simply not justified to make the claim that practice should change on the basis of this evidence, but here at St.Emlyn’s we do like challenges to our practice. We can find some common ground with the authors.
- Patients with clear neurovascular or skin compromise should be urgently manipulated.
- Patients with clear skin compromise should be manipulated.
In addition we at St.Emlyn’s believe that a displaced fracture with potential to cause any of the above should also be urgently manipulated. We also accept that some displaced fractures will not be obvious clinically and will be identified on X-ray only. Unsurprisingly, the minor displacements are easier to achieve good anatomical alignment, so this will almost certainly skew the results in favour of those that are not X-rayed before manipulation. [/DDET]
[DDET So just remind me…, why did we even bother to look at this paper?]
Good question, let’s go back to the beginning and our reasons for looking at papers. This paper failed the test on the basis of poor design. The reason we chose to look at this was because it is a really interesting and quite controversial subject. I especially have anxieties that this may be used as evidence in the future to not X-ray. There is no evidence here top change practice as this research design cannot answer the question of whether pre-manipulation X-rays are important.[/DDET]
THE ANSWER IS….?
To answer my original question then yes, in the opinion of the St.Emlyn’s team we should manipulate an ankle dislocation before X-ray when we believe it to be clinically indicated. If anyone quotes this paper to me as a reason to change then I will ‘learn them’. This paper changes nothing and confirms once again that’s it’s essential to critically appraise papers carefully. We simply cannot rely on abstracts to give us the full picture, and this paper has made me a bit cross (though probably because our field hockey team lost 6-0 today).