Before CA, i rated this paper: 7/10
1
Objectives and hypotheses
1.1
Are the objectives of the study clearly stated?
Yes. It is to look at the ability of the USS and CXR to detect traumatic PNXs. CT was the gold standard.
2
Design
2.1
Is the study design suitable for the objectives
Yes. it is a diagnostic study, albeit a difficult one to do in a perfect fashion.
2.2
Who / what was studied?
Patients were recruited from September 2003 to May 2004. Inclusion criteria were age >17, blunt trauma, and receiving a FAST scan, AND going on to CT scanning. Quite a select group in fact. This is bad in some respects as it is a spectrum bias. But at least they all had the gold standard (CT).
They did exclude patients in whom they could not do the scan. So this was not an intention to diagnose study.
2.3
Was this the right sample to answer the objectives?
Probably. It is clearly a spectrum of patients at the more serious end of the spectrum as this was a level 1 trauma centre in the US, and patients were going to get a FAST scan so it implies a higher level if injury.
2.4
Is the study large enough to achieve its objectives? Have sample size estimates been performed?
No sample size calculation was performed. Confidence intervals are reported in the results and these vary by up to about 10%.
2.5
Were all subjects accounted for?
All included patients were accounted for. Not surprising when they excluded everyone else!
2.6
Were all appropriate outcomes considered?
I think so. It was for the detection of PNX. They could have reported other findings such as haemothorax, but you get the feeling that we may see another paper on that subject from the same authors in the future :-)
2.7
Has ethical approval been obtained if appropriate?
It was waived by their institutional review board
2.8
Was an independent blinded gold standard test applied to all subjects?
Yes. CT was the gold standard, or a rush of air on chest tube placement. These seem reasonable.
Placement of the chest tube was almost certainly not blinded. Radiologists and EPs were blinded to the CXR results when interpreting the CT and USS.
3
Measurement and observation
3.1
Is it clear what was measured, how it was measured and what the outcomes were?
Yes, the sliding lung sign is well described. As it is a dynamic measurement it would have been better to see a video to really understand the concept, but there is enough information here.
3.2
Are the measurements valid?
Yes. It reflects an anatomical structure.
3.3
Are the measurements reliable?
probably
3.4
Are the measurements reproducible?
Unknown from this study.
4
Presentation of results
4.1
Are the basic data adequately described?
Not really. The patients are only briefly described in reference to their chest findings. There is no table 1 of demographics for us to decide what the patients were like or if they reflect our own practice. The PNX rate of 53/176 is quite high suggesting a relatively severely injured spectrum of patients.
4.2
Are the results presented clearly, objectively and in sufficient detail to enable readers to make their own judgement?
Hmmmm not so sure that this is as good as it could have been.
The results are presented a little oddly in that the text suggests that USS and CT have the same pick up rates (30%). In fact the USS had one false positive and one false negative. I would have preferred to see a 2x2 truth table rather than the confusing text and the summary statistics given in table 1.
The correlation assessment of PNX size using Spearmans correlation is probably not helpful. They would have been better assessing this with something like a kappa score.
4.3
Are the results internally consistent, i.e. do the numbers add up properly?
Seem to.
5
Analysis
5.1
Are the data suitable for analysis?
Yes
5.2
Are the methods appropriate to the data?
Simple descriptions of diagnostic data and confidence intervals is OK. They could have compared the overall results of the USS and CXR with something like a McNemars test. This would have told us if CXR and USS were statistically different, the data certainly suggests that this is the case.
5.3
Are any statistics correctly performed and interpreted?
As far as they go yes. Correlation is probably not a good measure (it will almost always be reassuring in these studies) better to measure agreement (kappa in categorical or bland altman in continuous data)
6
Discussion
6.1
Are the results discussed in relation to existing knowledge on the subject and study objectives?
Yes
6.2
Is the discussion biased?
Not really. They harp on a bit about the size of PNX. I am not sure how much more that adds as in trauma a PNX is a PNX to a large extent, the treatment is going to be the same pretty much regardless of size in the patients described.
7
Interpretation
7.1
Are the authors' conclusions justified by the data?
Yes
7.2
What level of evidence has this paper presented? (using CEBM levels )
Can never answer this one.
X3 BMW???
7.3
Does this paper help me answer my problem?
It makes me want a USS machine. Yet more reasons why it might be helpful in the resus room for very early diagnosis (pre CXR - which they will need anyway)
After CA, i rated this paper: 7/10
8
Implementation
8.1
Can the test be implemented in practice?
Yes.
8.2
What aids to implementation exist?
Increasingly widespread use of USS in UK EDs
8.3
What barriers to implementation exist?
Money Money Money
8.4
Are my patients the same as the patients tested?
Some are.
8.5
Will the test improve diagnosis in my patients?
It might speed diagnosis in the resus room. This could be very useful and clinically important.