List Archives

Mon Nov 3 23:15:22 GMT 2008

From: dburgess at mhg.com
> here is one study I have on file.

--> Thanks for that.
I started to have a look at it... This caused a few ideas to enter my little brain...

1. "CONCLUSIONS: The therapeutic shift to POD has substantially reduced transfusion requirements and length of hospital stay, and has reduced mortality in patients with unstable pelvic fractures..."

"From the Department of Surgery, University of Tennessee Health Science Center, Memphis, TN."

"Early in the study series, EPF was used... Later in the series, the POD was placed

Judging by the dates included in the publication, this study looked at cases from as far back as 13 years ago. It is RETROSPECTIVE with no mention of randomisation, sample size calculations, etc (unless I missed these in my quick scan through). And yet, from the conclusion above, it seems that these great benefits have been DIRECTLY ATTRIBUTED to this ONE SINGLE change - the introduction of POD!

Wow (with a capital "W")!

I am sure we've all noted the confusion between causation and association!

But things look really strange even if we were to assume that POD was the hero of this fairytale... Are we really expected to believe that this great institution in Memphis introduced NO OTHER CHANGES into the management of these trauma victims during a whole decade, while the rest of the planet was inventing stuff left, right and centre? Is there anyone else here who has not seen anything new in the management of hypovolaemic shock since the mid-90s?

To put it another way - THERE IS evidence that with ALL OTHER FACTORS equalised, most trauma centres use LESS blood in the mid-00s than they used in the mid-90s for the same clinical presentations! Can this not be what caused the LATER (POD) group to be given less blood?

2. "Early in the study series, EPF was used... usually performed in the operating room... Later in the series, the POD was placed immediately on recognition of the unstable pelvis..."

i.e. It seems that the EPF group had a DELAYED application (had to get to OR first), while the POD was placed early - could that not account for some of the benefit? This confounding variable seems to have been ignored.

3. Additionally EPF was thus put ONLY on patients deemed "bad" enough to get to the OR. POD was applied to patients, some of whom were NOT bad enough to go to the OR - so the study groups DID differ on this major aspect...

4. "After pelvic stabilization, additional hemodynamic instability mandated laparotomy for patients with a positive ultrasound or grossly positive lavage..."

i.e. in both groups surgeons made attempts to use surgery to deal with bleeding AFTER THE STUDY INTERVENTION, but they get no credit for this! Any reduction in bleeding is attributed to the POD, even though the patients it did not fix got surgery, which would (one hopes) have affected the outcome.

5. "Although there was decreased mortality with POD (26%) versus EPF (37%), it was not statistically significant... CONCLUSIONS:... POD ... has reduced mortality"

Actually the first statement indicates that it FAILED, upon statistical analysis, to show a mortality reduction of any significance! An insignificant positive is not a "slight positive" or a "nearly positive" - it's more of an "I gave it this much opportunity to show a positive benefit and it did not". So this is a false conclusion.

6. In comparing the 2 groups, we are given a SBP reading of 112.5mmHg in the POD group and 101.6mmHg in the EPF group. I ASSUME that this is the average BP of each group. Whether a 10% difference is significant or not is less of a matter here. What we really want to know is how many patients had a BP which APPEARED TO REQUIRE TREATMENT in the eyes of the surgeons. It is likely that there would be significantly more patients with SBP that LOOKS LIKE SHOCK in the EPF group! Imagine the 2 distribution graphs - the means are only 11mmHg apart, but how much of the tail of the EPF group is below what the surgeons thought was "shock"? THAT is what we should know! THAT is what would lead to transfusion!

Oh, look! Later on we read "...Patients treated with POD had notably fewer resuscitation transfusions despite the equivalent severities of shock on presentation" - but, as shown, they only had SIMILAR-ISH AVERAGE SBP READINGS, but that does NOT prove they had "equivalent severities of shock"

I'll stop here. I stopped reading the article at this stage.
Apology in advance if I have missed anything important by not reading to the end, but I generally stop reading articles once I decide that they cannot be used to guide my practice. In other words - "I've seen enough".

I do not mean to be critical of EVERYTHING in this article - much work seems to have gone into the study. But I must admit to have acquired a bias against it early on. Why? Well, it was over a 10-year period. With no infomation to the contrary, it appears that about halfway or so through that period, the institution switched from EPF to POD. What reason did they use at the time to justify this major change from what must have been standard treatment? What justification? And what would this decision look like if the results of the analysis, about 5 years later, showed that POD was NOT better? Would it still have been published? "Look guys - we changed to this POD thingie and our analysis now shows that it caused us to lose this many more patients since then while giving them more transfusions..."

If someone really wanted to evaluate the benefit of POD, they could have processed these two groups simultaneously over a few years, with randomisation...
_________________________________________________________________
Discover Bird's Eye View now with Multimap from Live Search
http://clk.atdmt.com/UKM/go/111354026/direct/01/

List Archives

For your convenience, find the Trauma List Archive Below

Pelvic fractures