(Side note: This is the second paper I have reviewed that has Jon Studnek as an author. He's a paramedic who also has a PhD, and is faculty at Carolinas Medical Center. He writes a lot, and I probably could fill all of my posts with reviews of his publications.)
|AKA "Dr. Medic"|
The investigators used data from the 14 hospitals in North Carolina that acted as the receiving centers for STEMI patients transferred for emergent percutaneous coronary intervention (PCI).They looked at sub-groups of patients, broken down in three different ways:
- Patients who first presented to a PCI center, or to a non-PCI center;
- Patients who had the cath lab activated by EMS, or in the ED; and
- Patients who used EMS, or those who "walked in" to the ED.
Another aspect of the study that gives it some "real-world" applicability is how they defined an "inappropriate" activation of the cath lab. While other authors have described a "clean cath" as an inappropriate activation, the authors acknowledge that there are many scenarios where PCI for presumed STEMI is appropriate, despite the 20/20 hindsight of a negative cath. Takotsubo cardiomyopathy, for example, often requires an emergent angiogram to clarify the diagnosis.
So, instead they defined an inappropriate cath lab activation "If catheterization was canceled because of ECG reinterpretation or if the patient was deemed not to be a candidate" for PCI. Clinical factors, such as age or DNR status, were used to determine candidacy.
The overall results, comparing paramedics and ED physicians were that 15% of activations were inappropriate:
They analyze the results further, breaking down the data into the subgroups described above. The group of interest is all the patients who were transported by EMS, and had their initial activation by EMS. In other words, none of these patients were "walk-ins," but it included both patients who were brought to PCI and to non-PCI centers (initially).
They compare these activations against all the patients who had cath lab activation performed by the ED physicians (both at PCI centers and non-PCI centers), with patients who either came in by EMS or car.
There are a couple different ways to analyze these results, but overall the physicans performed better that the medics. Well, 7+ years of training ought to pay off somewhere, and and incremental accuracy in ECG interpretation is a reasonable expectation.
However, you can't even conclude this from the data presented, since an activation may have been deemed "inappropriate" because of a patient's DNR code status, say, or severe comorbidities (e.g. sepsis, or terminal disease). Specifically, we don't have the break-down for ECG accuracy versus judging cath lab candidacy for the 2 groups - it may well be the case that medics are just as good as emergency physicians at reading ECGs, but the physicians are better at judging which patients actually warrant an emergent catheterization.
The last table emphasizes the point that, while this sort of study is great at generating statistically-significant results, there is a lot of "granularity" that is not accessible to us.
Clearly, not all EMS agencies or EDs are equal - some systems are better than others. In this table, note the range of appropriate activations:
On the other hand, it is concerning that some EDs, even at the big hospitals with cath labs, have a "false-positive" rate of 25%. Similarly, some EMS agency inappropriately activates the cath lab 1/3 of the time!
The Bottom Line
This isn't a study that you can use to change your clinical practice in the next shift. It isn't even very useful at changing practice at your EMS agency or ED. However, it points the way to doing the more practical research, by highlighting important aspects.
For example, how do paramedics at different agencies decide to activate the cath lab, and how do these methods correlate with accuracy? Could a closer look at the 65% - 100% range in appropriate activations suggest a "best practice" for EMS? Should we rely more on intensive continuing education for paramedics? Alternatively, should there be more emphasis on computerized and/or human algorithms for ECG interpretation?
Furthermore, since the "Not Cath Lab Candidate"group accounted for such a large proportion of the inappropriate activations (4.3%), might their be a better way to anticipate this exclusion? To a large degree, the cardiologist is the individual who is deciding the patient's candidacy for the cath lab, and it is often difficult for the emergency physician, let alone the paramedic, to anticipate their decision. I'm not sure that the accuracy of prehospital STEMI activation should be judged using such "soft criteria."
So, more research is called for, as usual. But this paper serves as a very useful guide for the future.