The paramedic vs physician debate is probably as old as the respective professions and can be traced back to ancient Egypt.
Like many EMS folks, I’d say it is training, skills, experience and a supporting governance structure which is important, rather than the job title. However, much research still works on the physicians vs paramedics level.
When it comes to out-of-hospital cardiac arrest (OHCA), we now have what is considered ‘best available evidence’: A meta-analysis by Boettiger et al. The conclusions of this important work: ‘EMS-physician-guided CPR in out-of-hospital cardiac arrest is associated with improved survival outcomes’ (when compared to paramedic care).
Problem is, it is all observational studies. All observational studies are prone to confounding and bias, and when it comes to the value of EMS physicians, both factors almost always favor EMS physicians. Letting EMS physicians do research on the value of EMS physicians is about as objective as asking an Irish person to judge the best beer in the world…
So when you meta-analyse a bunch of confounded and biased observational study, you get… bias and confounding. As a counter argument, I’d like to do invite you to a narrative review of the papers which are included in the meta-analysis by Boettiger et al. I’d suggest you read the meta-analysis (it’s open access) and then the rest of this post.
I will go through each paper. Which is a thrilling prospect, I know. Bear with me.
We will go from largest to smallest study.
Yasunaga et al. Huge (n=95,072) study from Japan. It only included witnessed OHCA and indeed, physician care yields better outcomes than paramedic care, albeit at the price of also more patients surviving in a vegetative state. The authors point out that ‘hospitals with [prehospital] physician [treatment] typically provide more optimal post-return of spontaneous circulation treatments, including therapeutic hypothermia and percutaneous coronary intervention’. So the study that provides over 70% of the data for this meta-analysis is essentially a subgroup analysis, and we don’t know if survival benefit is due to EMS physicians or treatment at cardiac centres.
Hagihara et al., again from Japan, n=18,462. The same issue. EMS physicians associated with better outcome, but also with better hospital care. The authors state that findings ‘need confirming with consideration of in-hospital resuscitation data’. To be fair, ROSC rates are quite a lot higher in the EMS physician group, but is it really the outcome we want?
Kojima et al., from, you guessed it, Japan. Just over 4,000 cases. Details are limited, as this is a conference abstract. Of note: All of these three Japanese papers state they are using a national database. Data collection was from 2005-2007 (Yasunaga), 2005-2010 (Hagihara) and 2005-2008 (Kojima). I can’t say one way or the other, but there is at least the possibility of data source overlap between the studies.
Fischer et al. (n=4,000) is interesting, because it doesn’t actually compare EMS physicians and paramedics. It compares survival after OHCA in Bonn, Germany (which happens to have EMS physicians) and Birmingham, UK (paramedics). The authors also didn’t actually collect data for this but pulled survival rates from two previous publications. While survival rates after OHCA in Germany were higher, they also had shorter EMS response times. Of course there is also the issue of comparing different populations, EMS structures and hospitals…
Soo et al. (n=1,547) is a bit difficult to interpret, but basically survival after OHCA is more likely if a physician is present at the time of OHCA. It is slightly unclear how the physicians in this study ended up at the scene of OHCA, but it seems that these were OHCA occurring during GP reviews, public events or were a GP happened to be a bystander. Of course, all of these factors are associated with higher chances of survival. In addition the authors note the possibility of ‘[…] selection bias, but this is difficult to establish for two reasons. First, the medical practitioner may recommend termination of or persistence with resuscitation efforts, based on knowledge of the patient and the medical background; and second, Nottinghamshire Ambulance Service protocols require ambulance personnel to maintain resuscitation, once initiated, until arrival at hospital or unless otherwise advised by the medical practitioner.’
[I’ve run out of Ryu versus Ken screenshots. You get the idea. They are both the same, they have the same moves, strengths and weaknesses. Just different uniforms.]
Fischer et al. (n=1,146). Again. This time comparing survival to hospital admission (not discharge) between four different European regions. EMS in Germany and Spain were physician-staffed and had significantly higher rates of survival to hospital admission than did the paramedic-staffed EMS of the UK and USA. Two issues: Is ROSC a patient-centred outcome? This is a comparison between different populations and EMS structures, there are many differences other than the provider profession that could be associated with the difference in outcome.
Mitchell et al. (n=1,038). In the meta-analysis, survival rates to hospital discharge are quoted as 12.4% (Edinburgh, physician-staffed) and 7.2% in Milwaukee (paramedic-staffed). Other than the recurrent issue of comparing different populations, EMS and hospitals, I just copy paste from the abstract: However patients were more likely to have a witnessed collapse in Edinburgh 65.7%, compared with 25% (p<0.001) and significantly more of those patients received bystander cardiopulmonary resuscitation (CPR) 42.3%, compared with 27.1% (P<0.005). When these two effects are accounted for there is no difference in outcome.
Olasveengen et al. (n=973) adjust for differences in prognostic factors between their physician and paramedic group, and didn’t find a difference in survival. Arguably, the sample size is too small.
Frandsen et al. (n=393) compared survival following OHCA with a one-tiered BLS response (5%), a two-tiered BLS response including defibrillation (1%) and two-tiered physician-provided ALS (13%).In the non-physician groups, patients in OHCA were transported to hospital in cardiac arrest, rather than treated on scene. Anybody who has tried CPR in the back of a moving ambulance will be able to confirm that it is notoriously difficult and this practice has been stopped in most modern ALS-level EMS. The authors note that survival rates for the two-tiered BLS model in another Danish region is 18%…
Oshige et al. (n=342) from Japan. This is mind-bogglingly complex and my head hurts from trying to understand it. Basically, they tried to figure out if Adrenaline, Lidocaine and or Atropine is beneficial in OHCA. Because the Japanese BLS-technicians (not paramedics) were not allowed to give any drugs, the drug-giving arm of the study was done by EMS physicians. See round 1-3 for the issues of Japanese EMS physicians. Not surprisingly, survival was better in the physician group, as BLS-technicians essentially followed a scoop and run approach (see round 9). The authors note that ‘early restoration of spontaneous circulation rather than early transport, might benefit prehospital CPR.’
Yen et al. (n=158), not from Japan but from Taiwain. A cost effectiveness-analysis of EMS physicians. Survival in physician group was actually worse, but overall very small numbers.
Eisenburger et al. (n=118). This doesn’t actually compare physician and paramedic-staffed EMS. All OHCA patients in this remote alpine area received a first-tier BLS response, followed by a second tier EMS physician response. The comparison was between the first tier BLS-providers being able to defibrillate prior to EMS physician arrival (n=13) or not (n=105). Survival to hospital discharge rates are the same in both groups. Of the three survivors in the BLS defibrillator group, two died, resulting in a better one-year survival for the physician defibrillation group. Small numbers…
Hampton et al. (n=65). Even smaller numbers… This was published before I was born. And I can’t get the full text. I’m not even sure if this is relevant at all, I’ll just copy-past the abstract and let you decide for yourself. The survival to hospital discharge rates given in the meta-analysis are 3/19 (15.8 %) for physician treatment and 2/46 (4.3 %) for paramedic treatment.
“During a 20-month period a “cardiac” ambulance was manned on alternate days by specially-trained ambulance personnel only, or by such personnel plus a doctor. The presence of a doctor did not lead to any reduction in the mortality of patients with heart-attacks. Although transport to hospital by the special service was associated with a low pre-hospital mortality, this was balanced by a high pre-hospital mortality in the group of patients brought to hospital by routine ambulances at times when the special vehicle was manned, but for some reason was not used. There was evidence of unintentional selection of low-risk cases for transport by the cardiac ambulance. The number of lives saved by the special service was too small to cause any significant reduction in the overall mortality from heart-attacks in Nottingham.”
Dickenson et al. (n=49). Tiny numbers… Of the nine patients (!) in the EMS physician group, 44% survived to hospital discharge, 5% in the paramedic group. That is an absolute difference in survival of 39%! Such a tremendous effect of EMS physicians has never been replicated (see round 1-13), even in the most biased studies.
Congratulations, you made it this far!!
If you read the meta-analysis by Boettiger et al. and this post, you now have a good idea of how different approaches to reviewing the literature can lead to different conclusions.
The truth? I’d say somewhere in between. Maybe what matters more is patients, and working together.