Problems With EMS Research

Every now and then, we get the latest news about some “groundbreaking” study involving EMS or even tangentially connected to EMS.  The older I’ve gotten (and hopefully a bit wiser as well), I’ve come to take most of these studies with a grain of salt.  I present to you a couple of reasons why.

First, most of these academic studies are done where the academic medical institutions are — typically big urban areas with ready access to full-service teaching hospitals. When you have a five minute transport time to a facility where every specialty of medicine is present and each attending physician goes by “Professor,” there is something to be said for rapid transport being treatment.  That goes even more so when your large urban EMS system is comprised, in part, of providers who have little interest in the clinical practice of medicine.

Next, most EMS research seems to have almost a fetishistic focus on resuscitation.  I’d surmise that the reasons are twofold.  First, the research outcomes are easy.  Unlike The Princess Bride, there’s no mostly dead.  It’s either dead or not dead.  It makes for easy endpoints in research.  Second, since American EMS was initially based on accidental trauma and out of hospital cardiac arrest, the resuscitation interest seems logical.

As such, we have ended up with some of the research questioning the value of paramedic-level providers.  And it’s true to an extent.  A BLS provider is perfectly capable of transporting a traumatic injury to a trauma center for the one intervention that matters — surgery.  And in the majority of cardiac arrests, BLS interventions of chest compressions and defibrillation are what matters.  And doubly so in both instances if you live in a large urban area with a plethora of hospitals.

But here’s what little of the research addresses.   Firstly, suburban areas, much less rural or frontier areas, where transport to definitive care is not measured in single digits.  In some rural and frontier areas of my home state, the highest level of medical care available in the county after normal business hours is a paramedic.  In these cases, there’s undoubtedly a benefit to providing the advanced level of care that a paramedic brings — but because there’s no “science” to definitively support this assertion, there’s skepticism at best about providing advanced level care. Additionally, some paramedic level interventions defy easy measurement.  Unlike the outcome of cardiac arrest – dead or not dead – it’s a bit more challenging to quantify and measure symptom relief. “Sir, earlier you described your ingrown toenail pain as 12 out of 10. Now that you’ve had some pain relief, what’s the pain level now?”  Meanwhile, the femur fracture patient says they are fine and don’t need pain management.   Similar measurement difficulties can occur with other medical crises such as respiratory distress and chest pain.   It’s hard to quantify subjective measurements.

What would go a long way toward providing meaningful EMS research for the rest of EMS would be to develop some research consortiums and studies that occur outside large urban medical care systems.  Factoring in the distance to transport to definitive care and removing the maniacal fascination with cardiac arrest resuscitation would be a huge step in providing meaningful data to EMS providers in suburban, rural, and frontier EMS systems.  One thing we know about EMS is that one system solution doesn’t fit all.  It’s time that our research agendas reflect that reality as well.


  1. Bob Kellow says

    The fixation on cardiac arrest research can be attributed to the fact that the population studied requires no control group. They all enter the study dead. Then, the dead are culled by the population of dead patients who are least likely to be resuscitated. There’s your study group.

    There have been few randomized controlled trials (RCT’s) on EMS because it’s next to impossible to create a control group for whom treatment will be withheld. It’s inhumane. Another problem is trying to control for all of the variables that could skew the findings.

    Finally, research, when done properly, is very expensive and time consuming. Rural and frontier areas always draw the short straw because of their population density, ergo lack of adequate trial base and too sparse a population for predictive sampling. I’m not encouraged that there’ll be a sudden groundswell of research interest anytime in the foreseeable future.