Tuesday, July 31, 2012

Trying Hard - Trials in EM Research

The Gist:  If one follows EM blogs, podcasts, or tweets for any period of time, it's immediately apparent that application of evidenced based medicine is the expectation.  Within the EM context, however, research and evidence face incredible challenges, partially due to the improvisational, desperate nature of the job.  Global discourse over TXA in trauma and epinephrine in cardiac arrest highlight learning points regarding the complexity EM research.  Note: This was part of a behemoth post, so this is Part 2 (Part 1 takes a gander at the CRASH-2 trial and under-utilization of TXA in the context of study design/result interpretation).

The papers behind many of these musings:  Effect of adrenaline on survival in out-of-hospital cardiac arrest: A randomised double-blind placebo-controlled trial. and Prehospital Epinephrine Use and Survival Among Patients With Out-of-Hospital Cardiac Arrest.  
  • On the PHARM Podcast, Episode 12, when Dr. Minh Le Cong and Tim Noonan highlighted some problems with trials in EM with medications like epi and they do a great job reviewing some of the issues below.  
What makes evidence based medicine different in EM?

Diversity of the Population - ED patients often have trauma or an undifferentiated illness/complaint that causes them to present to the ED.  Trauma, naturally, is not uniform and certainly not predictable in terms of enrolling patients. 
  • Mechanism of action -  There are seemingly infinite ways in which people manage to harm themselves or others.  Every study that investigates trauma will have some degree of variation (heterogeneity) in penetrating and blunt trauma, burns, etc.  Thus, inclusion criteria may rely on a clinician's judgment or an objective systemic endpoint (BP, HR, etc).
  • Predictability - Within a population, estimates exist regarding the incidence/prevalance of the disease.  One of the beauties of EM is never knowing what type of patient is going to roll through the doors next.  Sometimes we can guess (or hope), but that's about the extent.
Cost - In EM, many therapies are relatively inexpensive:  chilled saline, epinephrine (epi) , and TXA are all fairly cheap.
  • Pharmaceutical companies are not inclined to sink significant funds into examining the efficacy of these therapies.  In stark contrast, some therapies like activated protein C (Xigris, drotrecogin alfa) seems to get a great deal of journal/media attention for an intervention with limited clinical utility and no proven benefit.  Activated protein C,  far more expensive than drugs such as TXA/epi, has a clear pharmaceutical backing drawing more research and publicity.
Vulnerability of the Population  - The day a patient requires emergency care is often the worse day of their life.  
  • Oftentimes, one is unable to obtained informed consent for study participation due to the nature of a patient's injuries, mental status, or the critical nature of events.   The final rule on Exception From Informed Consent (EFIC) in the U.S. waived the need for informed consent in EM research in 1996. Note:  The Nov 2005 volume of Acad Emergency Medicine details this topic thoroughly.  Although in emergent situations, informed consent may be waived legally, IRBs often have issue with approving these studies.  This rule does not apply internationally.  For example, the authors of CRASH-2 reported that some of the delay in TXA/placebo administration existed in the delay in obtaining informed consent.  
  • The population is comprised of pregnant women and children, two groups that experimenters are already reticent to include in trials. 
Unique Environment - A certain MacGyver-esque quality exists in EM where there's some agreement that nearly anything (reasonable) may be done in order to save a life.  Due to the oft undifferentiated and improvisational nature of the trade, it's much more difficult to go through many of the standard research protocols.   
  • Time proves critical in many EM situations.  Thus, a physician's impression of a situation (or gestalt) may be the only thing guiding critical care, as some diagnostic queries may take too long.  This is especially notable in the time frame that TXA works best (within 3 hours of trauma, but <1 hour is superior).  
  • Along these lines, a fraction of EM practice incorporates salvage therapy.  Individuals may have a lower threshold for utilizing a treatment due to the notion that the intervention is a "last-ditch effort."  The use of epinephrine in cardiac arrest is an example in which many people have questioned, "what can be the harm if the patient is already dead?"  Unfortunately, epinephrine's primary benefit seems to be in establishing ROSC so the patient's family (and the patient's body) then bears the burden of an alive patient with little to no neurological recovery.  Certainly some patient's fall into the area where they can have both ROSC and retained neurological function (therapeutic hypothermia!); however, the studies don't look promising that this is positive at the population level.  It's extremely difficult to refrain from doing something.  
  • The population is (generally) very sick.  Many cohorts have confounders or confusing secondary endpoints where survivor bias plays a significant role.  TXA's failure to significantly reduce blood transfusions in CRASH-2 may serve as an example of this.
  • Much of EM is protocol driven, particularly in the pre-hospital realm.  Thus, study design may need implementation/approval at the system level.  The initial study in Houston, TX regarding permissive hypotension in the penetrating trauma patient is an example of a successful study with a protocol based intervention. 
Standard of Care - The current dialogue regarding epi in cardiac arrest serves as an excellent example of how a standard of care, widely accepted prior to rigorous studies and modern research design, requires one to assume the burden of disproof.
  • Placebo controlled trials may be difficult to get past an IRB with something that has become a "standard of care," even if the "standard of care" has never been adequately investigated.  
This paper, by Dr. Tim Coats in the the BMJ, discusses the challenges of conventional research in the EM realm and advocates for emphasis on pragmatic trials over explanatory trials in EM.  It really is helpful.

No comments:

Post a Comment