Introduction

Trauma is a very common EMS response. According to the National Emergency Medical Services Information System (NEMSIS) database, inuries accounted for 14.6% of EMS calls in 2012 [1]. Despite advances in the management of injuries over the last several decades, trauma remains the most frequent cause of death for Americans aged 1–44 and the fifth largest overall cause of death nationwide. Nearly 120,000 Americans died from unintentional injuries in 2010 [2].

Improvements in trauma care require research. Many of our beliefs and practices over the years have changed as a result of research that has challenged conventional medical wisdom. There are many examples of trauma therapies that were previously used but later were no longer considered to be the standard of care following research. These include pneumatic antishock garments, aggressive fluid resuscitation, and certain forms of airway management. Our recent wars in Iraq and Afghanistan have provided a new venue for observational trauma research that has yielded valuable insights. Of note, many of these trauma therapies can be used both in military and civilian trauma settings.

Prehospital trauma study types

There has been great interest in the last decade in providing evidence-based care in the EMS setting. This requires ongoing research efforts to establish a knowledge base that will support reliable conclusions regarding optimal EMS care and the development of valid and verifiable clinical treatment guidelines for providers.

Similar to any clinical research, prehospital trauma studies can be of a variety of types [3]. These include (in descending order of rigor):

• prospective randomized clinical trials (RCTs) (double-, single-, or unblinded)
  
• prospective/retrospective observational/cohort trials
  
• retrospective reviews/database analysis/case–control studies
  
• case series/case reports.

The “gold standard” in clinical research has been the prospective randomized double-blind clinical trial. In this type of study, patients are randomly assigned to one treatment arm or another and the outcomes of each treatment group are analyzed. If the clinical trial is designed so that the only difference between the two arms is the treatment modality being randomized, it makes it possible to know if the therapy provides patient benefit. The “blinding” ensures that the patient, the researchers, or both do not know into which group each patient is assigned. This blinding minimizes bias that could change other ways in which the patients in the two study groups are treated.

Randomized and blinded trauma clinical trials are difficult to conduct in the EMS setting because the consent and randomization procedures in this emergency situation are often not feasible. One early example of a prehospital trauma RCT is the 1994 study by Bickell et al. looking at the effects of the administration of IV fluid in trauma patients [4]. In this trial, hypotensive trauma patients who had sustained penetrating trauma were randomized on even days to receive standard infusions of IV fluid and on odd days to receive no IV fluids until mechanical control of hemorrhage was achieved in the operating room. In this study, consent to participate was “implied,” therefore informed consent requirements were waived, greatly simplifying enrollment in this study. (This approach to consent is no longer permitted in the United States.) A more recent example of a prehospital RCT is the 2009 study by Moore et al. of human polymerized hemoglobin in the treatment of traumatic hemorrhagic shock patients in the EMS setting [5]. This trial randomized patients to be resuscitated with either crystalloid or Polyheme, and was conducted using an exception to informed consent as per the federal regulation 21§CFR50.24 [6].

Cohort/observational studies provide clinical information despite the possibility of treatment and/or observational bias. In these studies, there is no “active” treatment intervention being randomized and evaluated. Instead, one or more groups (cohorts) are defined and patient outcomes evaluated based on the use of specific treatments. It is important to note that cohort/observational studies can be either prospective or retrospective. In a prospective cohort study, the study parameters, patient cohorts, and outcomes are defined before each group undergoes the treatment. In a retrospective observational study, the information being studied has already been collected but the study questions and outcomes of interest have not been established. In order to minimize bias in these retrospective studies, the researchers must define the patient cohorts and study outcomes prior to analyzing the database. Retrospective studies, although more prone to bias than prospective studies, are often easier to conduct because the data already exist, for example in a state trauma registry. An example of cohort analysis in trauma is a study by Holcomb et al. [7]. In this multicenter study, several groups of trauma patients receiving various transfusions of blood products were followed and their outcomes analyzed.

There are multiple databases available for analysis to EMS researchers. Because of the popularity of electronic patient care records, many EMS agencies now have databases available for research analysis of the patients to whom they have responded. Many states require hospitals to report to a state-wide trauma registry. In addition, many state EMS offices collect EMS data and provide a state-wide view of EMS activities. The NEMSIS database is a national compilation of EMS data, supplied by participating state EMS offices, and offers the broadest overview of EMS nationally.

While these databases can be a very valuable resource for trauma research, great care must be taken in study design to ensure valid conclusions can be drawn from the data. Database analysis can be prone to bias if the study parameters are not strictly defined before analysis takes place. “Data mining” is the analysis of a database for patterns prior to the formation of a hypothesis and strict study parameters. Although care must be taken to avoid unsupported conclusions, data mining can assist in developing hypotheses for future clinical trials if any observed differences in patient outcome are believed to be based on treatment differences.

Case–control studies compare two groups of patients whose outcomes differ. This research format is especially useful when the adverse outcome of interest is observed infrequently. The case group is matched to the control group only on a minimum number of variables, so that it is possible to examine which variables differ in the two patient outcome groups. These studies may quickly identify a characteristic or treatment that imparted the observed difference in patient outcomes between the cases and the controls. This format was used in the study by Shayne that examined the need for angiography in penetrating thigh trauma [8].

Case reports often are the starting point for the analysis of a case series or the initiation of a case–control study. The most important aspect of these reports is either the observation of a new disease or complication, or the potential benefit of a newly attempted therapy.

Data gathering and analysis of EMS patient care are performed as part of the quality assurance (QA)/performance improvement (PI) process needed to improve patient outcomes. In this process, a new protocol, medication, procedure, or device is implemented by the EMS agency after review of current medical literature and approval by its medical director, training officer, and other senior administrative personnel. After training of the EMS providers, this change is implemented and the outcome of the intervention is monitored as part of the QA/PI process. There is no control group, although sometimes the outcome may be compared to that of “historical controls,” a similar group of patients treated differently prior to the initiation of this new therapy. When studying the effects of a new therapy, it is often necessary to wait some period of time, after which it is likely that the new therapy being reviewed for quality has been fully integrated into the EMS system of care. This non-studied “wash-out period” helps to remove any inaccuracies that may influence the study outcome as a result of incomplete or inaccurate adoption of the treatment being studied as the system and providers become accustomed to it. This QA/PI process allows for the implementation of therapies with proven efficacy in the research setting. The subsequent QA/PI monitoring of patient outcomes allows EMS providers to compare current outcomes to those observed when other therapies were provided in order to establish that the new treatment approach enhances patient outcomes.