Despite anecdotal evidence seen on television and in movies, there is zero actual research to support penetrating traumatic brain injury (TBI) as an intervention to kill zombies (eg, walkers). For the moment, we shall dispense with the obvious conundrum that one cannot kill (end the life of) someone or something that is already dead. Rather, we shall focus on the concept of evidence.
Evidence is nothing more than a collection of facts. These facts may be minor or major and may stem from a variety of sources. A wide variety of models have been proposed to describe the hierarchy of evidence types. Of these various models, the pyramid shape is probably most familiar to nursing. The base and least rigorous form of evidence is expert opinion (eg, Dr Daryl Dixon reports that the bolt from a crossbow will kill walkers).1 It is important to note that the expert is often simply the person who happens to have the stage or the pen or to be a bully pulpit.
The next level of evidence is typically case studies and nonrandomized research. At this level, you may read a report by Grimes et al describing a case in which "...a knife thrust through the anterolateral skull proximal to Kocher's point resulted in death." Additional evidence comes in the form of nonrandomized research such as observational trials without control. For example, a group of nurses may collaborate to collectively clear a hospital wing of walkers via skull penetration with hemostats. Despite observing a 100% kill rate, there are key questions that remain unanswered: Is another method equally effective? How deep must the instrument penetrate (eg, dura mater, cerebral cortex, basal ganglia, or brainstem)? Does instrument size moderate the effect (do large diameter crowbars have higher odds ratio of killing compared with small bore arrows)?
The prospective randomized clinical trial is the next level of evidence and often used to test hypotheses that were developed from nonrandomized studies. For example, a randomized clinical trial may be designed to only stab half the walkers in the head and hit the other half with baseball bats (blunt nonpenetrating TBI) and thereby test the hypothesis that the proportion of walker death is statistically significantly higher after open TBI versus closed TBI. One could design a test of open TBI versus no injury (control), but this would probably not receive institutional review board approval because of an increased risk to the research staff.
The highest level of evidence is systematic review or meta-analysis. These articles synthesize and statistically summarize the accumulated data from all the available research. Intellectual honesty enhances the value of meta-analyses by including results from every study (those that reject the null hypothesis as well as those failing to reject the null hypothesis). For example, we would expect to read a systematic review that includes data from the highly successful Hilltop Study, mixed results from the Alexandria Effort, and results from the failed Sanctuary Trial.2
Thanks to Internet and high-speed computing, nurses can quickly and efficiently determine for themselves what evidence is available to support any given position. An October 2018 search of PubMed found 138 articles included the keyword "zombie." Of these, only 1 was listed as a clinical trial.3 Upon inspection, this was a study about the zombie entertainment industry and did not provide evidence for mortality.
Again, it can be noted that there are difficulties involved with evidence collection regarding the issue of the definition of death on a nonliving being. Perhaps, the limitation is that this is not an intervention with adequate preclinical data, or perhaps the theory is flawed. Because of the lack of research data, the only evidence available is expert opinion. Thus, we conclude that, if we were attacked by a horde of walkers, our best evidence-based practice option would be to use a skull-penetration intervention (class of evidence, level IIB; strength of recommendation, level C).
The authors declare no conflicts of interest.
References