Page 68 - Journal of Special Operations Medicine - Fall 2014

was CRoC, AAJT, JETT, and SJT, based, as noted ear- Figure 1 Effectiveness percentages by model of junctional
lier, on their date of FDA clearance for the inguinal indi- tourniquet. The SJT and the CRoC had the highest
cation. The combat uniform was worn, although boots effectiveness percentages; their difference was not statistically
and socks were removed during testing. Testers had a significant. The JETT and the AAJT had significantly lower
5-minute rest period between test iterations. Hence, all effectiveness percentages than the SJT and CRoC.
tests were conducted on a given subject within a 3-hour
time period. The test location was a work room of the
Department of Combat Medic Training at the US Army
Medical Department Center & School.

Test results included effectiveness percentages, time of
application, and ranked preferences. After testing, users
ranked tourniquet performance subjectively. 16,17 Users
ranked their preference of the tourniquets by answering
the question, ”If you had to go to war today and you could
only choose one, which tourniquet would you choose to
bring?” The rank, R, was a whole number ascribed by a rate was attributed to the fact that the AAJT hurt so
user to a given device (rank = 1, “best”; 4, “worst”). The often and to such a degree that the simulated casualties
rank was converted into a score (score = 5 minus rank). commonly stopped the iteration of use early before ef-
The users’ scores by model were summed. For nine users fectiveness was attained.
with 4 points allotted to the best rank, the best possible
score was 36 and the worst possible score was 9. Of the nine users, only two were able to make effective
use of each of the four models of junctional tourniquet
Statistical testing included repeated measures analysis of in one or more tests; however, 16 of the total 19 inef-
variance (ANOVA) to see if any device was different from fective tests were made with use of the AAJT. For all 72
the others (SAS Institute, Cary, NC). Pairwise compari- tests, the average effectiveness rate was 74% (53 of 72).
sons were adjusted using Tukey’s method. Comparison of
proportions such as effectiveness percentages was made The effectiveness rate for the left side was 69% (25 of
with χ test using SAS (SAS Institute, Cary, NC) and MS 36), and the effectiveness rate for the right side was 79%
2
Excel 2003 (Microsoft, Redmond, WA). Descriptive sta- (28 of 36). Although the order of testing was always left
tistics were used to portray results. Significance for results first and right second, so that the users had fresh experi-
was established when p values were < .05. ence before they used the junctional tourniquet models
on the right side, the difference in effectiveness rate by
side (left versus right) was not statistically significant
Results (p = .643).

Safety Results Time to Effectiveness Results
Based on the definition of safety used in the present Because the AAJT hurt to such a degree that the simu-
study, all tourniquet uses were safe in the absence of lated casualties stopped early the iteration of use, the
adverse events during the 3 hours of testing. The four AAJT was removed from further statistical analysis of
models of tourniquets were equally safe. time to effectiveness. For the three remaining models,
the times to effectiveness varied by junctional tourni-
Effectiveness Results quet model (p < .003; Figure 2). Both the CRoC and
The effectiveness percentages varied by model of junc- SJT models (which did not differ [p = .090]) were more
tional tourniquet (p < .003). Effectiveness percentages rapidly effective than the JETT (p ≤ .008). Average time
by junctional tourniquet model were statistically strati- to effectiveness by side (left versus right) did not differ
fied into two groups with a pair of models of tourniquet (p = .094). However, the left side, being first in order
in each group. The most effective junctional tourniquets of testing, took longer (average ~150 seconds) than the
were the SJT (100%) and the CRoC (94%), which did right side (average ~90 seconds; data not shown).
not differ significantly from each other (p = .187; Figure
1). However, the CRoC was more effective than both the Subjectively Ranked Performance Results
JETT and AAJT (p < .001), and the SJT was also more Preferences of users for junctional tourniquets were dif-
effective than both the JETT and AAJT (p < .001). Dif- ferent (p < .001). Users most often preferred the CRoC
ferences in the effectiveness of the JETT and the AAJT and SJT over other models (p < ANOVA probability) but
were not statistically different (p = .991). The effective- had no preference between these two models (p = .187;
ness rate of the AAJT was 11%; this low effectiveness Figure 3, Table 2).

60 Journal of Special Operations Medicine Volume 14, Edition 3/Fall 2014