Page 56 - Journal of Special Operations Medicine - Spring 2016

Figure 2 Time to effectiveness results by junctional Figure 3 Ranking point results by junctional tourniquet
tourniquet model. model.

Mean time to effectiveness (in seconds) differed by model of junctional
tourniquet. In pairwise comparisons, the SJT and JETT had the fast-
est mean times to effectiveness; the difference between the two was
not statistically significant. The CRoC had the slowest mean time to
effectiveness, in part because the testing required CRoC assembly be-
fore each use. Each column height represents the mean, and each bar
length notes the standard error of the sample.
The first method of analyzing ranks used the rank
number to generate points for all models assessed, and The chart shows that by all available rankings, users most preferred
the SJT and AAJT. Preference for junctional tourniquet model was an-
results differed by model (p < .05). The Steel–Dwass alyzed by all rankings of each model assessed (most preferred, points
nonparametric analysis of such ranking indicated that = 4, best) for each user. The SJT and AAJT were most preferred. The
users’ preferred junctional tourniquet models stratified top and bottom of each box represents the 75th and 25th percentiles,
respectively; whiskers represent the range, and the line in the box rep-
into two tiers—a first tier of one pair of models and a resents the median value of the distribution. Since the JETT was so
second tier of three models. The first tier included the often ranked with 2 points, the box plot collapsed to a line as the box
two most preferred models: the SJT and AAJT. The top and bottom were overlaid with the median at 2. The CRoC had
only one medic ranking it as best.
second tier included the least preferred models: AAJT,
JETT, and CRoC (Figure 3). The reason the AAJT was JETT, and CRoC). The only pairwise comparison that
in both groups was that because it had so few tests, its was significantly different was SJT–CRoC (p = .047;
ranked variability was too high to differentiate it from p > .12 for all others) (Figure 4). Thus, for the second
other models. The SJT–JETT and SJT–CRoC compari- method, the SJT, AAJT, and JETT were most preferred.
sons were significantly different (p = .0013 and .01, re-
spectively; p > .5 for all four others). Thus, for the first Results by Training Group
method, the SJT and AAJT were most preferred. An unplanned opportunity for an after-the-fact analysis
arose by circumstance for comparing effectiveness per-
The second method of analyzing user preference was centages by training group. Since this was not planned
to assess only the highest-ranked model (most pre- ahead of time, we only made comparisons using descrip-
ferred) for each user instead of analyzing all rankings. tive methods. The division of training involved the first
Such ranking differed by model stratified into two tiers: nine users versus the final five users, because they were
a most preferred tier of three models (SJT, AAJT, and trained at different times. Only one of the first nine us-
JETT) and a least preferred tier of three models (AAJT, ers had 100% effectiveness in every test irrespective of

Table 2 Ranked Preferences of Junctional Tourniquet by Model
User Preference by Model Score* by Model
Rank † Score CRoC AAJT JETT SJT CRoC AAJT JETT SJT
1 4 § 1 ‡ 4 2 7 4 16 8 28
2 3 7 0 0 7 21 0 0 21
3 2 3 1 10 0 6 2 20 0
4 1 3 4 2 0 3 4 2 0
Sum 14 9 # 14 14 34 22 30 49
*Score for each tourniquet (number of assessments × score).
† 1 = best; 4 = worst.
§ Points assigned to the highest rank of 1.
‡ Number of users who gave this tourniquet a rank of 1 (best).
# Only nine users ranked the AAJT; the other five users did not rank the AAJT because they did not use it.

40 Journal of Special Operations Medicine Volume 16, Edition 1/Spring 2016