Page 43 - Journal of Special Operations Medicine - Fall 2016

Figure 2 Time to stop bleeding, by tourniquet model. Figure 3 Blood loss, by tourniquet model.

The fifth and 95th percentiles are represented by the whiskers. Box The fifth and 95th percentiles are represented by the whiskers. Box
tops are 75th percentiles, box bottoms are 25th percentiles, the box tops are 75th percentiles, box bottoms are 25th percentiles, the box
cross-line represents the median, and the hashed line represents the cross-line represents the median, and the hashed line represents the
mean. TPT3 has mean and median lines overlaid. EMT, Emergency mean. TPT3 has mean and median lines overlaid. EMT, Emergency
and Military Tourniquet; TPT2, Tactical Pneumatic Tourniquet and Military Tourniquet; TPT2, Tactical Pneumatic Tourniquet
2-inch; TPT3, Tactical Pneumatic Tourniquet 3-inch. 2-inch; TPT3, Tactical Pneumatic Tourniquet 3-inch.

64 seconds for TPT3 and EMT, respectively). In pair- Table 1 Contingency Table of Composite Outcome Results
wise comparison, the mean differences of TPT2-TPT3 by Model
(33 seconds) and TPT2-EMT (47 seconds) were statisti-
cally significant (p = .0014 and p < .0001, respectively), Composite Outcome, No.
(%)
but the mean difference of TPT3-EMT (13 seconds) was
not (p = 0.33). Tourniquet Bad Good Sum, No. (%)
EMT 3 (10) 27 (90) 30 (100)
Differences in mean after time by model were not statis- TPT2 26 (87) 4 (13) 30 (100)
tically significant in one-way analysis because all three TPT3 16 (53) 14 (47) 30 (100)
models were in one tier. In pairwise comparison, the
mean differences (from 4 seconds to 8 seconds) of all Sum, No. 45 45 90
models were not statistically significant (p > .2 for all). EMT, Emergency and Military Tourniquet; TPT2, Tactical Pneumatic
Tourniquet 2-inch; TPT3, Tactical Pneumatic Tourniquet 3-inch.
Differences in mean pressure by model were statisti-
cally significant in one-way analysis; results were in two composite outcome count (i.e., score for TPT2-EMT [1.1]
tiers with the EMT being intermediate and falling into and for TPT3-EMT [0.8] were statistically significant; p
both tiers. In pairwise comparison, the mean difference < .0001 and p = .003, respectively; Figure 4), whereas the
of TPT2-TPT3 (6mmHg) was statistically significant (p difference in mean composite outcome count for TPT2-
= .0485), but other mean differences were not (TPT3- TPT3 (0.3) was not (p = .2).
EMT: 1mmHg, p =.9; and TPT2-EMT: 5mmHg, p =.1).
Differences in mean blood loss rate by model were sta-
Differences in mean blood loss by model were statis- tistically significant in one-way analysis: results were in
tically significant in one-way analysis: results were in two tiers, with the TPT2 solely in the low tier with low-
two tiers, with the EMT solely in the low tier with least est mean (5.6mL/s), whereas the other tier had statisti-
blood loss (320mL). In pairwise comparison, the mean cally higher means (TPT3, 7.3mL/s; EMT, 6.9mL/s). In
differences of TPT2-EMT (152mL) and of TPT3-EMT pairwise comparison, the differences in mean rates for
(147mL) were statistically significant (p < .0001 for TPT2-EMT (1.4mL/s) and for TPT3-TPT2 (1.7mL/s)
both; Figure 3). were statistically significant (p < .0001 for both),
whereas the difference in mean rates for TPT3-EMT
Differences in composite outcome by model were statisti- (0.4mL/s) was not (p = .3).
cally significant in contingency analysis and the results
were in two tiers, with the EMT solely in the first tier Differences in mean pump counts by model were sta-
with good outcomes (90%), whereas the second tier had tistically significant in one-way analysis: results were
good outcomes in statistically lower proportions of it- in two tiers, with the EMT and TPT3 in the low tier
erations (TPT2, 13%; TPT3, 47%; likelihood ratio, 40; with lesser means (n = 39 and 42, respectively), whereas
Table 1). In pairwise comparison, the differences in mean the high tier had TPT2 with a statistically greater mean

Evaluation of Pneumatic-Tourniquet Models 25