Page 32 - Journal of Special Operations Medicine - Spring 2016

Results for all). For time to determination of bleeding control,
total time, pressure, and blood loss, the mixed statistical
Comparison of Results by Tourniquet Design model determined that 42%, 45%, 5%, and 60%, respec-
The difference (C-A-T 7 minus C-A-T 6) in percentage of tively, of the variability of results was attributable to user
effectiveness in bleeding control by C-A-T design was not effects showing the validity of the model with user effect.
statistically significant (96% for C-A-T 6 versus 97% for
C-A-T 7; p = .7). Similarly, the mean differences were not The mean difference in blood loss was statistically sig-
statistically significant in time to bleeding control deter- nificant (141mL versus 129mL for the C-A-T 6 and
mination (23 versus 22 seconds for C-A-T 6 and C-A-T C-A-T 7, respectively; p = .03) in that there was less
7, respectively; p = .1), in total time (32 versus 31 sec- blood loss with the C-A-T 7 than the C-A-T 6, but
onds for C-A-T 6 versus C-A-T 7, respectively; p = .2), only in the mixed statistical model ANOVA, which ac-
and in pressure (205mmHg versus 205mmHg for C-A-T counted for user effects.
6 versus C-A-T 7, respectively; p = .9). Only in the un-
mixed statistical model was the difference in mean blood Discussion
loss not statistically significant (141mL versus 129mL for
C-A-T 6 versus C-A-T 7, respectively; p = .05). However, In the present study using a manikin, the two C-A-T
ease of use results differed. The difference in ease of use designs showed differential performance favoring the
was statistically significant, favoring the C-A-T 7 (p = C-A-T 7 over the C-A-T 6. The difference is accentuated
.002) (Table 2); most of the results (82%) for the C-A-T 6 by an unpublished analysis of two previous C-A-T 7 pro-
were either neutral or easy (29% and 53%, respectively), totypes analyzed by the present investigators in December
whereas most of the results (74%) for the C-A-T 7 were 2014, in which the two previous prototype designs did not
either easy or very easy (41% and 33%, respectively). result in superior performance over the C-A-T 6. The feed-
back from the 2014 assessment led to the rejection of the
Comparison of Results by User of Tourniquets previous designs and to further spiral development of the
Results by user differed significantly (p = .04). When C-A-T design to what was assessed in the present report.
analyzing all users individually for mean differences in
blood loss by C-A-T design, only one user had a statis- Some differences in performance between the C-A-T 7
tically significant difference and this result favored the and C-A-T 6 were statistically significant. These were
C-A-T 7 (p = .003; eight others, p > .058); nine differ- few, however, and the sizes were small; such results are
ences were negative, and one was positive. When all common in design refinements of satisfactory but imper-
data of the 10 users were pooled, the mean difference fect medical devices. There was no worse performance
in blood loss by user remained statistically significant for C-A-T 7 by any parameter.
(p = .02), and this result favored the C-A-T 7. All users
preferred the C-A-T 7 (10 of 10; 100%). On the other hand, interuser differences were often
large and affected outcomes; 42% of the variability of
Comparison of Results in a Mixed Statistical Model overall results was attributable to user effects. In specific
Due to the presence of significant user effects, a mixed examples, 42%, 45%, 5%, and 60% of the variability
statistical model also was used to analyze the data. This of performance results were attributable to user effects
model determined that 42% of the variability of all re- for time to determination of bleeding control, total time,
sults was attributable to user effects. Again, the differ- pressure, and blood loss, respectively. Although the dif-
ential performance by C-A-T design was not statistically ferences in means, such as for blood loss, were small for
significant (p = .3). individual subjects researched, when applied to millions
of Americans, such differences would become more im-
Similarly, time to bleeding control determination, total portant. Such is the distinction between care of an in-
time, and pressure were not statistically significant (p > .3 dividual casualty and public health policy of a nation.

Table 2 Results of Ease of Use by Combat Application Tourniquet Design
Ease-of-Use Score,* No. (%)
1 2 3 4 5
Design (very difficult) (difficult) (neutral) (easy) (very easy) Total
C-A-T 6 1 (1) 5 (5) 29 (29) 53 (53) 12 (12) 100
C-A-T 7 0 (0) 9 (9) 17 (17) 41 (41) 33 (33) 100
Total 1 14 46 94 45 200
*Data represent 10 uses of each tourniquet by 10 users. C-A-T 6, Combat Application Tourniquet Generation 6; C-A-T 7, Combat Application
Tourniquet Generation 7.

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