Page 96 - JSOM Winter 2023

Statistical Analysis Physical Domain
Statistical significance was accepted at p≤.05. All analyses Differences in the PD measures by PD participation are shown
were run using SPSS 26.0 (Statistical Package for Social Sci- in Table 3. The PD ≥4 group showed significantly greater up-
ence, SPSS Inc. (https://www.ibm.com/products/spss-statistics) per and lower body strength in bench press, weighted pull-up,
for Windows. Thematic analysis was used to analyze the qual- squat, and deadlift. PD ≥4 also showed greater overall strength,
itative data. Descriptive statistics (means, 95% CIs, medians, as evidenced by their strength-to-weight ratio (p=.01 vs. con-
interquartile ranges, and proportions) were calculated for all trol and p=.01 vs. PD <4). Finally, PD ≥4 also had greater
variables for the total study population and by level of partic- power measured by standing broad jump (255.5-cm [250.4–
ipation in the PD. The control group was compared with the 260.6] PD ≥4 vs. 244.1-cm [238.8–249.2] control, p=.01).
PD ≥4 and PD <4 groups.
The average body mass index for all groups was about
Means were compared across levels of participation in the PD 27kg/m , with percent fat being low at 11.4% (95% CI 10.9–
2
using a one-way analysis of variance. Homogeneity of variance 11.9) overall. As expected, significantly more participants in
was assessed with the Levene test. Bonferroni post-hoc tests the control group reported never meeting with the program’s
were used when the variances were equal, and Games-Howell performance dietitian. Effective mindset scores (as measured
post-hoc tests were used when the variances were unequal. by the GSE self-efficacy scale) were high, with a median of
26.0 (IQR 7.00) on a 30-point scale, with little difference by
For non-parametric values, medians were compared across participation in the PD. We did not find significant differences
levels of participation in the PD using the independent sam- in sleep quality (PSQI) by PD participation.
ples Kruskal-Wallis test. Significance values were corrected us-
ing Bonferroni for multiple tests. Proportions were compared There were no significant differences in injury or reported pain
across levels of participation using z-tests with p-values ad- by PD participation. PD ≥4 participants were significantly
justed using the Bonferroni method. more likely to have been injured during combatives (29.2%
PD ≥4 vs. 0% control, p=.01) and at training school (e.g., high
It should be noted there was some missing data for a few altitude-low opening (HALO) parachute techniques, survival,
individuals, there were about 15 data points missing for the evasion, resistance, escape (SERE), Airborne, Ranger) (37.5%
physical measures, about 10 data points missing on the ques- PD ≥4 vs. 11.1% control, p=.04, 7.9% PD <4, p=.01). Bio-
tionnaires. There were about 70 participants that did not com- markers were also considered in the PD (Table 4). Overall,
plete any of the blood draws. mean levels were within normal ranges, and there were no sig-
nificant differences by PD participation.
Results
We expanded our question about PD participation mid-way
All units at the selected post (about 420 active Operators) through the study (subset of the last 112 participants, Appen-
were informed of the study. As shown in Table 1, a total of dix A). Over 70% used the HPTC for their workouts, while
231 Soldiers participated in this study: 63 reported they did about 50% used the strength and conditioning specialists. On
not participate in the PD (control group), and 159 indicated average, 78% utilized sports medicine, including a physical
they did (66 PD ≥4 and 93 PD <4). therapist (35%) or athletic trainer (43%).
TABLE 1 Participation in the Physical Domain (PD) Psychological, Social/Family, Spiritual,
No. (%) and Health/Lifestyle Domains
Participation in the PD? Table 5 shows the psychological domain measures by PD par-
ticipation. Significant variation by PD participation was evi-
Yes 159 (68.8) dent in positive affect, generalized anxiety, perceived stress, and
No 63 (27.3) depression. The PD ≥4 group reported significantly higher pos-
Did not answer 9 (3.9) itive affect), lower levels of PHQ-8 depression), lower GAD-7
Total 231 (100) anxiety), and less Cohen PSS stress) than the control group.
How often?
Control (did not participate) 63 (27.3) The social and family domain measures (Table 6) did not signifi-
PD <4×/week 93 (40.3) cantly vary by PD participation except for family support; the
PD ≥4×/week 66 (28.6) PD ≥4 group reported significantly better family support than
the PD <4 group. Like the social/family domain, there was little
Did not answer 9 (3.9)
variation in the spiritual domain by PD participation (Table 7).
Table 2 shows demographic information by PD participation We also asked about several health behaviors, including ciga-
level. The average age of Soldiers was about 31 years, with rette and smokeless tobacco use, caffeine consumption, seat-
more than half of the participants being married with chil- belt use, vitamin D intake, nutrition, and physical activity
dren. Years in SOF were significantly lower among PD ≥4 (2.9 (Appendix B). Few participants reported smoking (2.7%). Al-
[95% CI 2.1–3.8] y) than PD <4 (5.1 [95% CI 4.1–6.2] y). most 25% of participants reported using smokeless tobacco.
Staff Sergeants made up the largest percentage of participants, Nearly 40% of the PD ≥4 group reported taking caffeine sup-
followed by Sergeants First Class. Significantly more non- plements compared with 21% of the control group.
participants (i.e., the control group) reported being deployed
within 90 days of testing, while a greater percentage of PD Qualitative Comments
participants had not deployed for more than 365 days before Sixty participants answered the open-ended question: “Are
testing. there any comments or thoughts about the THOR3 (PD) that

94 | JSOM Volume 23, Edition 4 / Winter 2023