Page 80 - Journal of Special Operations Medicine - Fall 2015

Somatic symptoms range, 0–2,230mg/day). Relative to body weight, the
To assess severity of somatic symptoms, participants mean caffeine intake was 3.05mg/kg per day (SD, 4.0;
completed the Patient Health Questionnaire-15 (PHQ- range, 0–24.58mg/kg).
15), a 15-item self-report questionnaire that assesses
9
the extent to which the respondent is bothered by vari- Participants reporting alcohol consumption were signif-
ous somatic experiences or problems. Because the pres- icantly more likely to be tobacco users than participants
ent sample was composed entirely of male participants, who denied alcohol consumption [31.9% versus 9.4%;
the PHQ-15 item assessing problems with menstruation χ (1) = 6.70; p = .010]. Participants reporting alcohol
2
was omitted. The scale has demonstrated adequate in- consumption also reported significantly greater caffeine
ternal consistency and has been validated against other intake (mean, 3.37mg/kg; standard error [SE], 0.36)
measures of somatization and physical health. than participants denying alcohol consumption [mean,
1.95mg/kg; SD, 0.52; Wald χ (1) = 5.03; p = .025]. The
2
interaction of tobacco use by alcohol consumption as a
Data Analysis
predictor of caffeine use was not statistically significant,
An initial inspection of the data revealed considerable however [Wald χ (1) = 0.23; p = .632].
2
positive skew in most continuous variables; therefore,
generalized linear modeling with robust maximum Associations of Alcohol Consumption,
likelihood estimation was used for multivariate regres- Tobacco Use, and Caffeine Intake With Somatic
sion analyses. Regression models were specified to fit a and Psychological Symptoms
negative binomial distribution, which was determined The mean somatic symptom score on the PHQ-15 was
to yield best fit based on comparison of the Akaike in- 2.47 (SD, 2.13). We next computed the proportions of
formation criterion, using a smaller-is-better approach. participants reporting each type of somatic complaint.
Post hoc group differences in mean values were con- Results are displayed in Table 1. Back pain and extrem-
ducted using t tests. For univariate comparisons of cat- ity pain were the most common somatic complaints, be-
egorical variables, χ analyses were conducted. ing reported by 56.4% and 53.8%, respectively, of the
2
sample.
Results Results of regression analyses indicated that tobacco use
was associated with less severe headaches (B = −0.72;
Alcohol Consumption, Tobacco Use, SE, 0.39; p = .063), alcohol consumption was associated
and Caffeine Intake with more severe extremity pain (B = 0.39; SE, 0.21;
In terms of alcohol use, a total of 163 (83.2%) partici- p = .068), and caffeine intake was associated with sig-
pants reported any amount of alcohol consumption and nificantly more severe back pain (B = 0.04; SE, 0.01;
137 (69.9%) reported consuming at least 12 alcoholic p = .003) and extremity pain (B = 0.03; SE, 0.01; p = .010).
beverages per year. Of those reporting alcohol consump- When adjusting for age and deployment history, alcohol
tion, 149 (91.4%) reported drinking beer, 83 (50.9%) consumption and back pain were no longer associated
reported drinking wine, and 80 (49.1%) reported drink- with extremity pain and the association of tobacco use
ing liquor or spirits. On average, participants reported with headaches was weakened. Caffeine intake continued
drinking 1.24 (SD, 1.36) days per week and consumed to be associated with severity of back pain, however.
1.44 (SD, 1.57) alcoholic beverages per instance of con-
sumption, for an overall mean consumption rate of 2.85 In light of these results, we conducted two separate
(SD, 4.92) alcoholic beverages per week (range, 0–49 follow-up exploratory analyses, one for back pain and
drinks per week). Eight (2.1%) participants met criteria another for extremity pain, to compare mean caffeine
for “heavy drinking,” which, is defined as consuming intake according to pain severity group (i.e., not both-
14 or more alcoholic beverages per week for men. 10 ered, bothered a little, bothered a lot). Results indicated
a significant between-group difference in daily caffeine
In terms of tobacco use, a total of 55 (28.2%) partici- intake according to back pain group [Wald χ (2) = 11.39;
2
pants reported current tobacco use. Of those reporting p = .003] and extremity pain group [Wald χ (2) = 11.39;
2
tobacco use, 43 (78.2%) reported chewing smokeless p = .003]. As seen in Figure 1, mean daily caffeine con-
tobacco, eight (14.5%) reported smoking cigars, seven sumption was highest among those participants report-
(12.7%) reported smoking pipes or hookahs, and four ing they were bothered a lot by back or extremity pain,
(7.3%) reported smoking cigarettes. whereas mean daily caffeine consumption was lowest
among those reporting they were not bothered by back
In terms of caffeine use, a total of 174 (88.8%) partici- or extremity pain. These differences remained statisti-
pants reported regular caffeine consumption. The mean cally significant when adjusting for age and deployment
caffeine intake among was 253.70mg/day (SD, 334.95; history.

68 Journal of Special Operations Medicine Volume 15, Edition 3/Fall 2015