Study Design
As part of a study to determine key psychological stressors associated with deployment in support of a medical mission, soldiers completed anonymous and cross-sectional paper surveys(16). Study participants were all active duty soldiers supporting the U.S. response to the Ebola crisis in mainly non-medical roles. Surveys were administered in Liberia in February 2015. Participants were briefed in groups, provided an information sheet, and 84.8% (n = 173) provided informed consent and were included in the analytic sample. Study activities were reviewed and approved by the Walter Reed Army Institute of Research Institutional Review Board.
Measures
Demographics and Deployment Background
Demographic measures included age (18-24, 25-29, 30-39, 40 and older), rank (junior enlisted [rank E1-E4], non-commissioned officers [NCOs; rank E5-E9], and officer/warrant officer [rank O1 – O6/WO1 – CW5]) and gender (male, female). Background information about deployment included months on current deployment (ranged from less than one month to five months or more) and job category (medical, logistics, aviation, command and control, support, security/force protection, and other).
Health-Promoting Leadership
Health-promoting leadership measured soldiers’ perceptions of leader behaviors focused on the preventive medicine goals of the mission as well as overall psychological health. This scale has been used in previous research with military units(44). Individual items are described in Table 2. Study participants were asked to rate their current team/unit leadership on 13 items in terms of how frequently they occurred (1 = never to 5 = always). Mean response scores for each of the 13 items were calculated. Cronbach’s alpha for this scale was 0.96.
General Leadership
Soldiers’ perceptions of general leadership behaviors were assessed using the four-item Walter Reed Army Institute of Research Leadership Scale (WRAIR-LS)-Short Form, a scale frequently used in research with military units(35, 41, 46). A sample item includes, “leader tells soldiers they have done a good job,” and response options ranging from 1 = never to 5 = always. Negative items were reverse scored. Response scores were averaged. Internal consistency was 0.82.
Behavioral Health and Wellbeing
An adapted form of the 17-item posttraumatic stress disorder (PTSD) Checklist-Specific assessed PTSD symptoms(47, 48). The scale was dichotomized using a diagnostic algorithm for each cluster and overall score of 50 as the cutoff, which is consistent with other studies conducted with military populations(49). Internal consistency was 0.94.
The eight-item version of the Patient Health Questionnaire for Depression (PHQ-8) assessed depression symptoms of study participants(50). Items in the PHQ-8 are scored with four response options (0 = not at all to 3 = nearly every day). Items were summed, which resulted in a composite score ranging from 0 to 24. Respondents were indicated as a positive screen for depression if a summary score of 10 or more was met(51). Internal consistency was 0.90.
The seven-item Generalized Anxiety Disorder scale (GAD-7) assessed anxiety symptoms(52). Items were scored with four response options (0 = not at all to 3 = nearly every day). The seven items were summed, resulting in a composite score ranging from 0 to 21. Respondents were indicated as a positive score for anxiety if a summary score of 10 or more was met(53). Internal consistency was 0.92.
Sleep problems were measured using the seven-item Insomnia Severity Index(54). Items measured participants’ difficulty falling asleep, staying asleep, and problems waking up too early, in addition to their satisfaction and distress regarding sleep problems. Each item was scored on a scale ranging from 0 to 4. The seven items were summed, resulting in a composite score ranging from 0 to 28. The summed scores were categorized to include no sleep problems (0-7), subthreshold insomnia (8-14), moderate insomnia (15-21), and severe insomnia (22-28) based on validated literature guidelines(54). The scale was further dichotomized to no sleep problems (0-14) and sleep problems (15-28). Internal consistency was 0.91.
Burnout was measured using a single item which asked soldiers to rate their level of burnout on a scale from 1 = very low to 5 = very high. The item was further dichotomized so that a response of three or below was considered low burnout and a score of four or higher was considered high burnout.
A single item measured morale. The item asked participants to rate their personal morale from on a scale from 1 = very low to 5 = very high. Morale was dichotomized so that a response of three or below was considered low morale and a score of four or above was considered high morale.
Attitudes toward Disease Risk and Preventive Medicine Practices
Three items (“I understand the level of risk from disease”; “I know what to do to protect myself from disease”; “this deployment will make a meaningful difference in fighting the Ebola epidemic”) adapted from Castro(55) and Sipos(56) were selected to assess attitudes towards the deployment. Items were rated on a five-point scale from 1 = strongly disagree to 5 = strongly agree. The items were further dichotomized so that a score of four or higher was indicated as agreement.
A three-item scale developed for this study examined attitudes towards preventive medicine measures during deployment. Response questions included both positive (“taking our temperature twice a day makes sense to me”) and negative (“taking our temperature twice a day is a waste of time”; “preventive medicine measures recommended for this deployment are not practical”) items with response options ranging from 1 = strongly disagree to 5 = strongly agree. A positive score was indicated by a response of four or higher for the first item, and a score of two or lower for the two negative items. Internal consistency was 0.73.
A single item developed for this study (“I avoided unnecessary risks on this mission”) assessed preventive behaviors during deployment. Response options ranged from 1 = strongly disagree to 5 = strongly agree, and were dichotomized so that a score of four or higher was indicated as agreement.
Analysis
An exploratory factor analysis (extraction method: principal factor; rotation method: oblique promax with Kaiser normalization) of the health-promoting leadership scale was conducted. Two subscales were discovered: psychological health-promoting leadership and preventive medicine health-promoting leadership. Rotated factor loadings for each scale item are reported in Table 2. Internal consistency was 0.96 and 0.92 for the two factors, respectively.
The study aimed to examine the association between health-promoting leadership and behavioral health, attitudes, and behaviors of soldiers. The sample population is inherently hierarchical due to clustering effects by organizational unit, and therefore, assumptions of independence could not be met(57). Indeed, the main exposure of interest, health-promoting leadership, introduces non-independence, as clusters of participants in our sample had the same leaders. In order to account for non-independence, mixed effects logistic regression models were used to indicate associations between health-promoting leadership behaviors and outcomes of interest. Military units (companies nested within battalions) were modeled as the random intercepts. All models were adjusted for rank (junior enlisted, NCOs, and Officer/Warrant Officer) and general leadership (continuous). Other demographic and background covariates were tested for inclusion in the model but did not meet the criteria for confounding (i.e., associated with the exposure and outcome, change in odds ratios by > 10%). Listwise deletion was used to handle any missing data. All data analyses were conducted using STATA 14.