The Walter Reed Army Institute of Research (WRAIR) and the Army Public Health Center (APHC) Behavioral Health Advisory Team (BHAT) invited all US Army soldiers across three major commands from December 9th, 2020 to January 19th, 2021 to participate in the survey . This data collection occurred prior to the US Food and Drug Administration’s emergency use authorization of COVID-19 vaccines. Soldiers were invited to participate via a link sent through military communication channels as part of routine operations. The survey included an information page about participation and a screener to determine eligibility. Respondents were allowed to proceed to the full survey if they agreed to participate and were then asked if they consented to let their data be used for research purposes. Survey participation was voluntary, and soldiers were not compensated for participation. This survey was approved by the WRAIR human research protection branch and the APHC office of human protections.
Soldiers assigned to the three Army commands participating in the survey were eligible for inclusion. Respondents were active-duty military or activated reservists. Civilians and contractors were not eligible to complete the survey.
Mental health was measured by screening for depression and anxiety. Depression symptoms were measured with the two-item Patient Health Questionnaire (PHQ-2) . Anxiety symptoms were measured with the two-item Generalized Anxiety Disorder (GAD) scale . For both measures, soldiers rated each item on a 4-point scale ranging from 0 (not at all) to 3 (Nearly every day). These items were followed by a question on functional impairment (“How difficult have these problems made it for you to do your work, take care of things at home, or get along with other people?”) ; this item was rated on a 4-point scale ranging from 0 (Not difficult at all) to 3 (Extremely difficult). For both depression and anxiety, scores of 3 or higher accompanied by reports of any functional impairment (scores of 1 or more) were regarded as a positive screen.
Adherence to COVID-19 public health guidelines
Adherence to COVID-19 public health guidelines was assessed with eight items (eg, wearing a mask or face covering; coughing or sneezing into your elbow or using a tissue) developed for this survey. Soldiers reported their frequency of engaging in each behavior using a 5-point scale ranging from 1 (Never) to 5 (Always). Reliability was not calculated because these items were treated as separate outcomes rather than as a scale. High adherence to each public health guideline was defined as a score of 4 or 5 (frequently or always) and low adherence to each public health guideline was defined as a score of 3 or lower.
COVID-19 leadership behaviors
COVID-19 specific leadership behaviors were assessed with 14 items from the health-promoting leadership scale[18, 19] adapted based on an intervention to promote resilience during facility-based quarantine after exposure to COVID-19 . Soldiers rated their immediate supervisor on each item using a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). High scores on the COVID-19 leadership behaviors scale were defined as a mean of 3.5 or higher (ie, a mean score rounding toward “agree” or higher), and low scores were defined as lower than 3.5. This cutoff reflects a conceptually meaningful difference between low and high categories. Internal consistency was high (Cronbach’s alpha = 0.97).
Given the established relationship between rank and mental health and health-related behaviors [25, 26], we accounted for rank in our analyses. Military rank was classified into three groups: (1) junior enlisted soldiers (E1-E4), (2) non-commissioned officers (E5-E9), and (3) officers (O1-O9) and warrant officers (WO1-WO5). ).
Given the potential effect of gender on mental health and health-related behaviors [25, 27], we accounted for gender in our analyses. Gender was classified into three groups: male, female and prefer not to respond.
General leadership was included as a covariate to ensure that the relationships between COVID-19 leadership behaviors and study outcomes were not better explained by the quality of leadership in general. General leadership was measured with the five-item Perceived Leader Effectiveness scale . Soldiers rated each item (eg, “My immediate supervisor is an effective leader” and “My immediate supervisor displays strong leadership abilities”) on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). High scores on the general leadership scale were defined as a mean of 3.5 or higher (ie, a mean score rounding toward “agree” or higher), and low scores were defined as a mean score of lower than 3.5. This cutoff reflects a conceptually meaningful difference between low and high categories. Internal consistency was high (Cronbach’s alpha = 0.96).
COVID-19 concerns were included as a covariate to ensure that COVID-19 leadership behaviors explained outcomes over and above concerns related to the pandemic. COVID-19 concerns were measured with 20 items developed for this survey. For each item, participants rated the extent to which they were worried or concerned about a range of factors related to COVID-19 such as accessing medical care, engaging in social activities, and the changing rules, regulations and guidance related to COVID-19. Soldiers rated each item on a 5-point scale ranging from 0 (not at all) to 4 (extremely). High levels of COVID-19 concerns were defined as a mean score of 1.5 or higher (ie, more than “slightly” concerned on average), and low levels of COVID-19 concern were defined as a mean score of lower than 1.5. This cutoff was chosen to reflect meaningful differences in average level of concern related to COVID-19. Internal consistency was high (Cronbach’s alpha = 0.94).
COVID-19 status was included as a covariate to ensure that COVID-19 leadership behaviors explained outcomes over and above COVID-19 infection and/or illness given the potential link between COVID-19 status and mental health symptoms [29, 30]. COVID-19 status was measured with five items developed for this survey. Soldiers were asked if they had tested positive for the virus, been diagnosed with COVID-19 by a medical professional, become seriously ill with COVID-19, been hospitalized with COVID-19, or recovered from COVID-19. Soldiers who responded “no” to all items were deemed COVID-19 negative, while those who responded “yes” to any one item were deemed COVID-19 positive.
Among survey participants (n= 7,829), each mental health item had no more than 11.7% missing, and each item assessing adherence to COVID-19 public health guidelines had no more than 7.0% missing. Each item on the COVID-19 leadership behaviors scale had no more than 11.4% missing, and all other model predictors had no more than 11.1% missing. Multivariable logistic regression models removed missing data using listwise deletion. Frequencies were calculated for the outcomes of interest (positive mental health screens and adherence to COVID-19 public health guidelines) as well as COVID-19 leadership behaviors. Unadjusted and adjusted multivariable logistic regression models were calculated. All analyzes were conducted in R v.4.1.0 .