Abstract
Study Objectives
The present study characterized a sample of 4,667 Army soldiers based on their patterns of insomnia before, during, and after deployment, and explored pre-deployment factors predictive of these patterns.
Methods
Data were analyzed from the Army Study to Assess Risk and Resilience in Service members (STARRS)—Pre/Post Deployment Study (PPDS), using surveys that captured data approximately 1–2 months pre-deployment, and 3- and 9-month post-deployment from soldiers deployed to Afghanistan. Patterns of insomnia across time were examined. Theoretically derived variables linked to sleep disturbance were examined as predictors of the insomnia patterns.
Results
Five longitudinal patterns of insomnia characterized the majority of the sample: “No Insomnia” (no insomnia symptoms at any timepoint; 31%), “Deployment-related Insomnia” (no pre-deployment insomnia, developed insomnia symptoms during deployment and recovered; 40%), “Incident Insomnia” (development insomnia during or shortly after deployment that did not remit; 14%), “Chronic Insomnia” (insomnia both pre- and post-deployment; 11%), and “Other Insomnia” (reported insomnia at ≥1 timepoint, but no clear pattern across the deployment cycle; 4%). Several pre-deployment factors were predictive of insomnia trajectories, including lifetime major depressive episodes, traumatic brain injury history, posttraumatic stress disorder, and past year personal life stressors.
Conclusions
Distinct longitudinal patterns of insomnia were identified, with more than half of the sample reporting insomnia at some point in the deployment cycle. Identifying mental health conditions that are associated with different insomnia patterns prior to deployment can inform targeted interventions to reduce long-term sleep difficulty.
Keywords: insomnia, deployment, soldiers, Army STARRS, veterans, mental health
Statement of Significance.
Few studies on insomnia in active-duty soldiers have examined the nature of individual-level change over deployment, precluding the identification of transient versus chronic patterns of insomnia. In this sample of Army soldiers, we report distinguishable longitudinal patterns of insomnia, with the largest group reporting transient insomnia and a quarter of the sample reporting a variant of a chronic insomnia profile. Those with chronic insomnia were more likely to experience mental health conditions and life stressors, perhaps serving as ongoing insomnia triggers. Considering insomnia as an evolving process across the deployment cycle (rather than as a fixed entity) highlights the need to determine when appropriate interventions could reduce the development of symptoms or interrupt the transition from transient to chronic insomnia.
Introduction
Elevated prevalence of insomnia symptoms are frequently documented among deployed military soldiers and characterized as inevitable occupational hazards attributed to deployment-related factors (e.g. nighttime shifts, short sleep periods, stimulant consumption, combat exposure/threats) [1–3]. Unfortunately, sleep disturbances and insomnia diagnoses are not isolated to the deployment period, and have been shown to precede and persist beyond deployment [1, 4–6]. To date, most research on deployment-related insomnia is conducted cross-sectionally at various timepoints in the process (pre-, during, and post-deployment), precluding an understanding of how insomnia symptoms change within individuals across deployment.
The existing literature has demonstrated that there are several critical periods throughout the deployment cycle when insomnia symptoms may develop. First, many active duty soldiers may enter deployment with existing sleep disturbance. One study revealed that approximately 20% of soldiers entered deployment with pre-existing insomnia [5], which is a significant, independent contributor to the development of post-deployment mental health conditions, including posttraumatic stress disorder (PTSD) [4, 7], depression [4, 8], and suicidal ideation [7]. As insomnia interventions can reduce the risk of developing psychiatric symptoms [9], early identification and treatment of insomnia symptoms may improve operational readiness and post-deployment transitions.
Military deployments can shorten sleep duration and increase sleep disruption [1, 2], changes that could have a detrimental impact on mission-related performance. Thus, deployment itself represents a sensitive period for sleep disturbance, and one that could yield additive effects for those already experiencing sleep disruption prior to deployment. Moreover, for some soldiers, sleep disturbances can persist through the post-deployment reintegration process. In a retrospective review of post-deployment screenings of US service members [10], 41% of soldiers deployed to Iraq or Afghanistan reported some degree of insomnia upon deployment return, with 36% continuing to report insomnia symptoms at the 3-month post-deployment assessment. Notably, insomnia symptoms identified at 4-months post-deployment were found to be significant predictors of depression and PTSD at 12-months post-deployment [11]. Much like deployment itself, the post-deployment adjustment period can involve a number of life stressors that may precipitate mood disturbance (e.g. re-integration into work and home life, healing injuries). What is less understood, however, are the patterns of change in sleep disturbance throughout the course of deployment, and what predisposing factors may be associated with these patterns. Identifying when sleep disturbance emerges during the deployment cycle, and if or when it remits, could be crucial for understanding post-deployment functioning, and targeting interventions to improve functioning.
Finally, far less is known about the soldiers who remain relatively good sleepers despite exposure to the same environmental stressors that are frequently associated with sleep disturbance. Findings from a study using the Millennium Cohort Study participants found that a majority of the active-duty personnel (75.8%) did not self-report any insomnia symptoms at baseline and these individuals were more likely to be categorized as resilient at the follow-up (i.e. self-reporting very good or excellent health, fewer lost work days due to injury, and lower healthcare utilization). With increasing insomnia symptoms, the percentage of soldiers in the resilient categories decreased [12]. These data are a reminder that approximately 4 out of 5 soldiers report adequate sleep when entering deployment; however, conclusions are limited by assessment of sleep disturbance only at baseline and do not consider the fluctuations that can occur over time. Thus, a more nuanced characterization of “resilient” sleepers, those who maintain adequate sleep throughout the deployment cycle, could be important for identifying modifiable treatment targets.
The present study expands on existing cross-sectional data by aiming to identify patterns of insomnia including pre-deployment, during deployment, and 3- and 9-months post-deployment. Based on previous research that shows a proportion of the sample will have prior insomnia experiences [5, 7, 13] and that some soldiers will report sleep disturbances following deployment [10], three longitudinal patterns of insomnia were hypothesized to capture the majority of the sample: a “chronic insomnia group” (those who report pre-deployment insomnia symptoms and continue to report insomnia post-deployment), a “new onset insomnia group” (those who do not report pre-deployment insomnia, but report insomnia symptoms post-deployment), and a “resilient sleeper group” (those who report no insomnia prior to and following deployment). Next, we examined pre-deployment sociodemographic, health behavior, and mental health characteristics as predictors of the identified patterns of insomnia.
Methods
Overview of pre/post deployment study
The data source for the present study was the Pre/Post Deployment Study (PPDS) of the Army Study to Assess Risk and Resilience in Service members (Army STARRS). Complete descriptions of the design and methodology are described in detail elsewhere [14, 15]. In short, the PPDS is a prospective, four-wave panel survey that collected data from United States Army soldiers in three Brigade Combat Teams (BCTs), shortly before deployment to Afghanistan in 2012 (T0/baseline) and three additional times following return from deployment: within 1 month of return (T1), 3 months following return (T2), and 9 months following return (T3). This post-deployment phase ran through April 2014. At baseline, participants provided written informed consent for study participation, to link their administrative records to their survey responses, and to participate in future assessments. The PPDS recruitment, consent, and data protection procedures were approved by the Human Subjects Committees of all collaborating organizations.
For the purposes of the present study, time points T0, T2, and T3 were used. The T1 questionnaire, omitted from the present study, only briefly evaluated deployment-related experiences and did not inquire about past month insomnia. Therefore, in order to capture deployment-related sleep disturbance, the T2 survey questions specifically asking about sleep during the recent deployment were used as a proxy for that time point. The past month questions from T2 were used as intended to capture insomnia symptoms at the 3-month post-deployment period.
Participants
At T0, 8,558 soldiers completed the survey, of which 55% had complete data available for the longitudinal analyses in the present study. Therefore, the analytic sample for identifying and characterizing the longitudinal patterns of insomnia included 4,667 soldiers.
Measures
Insomnia
Past 30-day insomnia symptoms were assessed by self-report at each time point using items adapted from the Brief Insomnia Questionnaire [16]. Consistent with previous studies [7], past month insomnia was defined as (1) having problems getting to sleep, staying asleep, waking too early, or feeling tired after a full night in bed at least 3 or more nights per week, and (2) experiencing at least “some” interference of one or more of the following insomnia-related symptoms: daytime fatigue, sleepiness, or low motivation; headaches and upset stomach; moodiness; reduced school or work performance; and accident-proneness. Participants were categorized (yes/no) for probable insomnia at each time point.
Lifetime insomnia was considered present if the soldier indicated (1) ever having problems getting to sleep, staying asleep, waking too early, or feeling so tired after a full night’s sleep that it interfered with daily activities; (2) experiencing a whole month or longer when such sleep problems were present at least three nights per week; and (3) experiencing at least “some” interference of one or more symptoms described above during a typical month when the insomnia was worst.
Pre-deployment health behaviors
Soldiers were asked items adapted from the substance abuse module of the Composite International Diagnostic Interview Screening Scales (CIDI-SC [17]) to assess substance usage in the past 30 days. Caffeine (e.g. energy drinks, coffee, soda, pills, and gum) and tobacco (e.g. cigarettes, cigars, smokeless) use were each collapsed into three categories: no use, some use (reporting less than 1 day a week up to 4 days a week), and daily use. The use of prescription stimulant, sedative, and/or pain relievers in the past 30 days that occurred without a prescription or included consuming more than prescribed to get high, buzzed, or numb was categorized into never or at least once misusing prescription drugs. Lastly, past month problematic alcohol consumption (defined as drinking 5 or more drinks of alcohol on the same day) was categorized into: never, some (range of <1 to 1 to 2 days/week), or heavy use (range of 3–4 days/week to daily).
Pre-deployment lifetime mental health symptoms
At the pre-deployment survey, soldiers were asked about several domains of mental health functioning based on items adapted from the CIDI-SC and a 6-item screening version of the PTSD Checklist (PCL [18]). The CIDI-SC/PCL lifetime diagnoses included in the present study were major depressive disorder (MDD), mania/hypomania, generalized anxiety disorder (GAD), PTSD, alcohol or other substance use disorder, and adult attention deficit-hyper-activity disorder (ADHD; requiring symptoms during the preceding 6 months). In addition to lifetime diagnoses, the CIDI-SC/PCL past month diagnoses included: MDD, PTSD, and GAD.
Pre-deployment lifetime traumatic brain injury (TBI) (yes/no) was defined, following Adams and colleagues’ [19] previous definition, by the endorsement of 1 or more head, neck, or blast injuries at any point in the soldiers’ lives that led to loss of consciousness of any duration (i.e. <30 min, 30 min to 24 h, or >24 h).
The number of lifetime traumatic stressors was assessed through a list of 15 highly stressful experiences (e.g. physical or sexual assault, death of close friend). A total number of different types of trauma experienced was created (possible range = 0–15). Past year personal life stressors were assessed with 10 items that evaluate the severity of stress from financial matters, health, romantic relationships, family relationships, problems experienced by loved ones, work, and life overall. Each item was rated on a 5-point Likert scale (none to very severe), and all endorsements greater than “none” were summed for a total score (possible range = 0–10).
Other covariates
Self-reported sociodemographic variables were collected in the pre-deployment survey, which included age in years, sex, race and ethnicity, highest education level, marital status, and number of military deployments.
Data analytic plan
Identifying longitudinal patterns of insomnia over the deployment cycle
Based on participants’ insomnia (see Measures) pre-, during-, 3-, and 9-months post-deployment, participants were grouped into mutually exclusive categories to describe their longitudinal patterns of insomnia over the deployment cycle (e.g. no insomnia, chronic insomnia). With a binary yes/no categorization of insomnia at each of the 4 timepoints, there were 16 combinations yielding distinct trajectories. Trajectories were collapsed based on the most common patterns of response to the insomnia assessment over the follow-up.
Identifying predictors of longitudinal patterns of insomnia
We examined pre-deployment sociodemographic, behavioral, and mental health characteristics as predictors of longitudinal patterns of insomnia. To identify relevant predictors of these patterns, we first examined unadjusted differences in these characteristics across the insomnia patterns using chi-squared tests for categorical variables and analysis of variance for continuous variables. For mental health conditions, we examined both lifetime and past month diagnoses in the unadjusted comparisons. We chose to retain only lifetime diagnostic covariates in the multivariate models after observing more differences between groups on lifetime diagnosis and confirming overlap between lifetime and past month diagnoses. Finally, we fit a multivariate generalized logit model to compare sociodemographic, behavioral, and mental health characteristics across insomnia patterns. Factors that differed significantly across insomnia patterns, yielded effect sizes of at least medium magnitude (defined as 0.06 for eta-squared and 0.3 for Cramér’s V [20]) in unadjusted comparisons, and/or were previously identified risk factors (i.e. age, sex, prior deployment number, and relationship status) for insomnia, were entered into the adjusted model. The model was fit using the proc logistic procedure in SAS, specifying a generalized logit (glogit) link function. This model was selected to accommodate levels of the response variable (i.e. longitudinal pattern of insomnia) having no assumed ordering. The “No Insomnia” pattern was selected as the reference category. Pre-deployment characteristics were entered, employing a forward stepwise procedure, into the model as fixed effects. Due to small effect sizes in the unadjusted model, education level was not entered into the logistic model. For parsimony, problematic drinking status and past month abuse of the prescription medications were also removed as the responses to these variables were simultaneously captured by SUD status. The remaining variables were entered into the logistic model; however, during the stepwise selection procedure, marital status, number of prior deployments, GAD, and caffeine use did not show a significant effect on the group probabilities and were eliminated from the final model. Odds-ratios and confidence intervals were estimated for all comparisons on characteristics, including pairwise comparisons between each of the five insomnia patterns (i.e. changing reference category). The multivariate model was weight-adjusted using a weighting algorithm developed for the STARRS survey data [15], and has been previously applied to longitudinal analysis of insomnia and mental health [7].
Missing data
To examine potential bias in our sample attributed to attrition, we compared our analytic sample to participants who were excluded from this analysis (n = 3,891; 45%) due to missing data at any of the follow up time points on lifetime insomnia and demographic variables. Significant differences were observed on lifetime insomnia episodes, sex, and marital status between the two groups; however, the effect sizes were trivial (Cramér’s V = 0.03, 0.02, and 0.06, respectively).
All analyses were conducted in SAS, version 9.3 [21]. Resulting p-values from the final adjusted model were corrected for multiple testing using the Hochberg procedure (PAdj). p-Values less than 0.05 (two-tailed) were considered statistically significant.
Results
Participants of this sample were military soldiers between the ages of 18 and 64 years (M = 26.20, SD = 5.97). A majority of the sample were men (n = 4,398, 94%), identified as White (n = 3,408, 73%), and were married (n = 2,522, 54%). Additional sample characteristics can be found in Table 1.
Table 1.
Pre-deployment (T0) Sample Characteristics by Longitudinal Pattern of Insomnia
| Total N = 4,667 | No insomnia n = 1,437 | Deployment-related insomnia n = 1,851 | Incident insomnia n = 668 | Chronic insomnia n = 500 | Other insomnia n = 211 | F or Chi-Square | P | |
|---|---|---|---|---|---|---|---|---|
| n (%) or M (SD) | ||||||||
| Sociodemographic | ||||||||
| Age | 26.2 (6.0) | 26.5 (6.1) | 25.5 (5.6) | 27.3 (6.3) | 26.7 (6.0) | 26.3 (5.9) | 13.7 | 0.0001 |
| Sex (female) | 269 (6%) | 73 (5%) | 84 (5%) | 67 (10%) | 36 (7%) | 9 (4%) | 31.5 | 0.0001 |
| Race | 15.0 | 0.06 | ||||||
| White | 3,408 (73%) | 1050 (73%) | 1,358 (73%) | 479 (72%) | 367 (73%) | 154 (73%) | ||
| Black | 479 (10%) | 167 (12%) | 193 (10%) | 55 (8%) | 41 (8%) | 23 (11%) | ||
| Other | 780 (17%) | 220 (15%) | 300 (16%) | 134 (20%) | 92 (18%) | 34 (16%) | ||
| Ethnicity (Hispanic) | 759 (16%) | 222 (15%) | 309 (17%) | 117 (18%) | 84 (17%) | 27 (13%) | 3.7 | 0.45 |
| Education (≥ associate degree) | 982 (21%) | 370 (26%) | 350 (19%) | 162 (24%) | 74 (15%) | 26 (12%) | 49.8 | 0.0001 |
| Marital status | 51.8 | 0.0001 | ||||||
| Never married | 1,748 (37%) | 541 (38%) | 755 (41%) | 200 (30%) | 169 (34%) | 83 (39%) | ||
| Married | 2,522 (54%) | 807 (56%) | 948 (51%) | 395 (59%) | 270 (54%) | 102 (48%) | ||
| Divorced, separated, or widowed | 397 (9%) | 89 (6%) | 148 (8%) | 73 (11%) | 61 (12%) | 26 (12%) | ||
| Average number of deployments | 1.9 (2.4) | 1.8 (2.4) | 1.7 (2.3) | 2.3 (2.6) | 2.3 (2.5) | 2.2 (2.5) | 11.8 | 0.0001 |
| Pre-deployment health behaviors | ||||||||
| Tobacco use | 54.5 | 0.0001 | ||||||
| None | 1,845 (40%) | 639 (44%) | 681 (37%) | 287 (43%) | 177 (35%) | 61 (29%) | ||
| Some (1–4 days/week) | 841 (18%) | 253 (18%) | 369 (20%) | 113 (17%) | 75 (15%) | 31 (15%) | ||
| Daily | 1,981 (42%) | 545 (38%) | 801 (43%) | 268 (40%) | 248 (50%) | 119 (56%) | ||
| Problematic drinking | 83.1 | 0.0001 | ||||||
| Never | 2,059 (44%) | 706 (49%) | 798 (43%) | 319 (48%) | 166 (33%) | 70 (33%) | ||
| Low users (≤1 or 1–2 day(s)/week) | 2,368 (51%) | 681 (47%) | 969 (52%) | 313 (47%) | 281 (56%) | 124 (59%) | ||
| Heavy users (≥3 days/ week) | 240 (5%) | 50 (4%) | 84 (5%) | 36 (5%) | 53 (11%) | 17 (8%) | ||
| Caffeine | 52.6 | 0.0001 | ||||||
| None | 178 (4%) | 55 (4%) | 78 (4%) | 24 (4%) | 14 (3%) | 7 (3%) | ||
| Some (1–4 days/week) |
2,356 (51%) | 797 (55%) | 954 (52%) | 309 (46%) | 198 (40%) | 98 (46%) | ||
| Daily | 2,133 (46%) | 585 (41%) | 819 (44%) | 335 (50%) | 288 (58%) | 106 (50%) | ||
| Stimulant use (ever) | 46 (1%) | 12 (0.8%) | 8 (0.4%) | 9 (1%) | 11 (2%) | 6 (3%) | 22.1 | 0.0002 |
| Sedative use (ever) | 48 (1%) | 8 (0.6%) | 16 (1%) | 5 (0.8%) | 14 (3%) | 5 (2%) | 23.3 | 0.0001 |
| Pain reliever use (ever) | 81 (2%) | 12 (0.8%) | 30 (2%) | 15 (2%) | 16 (3%) | 8 (4%) | 19.5 | 0.0006 |
| Lifetime mental health factors | ||||||||
| Insomnia | 1,683 (36%) | 237 (17%) | 491 (27%) | 244 (37%) | 500 (100%) | 211 (100%) | 1,572.3 | 0.0001 |
| MDE |
419 (9%) | 21 (1%) | 126 (7%) | 57 (9%) | 173 (35%) | 42 (20%) | 542.7 | 0.0001 |
| Mania/hypomania | 187 (4%) | 10 (0.7%) | 65 (4%) | 22 (3%) | 63 (13%) | 27 (13%) | 181.4 | 0.0001 |
| GAD | 371 (8%) | 27 (2%) | 119 (6%) | 54 (8%) | 134 (27%) | 37 (18%) | 347.5 | 0.0001 |
| SUD | 966 (21%) | 195 (14%) | 371 (20%) | 138 (21%) | 186 (37%) | 76 (36%) | 158.1 | 0.0001 |
| TBI | 1,632 (35%) | 375 (26%) | 625 (34%) | 262 (39%) | 263 (53%) | 107 (51%) | 147.6 | 0.0001 |
| Past 6 month ADHD | 285 (6%) | 18 (1%) | 97 (5%) | 35 (5%) | 102 (20%) | 33 (16%) | 273.9 | 0.0001 |
| PTSD |
526 (11%) | 48 (3%) | 159 (9%) | 100 (15%) | 162 (32%) | 57 (27%) | 388.3 | 0.0001 |
| Average number of traumatic life events | 3.6 (3.2) | 2.8 (2.7) | 3.5 (3.1) | 4.0 (3.3) | 5.5 (3.6) | 4.5 (3.4) | 75.67 | 0.0001 |
| Average number of past year personal life stressors | 3.6 (2.9) | 2.7 (2.5) | 3.7 (2.8) | 3.9 (2.8) | 5.4 (3.0) | 4.7 (2.9) | 110.1 | 0.0001 |
This table presents unweighted ns and %s. Problematic drinking = 5+ alcoholic drinks per day. ADHD = attention deficit hyperactivity disorder, GAD = generalized anxiety disorder, MDE = major depressive episode, PTSD = posttraumatic stress disorder, SUD = any substance use disorder, TBI = traumatic brain injury.
Longitudinal patterns of insomnia over the deployment cycle
By collapsing the common patterns of insomnia over time, five longitudinal insomnia patterns characterized the 4,667 participants in our sample: (1) “No Insomnia,” n = 1,437 (31%), reporting no insomnia at any time point; (2) “Deployment-related Insomnia,” n = 1,851 (40%), reporting insomnia only during and up to 3-months after deployment (T2); (3) “Incident Insomnia,” n = 668 (14%), who developed insomnia during or up to 3-months after deployment (T2) and continued to report insomnia symptoms at 9-months (T3); (4) “Chronic Insomnia,” n = 500 (11%), had pre-deployment insomnia that remained through follow-up; (5) “Other Insomnia,” n = 211 (4%), reported insomnia at ≥1 timepoint(s), but had no identifiable pattern of insomnia across the deployment cycle.
Pre-deployment sociodemographic, behavioral and mental health predictors of longitudinal pattern of insomnia
Table 1 displays unadjusted prevalence of characteristics by insomnia group, and Table 2 displays results of the final weight-adjusted multivariate generalized logit model comparing the odds of pre-deployment characteristics between the identified longitudinal patterns of insomnia.
Table 2.
Weight-Adjusted Stepwise Multinomial Logistic Regression of Pre-deployment Factors Predicting Longitudinal Pattern of Insomnia
| Deployment-related insomnia vs. no insomnia | Incident insomnia vs. no insomnia | Chronic insomnia vs. no insomnia | Other insomnia vs. no insomnia | |||||
|---|---|---|---|---|---|---|---|---|
| Predictors | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI |
| Age | 0.97* | 0.96–0.982 | 1.01 | 1.00–1.03 | 1.00 | 0.98–1.01 | 0.99 | 0.96–1.01 |
| Sex (reference: male) | 1.08 | 0.78–1.51 | 0.47* | 0.33–0.671 | 0.63 | 0.40–1.01 | 0.99 | 0.48–2.06 |
| Tobacco use (reference: none) | ||||||||
| Some (1–4 days/week) | 1.30 | 1.07–1.59 | 1.06 | 0.81–1.39 | 1.11 | 0.79–1.57 | 1.27 | 0.79–2.04 |
| Daily | 1.26 | 1.08–1.48 | 1.07 | 0.87–1.33 | 1.40 | 1.08–1.80 | 1.93* | 1.36–2.73 |
| Lifetime SUD | 1.11 | 0.91–1.36 | 1.09 | 0.85–1.41 | 1.41 | 1.07–1.85 | 1.64 | 1.15–2.32 |
| Lifetime MDE | 2.40* | 1.47–3.92 | 2.65* | 1.54–4.57 | 7.56* | 4.53–12.631,2 | 3.57* | 1.93–6.60 |
| Lifetime Mania/hypomania | 2.52 | 1.27–5.01 | 2.02 | 0.92–4.41 | 3.06 | 1.47–6.40 | 4.77* | 2.16–10.56 |
| Lifetime TBI | 1.20 | 1.02–1.41 | 1.46* | 1.19–1.80 | 1.79* | 1.40–2.281 | 1.89* | 1.38–2.60 |
| Past 6 month ADHD | 2.18 | 1.28–3.69 | 1.89 | 1.03–3.47 | 3.63* | 2.06–6.41 | 3.42* | 1.78–6.60 |
| Lifetime PTSD | 1.58 | 1.11–2.23 | 2.52* | 1.72–3.69 | 3.03* | 2.05–4.471 | 3.59* | 2.25–5.741 |
| Average number of traumatic life events | 1.04 | 1.01–1.07 | 1.07* | 1.03–1.11 | 1.13* | 1.09–1.181 | 1.05 | 1.00–1.10 |
| Average number of past year personal life stressors | 1.11* | 1.08–1.14 | 1.11* | 1.07–1.15 | 1.19* | 1.14–1.241 | 1.15* | 1.08–1.22 |
This table presents weighted ORs. ADHD = attention deficit hyperactivity disorder, CI = confidence interval, MDE = major depressive episode, OR = odds ratio, PTSD = posttraumatic stress disorder, SUD = any substance use disorder, TBI = traumatic brain injury. *PAdj < 0.05.
1–4 indicates a pairwise difference (PAdj < 0.05) with: 1Deployment-related insomnia, 2Incident insomnia, 3Chronic insomnia, 4Other insomnia.
Deployment-related insomnia vs. no insomnia
Compared to the No Insomnia group, individuals who developed insomnia during or within 3 months following deployment that remitted by nine months post-deployment were younger (25.5 vs. 26.5 years (Table 1), Padj < 0.05 (Table 2). Individuals with deployment-related insomnia were 2.4 times more likely to have a history of major depressive episodes (95% CI: 1.4 – 3.9) and experienced more past-year personal stressors on average (3.7 vs. 2.7 events; Table 1; Padj < 0.05 (Table 2) compared to those in the No Insomnia group.
Incident insomnia vs. no insomnia
Compared to the No Insomnia group, individuals in the Incident Insomnia group, those who developed insomnia during or after deployment that did not remit by 9 months, were more likely to be women (10% vs. 5%; Table 1; Padj < 0.05, Table 2). With respect to mental health, individuals with Incident Insomnia were 2.6 times more likely to have a history of MDE and 2.5 times more likely to have a history or PTSD compared to the No Insomnia group (Padj < 0.05 for both, Table 2). History of TBI also was more likely in Incident Insomnia compared No Insomnia (OR: 1.5, 95% CI: 1.2–1.8). On average, the Incident Insomnia group experienced more lifetime traumatic events (4 vs. 2.8 events) and past-year life stressors (3.9 vs. 2.7 events) compared to the No Insomnia group (Padj < 0.05 for both).
Incident insomnia vs. deployment-related insomnia
Incident and Deployment-related Insomnia groups differed on demographic characteristics but not on pre-deployment health behaviors or lifetime mental health factors. The Incident Insomnia group was older (27.3 vs. 25.5 years; Table 1) and more likely to be women (10% vs. 5%; Table 1) than the Deployment-related Insomnia group (Padj < 0.05 for both, Table 2).
Chronic insomnia vs. no insomnia
The largest observed differences between groups on mental health factors were between the Chronic and No Insomnia groups. In the Chronic Insomnia group, prevalence of lifetime history of MDE was 35%, and was 7.6 times more likely than in the No Insomnia group (95% CI: 4.5–12.6; Table 1 and Table 2). Lifetime PTSD, past 6-month ADHD, and lifetime TBI were 3.0, 3.6, and 1.8 times more likely in the Chronic Insomnia compared to the No Insomnia groups, respectively (Padj < 0.05 for all; Table 2). On average, the Chronic Insomnia group experienced the highest number of lifetime traumatic events (M = 5.5 events, SD = 3.6) and past-year life stressors (M = 5.4 event, SD = 3.0) than any other group (Table 1), and significantly more than the No Insomnia group (Padj < 0.05, Table 2).
Chronic Insomnia vs. Incident and Deployment-related Insomnia
Differences between Chronic Insomnia and Deployment-related Insomnia followed similar patterns observed in Chronic vs. No Insomnia. Specifically, Chronic Insomnia was associated with increased likelihood of MDE, PTSD, TBI, lifetime traumatic events, and past year life stressors compared to Deployment-related Insomnia (Table 2). Chronic Insomnia also was associated with greater likelihood of lifetime MDE compared to Incident Insomnia (Padj < 0.05, Table 2). No other differences were observed between Chronic and Incident Insomnia.
Other insomnia vs. no insomnia
The Other Insomnia group followed a similar pattern to the Chronic Insomnia group. Specifically, Other Insomnia was associated with increased likelihood of lifetime MDE, PTSD, TBI, and past 6-month ADHD compared to No Insomnia (Padj < 0.05, Table 2). Additionally, Other Insomnia was associated with increased likelihood of daily tobacco use and lifetime history of mania or hypomania compared to No Insomnia (OR: 4.8, 95% CI: 2.2–10.6). Unlike Chronic Insomnia, there was no difference in number of traumatic life events in the Other Insomnia compared to the No Insomnia group.
Discussion
The results of the study extend cross-sectional research of pre- and post-deployment insomnia symptoms [1, 5] by demonstrating that patterns of insomnia symptoms across deployment can be distinguished. Five longitudinal patterns of insomnia were identified, with less than half of the sample (31%) denying insomnia symptoms across all timepoints, and the majority reporting insomnia symptoms at some point across the deployment cycle.
The largest group, defined as Deployment-related Insomnia (40%), was characterized by the presence of insomnia symptoms in close relation only to the deployment period. Given contextual factors associated with deployment (i.e. rotating shifts, reduced opportunity for sleep, combat exposure), the fact that such a large proportion of the sample experienced new-onset insomnia during deployment is not surprising. Although individuals in this trajectory experienced remission of insomnia symptoms post-deployment, the presence of insomnia is known to increase risk for work-related accidents and injuries [22], to detract from morale and intentions to remain in Army careers [13], and to confer additional risk for the development of psychiatric disorders, such as major depressive disorder and PTSD [11]. Additionally, for some of these individuals, the insomnia remained present up to 3-months post-deployment suggesting that the sleep disturbance acquired during deployment may take time to improve. Thus, while sleep disturbance is isolated to the deployment and early transition period, it can promote negative effects that outlast the individual’s military service, irrespective of continued insomnia symptoms.
Nearly a quarter of the sample had pre-existing insomnia or developed insomnia symptoms that remained throughout the post deployment transition and up to 9 months later. Similar to individuals with Deployment-related Insomnia, those whose insomnia persisted past the deployment period are at increased risk of developing or maintaining mental health disorders. These individuals also may experience increased difficulty with post-deployment reintegration, with negative impacts on their social lives [23] and ability to succeed in their post-deployment careers [24], relative to individuals whose insomnia symptoms remit upon returning home.
In secondary analyses identifying pre-deployment predictors of the longitudinal patterns of insomnia, there was shared vulnerability among the symptomatic trajectories compared to the No Insomnia group. As expected, the No Insomnia group was less likely to experience nearly all mental health problems, lifetime trauma, or past year stressors relative to all of the symptomatic trajectories. Within the symptomatic trajectories, we observed a semblance of a hierarchical structure, providing support for an increased vulnerability for negative outcomes in relation to increased chronicity of longitudinal insomnia symptoms. Fewer differences in correlates of group assignment were observed between the No Insomnia and Deployment-related Insomnia trajectories, suggesting this group is the least severe of the symptomatic patterns. Both Incident Insomnia and Chronic Insomnia patterns were associated with greater likelihood of lifetime MDE, TBI, and PTSD than the No Insomnia group. However, individuals in the Chronic Insomnia trajectory were more likely to experience a lifetime MDE compared to either the Incident or Deployment-related Insomnia trajectories, and were more likely to experience TBI, PTSD, and trauma-related stressors compared to the Deployment-related Insomnia group. This result suggests greater mental health burden associated with Chronic Insomnia, not only relative to the No Insomnia group, but relative to both the Deployment-related and Incident Insomnia trajectories. The Other Insomnia group, representing only 4% of the sample, was more difficult to characterize, as few differences were observed with this group compared to the other insomnia patterns. The lack of differences is potentially due to the small cell size limiting power to detect significant differences. Conversely, this outcome also could suggest that the unpredictable fluctuations in sleep disturbance may relate to the greater likelihood of lifetime history of mania or hypomania, or are idiopathic in nature.
Several of the lifetime mental health conditions (i.e. MDE, TBI history, and PTSD) and life stressors were significantly related to all or most of the symptomatic insomnia patterns. These results are consistent with factors associated with the stress-diathesis model for insomnia development, which posits that underlying individual vulnerability to insomnia (e.g. genetic predisposition, health conditions, early childhood experiences) interacts with stressful life events to promote the disorder state [25]. While the development of insomnia is a dynamic process that involves additional factors not assessed in this study, soldiers experiencing fewer of these predisposing vulnerabilities may possess a higher tolerance for the stressors of deployment (e.g. new sleeping conditions, chronic exposure to threat, and rotating shifts).
Our findings are particularly important with regard to women’s healthcare both during and post-deployment. Female soldiers are potentially more vulnerable for greater chronicity of insomnia developing during or shortly after deployment. In our sample, such chronicity is connected to the greatest overall physical and mental health burden. Women comprise approximately 17% of all military officers, cadets, and enlisted soldiers [26], a proportion that will likely increase over time. Although women are the minority in the military, they are significantly more likely to develop insomnia than men [27]. Women have also been shown to report greater frequency of military sexual trauma (14% to 1%) and greater rates of depression relative to men (48–39%) [28], factors which could potentially influence the severity and chronicity of insomnia in female soldiers. Indeed, results from the present study are consistent with insomnia rates in Veterans. Nearly half of women Veterans screen positive for insomnia, with insomnia symptoms associated with increased alcohol use, pain severity, mental health conditions, and outpatient mental health care utilization compared to those without insomnia symptoms [29]. Future work must explore the characteristics and life experiences of women that could interrupt this trajectory and optimize the health of women soldiers and Veterans most vulnerable to the development and effects of chronic insomnia.
The identification of distinct insomnia trajectories and correlates of group assignment throughout the deployment cycle underscores the importance of early assessment and intervention. Screening for insomnia prior to deployment and utilizing evidence-based treatments for improving sleep disturbance could potentially reduce rates of chronic insomnia trajectories in military personnel and possibly influence positive outcomes with regard to social and occupational functioning post-deployment. Pre-deployment psychoeducation about insomnia and behavioral strategies to ameliorate symptoms could buffer against the development of both Deployment-related and Incident Insomnia trajectories. Additionally, re-assessment and treatment of mental health conditions following deployment or prior to re-deployment may be an important factor in preventing Deployment-related insomnia from transitioning into a persistent pattern. Persistent precipitating factors more commonly experienced by military personnel, such as the negative effects of trauma or injury, could make a soldier more vulnerable to continued sleep disturbance without additional and timely intervention.
Limitations
While this study is strengthened by its prospective design that included assessment before and at multiple points across a combat deployment, several study limitations need to be considered. First, while the PPDS data come from a large, representative sample of active-duty Army soldiers, the present results are exclusive to the sample of soldiers who completed each of the surveys. Similarly, the predictors identified in this study are conditional on this sample and the variables selected into the model; therefore, these results may not generalize to other samples. The reliance of self-administered questionnaires of insomnia and other mental health symptoms may be less accurate than those collected by trained clinicians. In addition, the categorization of insomnia symptom status was created by questions that formed a proxy for insomnia diagnostic criteria; therefore, a full determination of diagnostic criteria for insomnia disorder could not be made. Given the use of data about deployment from the T2 survey, we cannot exclude the possibility that responses were influenced by recall bias. This survey format also did not provide an assessment of other undiagnosed or untreated clinically important sleep disturbances (e.g. obstructive sleep apnea, nightmares). Finally, it will be important for future research to consider how factors that occur during deployment (i.e. combat exposure, environmental conditions, team morale, changes to relationship or family dynamics) may influence shifts in soldiers’ insomnia trajectories.
Conclusion
While 31% of this sample of soldiers did not report insomnia symptoms across the deployment cycle, there remains a considerable number of soldiers who developed insomnia during deployment and groups of soldiers whose sleep disturbance persisted following the reintegration process. The largest group was those who developed insomnia during their deployment. Therefore, education on evidence-based insomnia prevention strategies and opportunities to practice healthy sleep patterns are needed among Army personnel to ensure safety and mental and physical health. Interventions focused on improving mental health symptoms simultaneously with healthy sleep behaviors may reduce the long-term development or exacerbation of insomnia.
Acknowledgments
The research also was supported with resources and use of facilities at the Cpl. Michael J. Crescenz VA Medical Center VISN 4 MIRECC and the VA Capitol Health Care Network (VISN 5) MIRECC. Dr. Miller’s, Boland’s, and Klingaman’s time was supported by the U.S. Department of Veterans Affairs, Veterans Health Administration (Clinical Science Research and Development Service—IK2 CX001874-PI:Katherine E. Miller, IK2CX001501-PI:Elaine M. Boland, and Rehabilitation Research and Development Service—1IK2RX001836—PI: Elizabeth A. Klingaman, VA Capitol Health Care Network), respectively. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the views of the Department of Health and Human Services, NIMH, the Department of the Army, the Department of Defense, or the Department of Veterans Affairs.
Institution where analyses were performed: Department of Psychiatry, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA
Funding
Data underlying this article were collected as part of the Army STARRS, sponsored by the Department of the Army and funded under cooperative agreement U01MH087981 with the U.S. Department of Health and Human Services, National Institutes of Health, and National Institute of Mental Health. Subsequently, STARRS- Longitudinal Study was sponsored and funded by the U.S. Department of Defense (DoD) (USUHS grant number HU0001-15-2-0004).
Disclosure Statement
Financial disclosure: none.
Non-financial disclosure: none.
Data Availability
The data are available through a restricted access web portal (https://www.icpsr.umich.edu/web/ICPSR/studies/35197).
References
- 1.Seelig AD, et al. ; Millennium Cohort Study Team. Sleep patterns before, during, and after deployment to Iraq and Afghanistan. Sleep. 2010;33(12):1615–1622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Troxel WM, et al. Sleep in the military: promoting healthy sleep among U.S. servicemembers. Rand Health Q. 2015;5(2):19. [PMC free article] [PubMed] [Google Scholar]
- 3.Lincoln ML, et al. Sleep disturbances after deployment: National Guard soldiers’ experiences and strategies. Sleep Health. 2018;4(4):377–383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gehrman P, et al. Predeployment sleep duration and insomnia symptoms as risk factors for new-onset mental health disorders following military deployment. Sleep. 2013;36(7):1009–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Taylor DJ, et al. ; STRONG STAR Consortium. Prevalence, correlates, and predictors of insomnia in the US army prior to deployment. Sleep. 2016;39(10):1795–1806. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bramoweth AD, et al. Deployment-related insomnia in military personnel and veterans. Curr Psychiatry Rep. 2013;15(10):401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wang HE, et al. Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers. Sleep. 2018;42(2). doi: 10.1093/sleep/zsy229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Koffel E, et al. Pre-deployment daytime and nighttime sleep complaints as predictors of post-deployment PTSD and depression in National Guard troops. J Anxiety Disord. 2013;27(5):512–519. [DOI] [PubMed] [Google Scholar]
- 9.Bei B, et al. Treating insomnia in depression: insomnia related factors predict long-term depression trajectories. J Consult Clin Psychol. 2018;86(3):282–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.McLay RN, et al. Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in U.S. service members returning from military deployments. Mil Med. 2010;175(10):759–762. [DOI] [PubMed] [Google Scholar]
- 11.Wright KM, et al. Insomnia as predictor versus outcome of PTSD and depression among Iraq combat veterans. J Clin Psychol. 2011;67(12):1240–1258. [DOI] [PubMed] [Google Scholar]
- 12.Seelig AD, et al. Sleep and health resilience metrics in a large military cohort. Sleep. 2016;39(5):1111–1120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Klingaman EA, et al. Prevalence, predictors and correlates of insomnia in US army soldiers. J Sleep Res. 2018;27(3):e12612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kessler RC, et al. Design of the Army Study to Assess Risk and Resilience in Service members (Army STARRS). Int J Methods Psychiatr Res. 2013;22(4):267–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ursano RJ, et al. ; Army STARRS collaborators. The Army study to assess risk and resilience in servicemembers (Army STARRS). Psychiatry. 2014;77(2):107–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kessler RC, et al. Reliability and validity of the brief insomnia questionnaire in the America insomnia survey. Sleep. 2010;33(11):1539–1549. [PMC free article] [PubMed] [Google Scholar]
- 17.Kessler RC, et al. The World Mental Health (WMH) survey initiative version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int J Methods Psychiatr Res. 2004;13(2):93–121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Weathers FW, et al. ed. The PTSD Checklist (PCL): reliability, validity, and diagnostic utility. Annual Convention of the International Society for Traumatic Stress Studies. 1993; San Antonio, TX. [Google Scholar]
- 19.Adams RS, et al. The association of lifetime and deployment-acquired traumatic brain injury with postdeployment binge and heavy drinking. J Head Trauma Rehabil. 2020;35(1):27–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Earlbaum Associates; 1988. [Google Scholar]
- 21.SAS Institute Inc. SAS/IML® 9.3 User’s Guide. Cary, NC: SAS Institute Inc; 2011. [Google Scholar]
- 22.Kessler RC, et al. Insomnia, comorbidity, and risk of injury among insured Americans: results from the America Insomnia Survey. Sleep. 2012;35(6):825–834. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kyle SD, et al. “.Not just a minor thing, it is something major, which stops you from functioning daily”: quality of life and daytime functioning in insomnia. Behav Sleep Med. 2010;8(3):123–140. [DOI] [PubMed] [Google Scholar]
- 24.Kucharczyk ER, et al. The occupational impact of sleep quality and insomnia symptoms. Sleep Med Rev. 2012;16(6):547–559. [DOI] [PubMed] [Google Scholar]
- 25.Spielman AJ, et al. A behavioral perspective on insomnia treatment. Psychiatr Clin North Am. 1987;10(4):541–553. [PubMed] [Google Scholar]
- 26.DMDC. Table of Active Duty Females by Rank/Grade and Service: May 2020. 2020. https://www.dmdc.osd.mil/appj/dwp/dwp_reports.jsp. Accessed July 27, 2020.
- 27.Suh S, et al. Sex differences in insomnia: from epidemiology and etiology to intervention. Curr Psychiatry Rep. 2018;20(9):69. [DOI] [PubMed] [Google Scholar]
- 28.Haskell SG, et al. Gender differences in rates of depression, PTSD, pain, obesity, and military sexual trauma among Connecticut War Veterans of Iraq and Afghanistan. J Womens Health. 2010;19(2):267–271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Babson KA, et al. Insomnia symptoms among female veterans: prevalence, risk factors, and the impact on psychosocial functioning and health care utilization. J Clin Sleep Med. 2018;14(6):931–939. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data are available through a restricted access web portal (https://www.icpsr.umich.edu/web/ICPSR/studies/35197).