A systematic review of wellbeing in children: a comparison of military and civilian families

Electronic literature databases were searched in September 2016 and again in February 2018 for relevant studies, including PsycInfo, EMBASE, MEDLINE, PubMed, Google Scholar, and Web of Science. Search terms included military (military OR army OR combat OR armed forces OR soldier OR navy OR air force OR marine OR veteran OR service personnel OR sailor OR airman OR military personnel OR military deployment), child (child* OR famil* OR offspring OR adolescen*) and wellbeing (resilien* OR hardiness OR wellbeing OR mental OR well-being OR health*) key words. Reference lists of relevant articles and review papers (e.g. [7]) and issues of journals (e.g. Journal of Traumatic Stress; Journal of Adolescent Health) were also examined for eligible studies.

To be considered for inclusion, studies had to include: a sample of children with a parent or sibling in the military compared to a sample of children without a military connected parent/sibling; a measure of child mental health or wellbeing; and a sample of child participants below 19 years of age. Excluded studies included:

We use the term ‘child’ throughout to refer to both children and adolescents under the age of 19 years. A child in a military family was defined as the legal dependent of a military serviceman/woman (of any nationality) or a child with a sibling in the AF. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart (Fig. 1) describes the systematic review process [23]. Nine studies met the inclusion criteria for this review.

The following data was extracted from each study, if available: (a) study information (e.g. study design, location); (b) child demographic information (age, family status [e.g. military, non-military], ethnicity, sex); (c) the assessment time points and retention rates for longitudinal studies; (d) aspect of child wellbeing assessed; (e) how wellbeing was measured (i.e. questionnaire, interview); (f) child wellbeing informant (i.e. child, parent, teacher); (g) findings; (h) ethical issues; (i) and sources of bias. Two authors (VW and SAMS) independently extracted and assessed data for accuracy. Any discrepancies were discussed and resolved.

The following child wellbeing outcomes were explored in this review: the prevalence of child mental health disorders (PTSD, depression, suicidal ideation, and substance use), quality of life (perceived stress, positive affect, quality of life) and externalising behaviour (physical fighting, carrying a weapon). We separately examined outcomes for: (i) children in civilian families, (ii) children with a primary caregiver in the military, (iii) children with a caregiver in the military who was deployed to a combat zone, and (iv) children with a sibling in the military. If child outcomes were available for pre- and post-parental deployment or at commencement or cease of major hostilities [24], we used the rate of child mental health disorders/behaviour problems post-deployment and following the cease of major hostilities to allow for this data to be compared to studies that did not make this distinction. We also examined whether there were any differences in child outcomes pre/post major hostilities [24]. Odds ratios (OR) or adjusted odds ratios (AOR) and 95% confidence intervals (CI), were extracted from the studies. Where the OR were not available, unadjusted OR were calculated from the data. The reference category for all effect sizes was having a civilian parent. For all studies, effect sizes were regarded as statistically significant at p = 0.05 if the 95% CI did not include 1.

The methodological quality of studies was independently assessed by two authors (VW and SAMS) using a 14-item checklist [25]. Studies were scored depending on whether they met the specific criteria (‘no’ = 0, ‘yes’ = 1). Studies had to at least meet criteria for items three (“Was the participation rate of eligible persons at least 50%?”), eleven (“Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?”), and fourteen (“Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?”; see [25]) to receive a quality score of ‘good.’ A study that met criteria on at least two of three items received a quality rating score of ‘fair’, while a study that met one or none of these items received a score of ‘poor.’ There was good agreement between reviewers. Any disagreements in quality rating scores were resolved following a re-examination of the data and discussion in a consensus meeting. Study quality ratings are provided in Table 1.

The nine studies identified were published between 2007 and 2016. Study quality ratings ranged from ‘good’ to ‘fair’. All studies were conducted in the US and recruited children via schools. Children were all in 6–12th grade (11–18 years, see Table 1). All studies collected data on child wellbeing using child self-report, often using non-validated measures [15, 26‐30] or questionnaires adapted from other measures [14, 24, 31]. Six studies [14, 15, 26, 27, 29, 31] used data from large-scale public-school surveys conducted in several waves (i.e. Washington State 2008 Healthy Youth Survey (HYS), [32]; California Healthy Kids Survey (CHKS), [33]).¹ Eight studies were cross-sectional [14, 15, 26‐31] and one study was longitudinal [24]. In all but three studies, information regarding parental deployment was provided [27, 29, 30]. Only two studies reported information about sibling service in the military [15, 31].

Three studies reported externalising behaviour data regarding school-based physical fighting and carrying a weapon (Table 2). Sullivan and colleagues [27] found that significantly more children with parents in the military reported having been in physical fights (AOR 1.67; 95% CI 1.62, 1.71) and carrying a weapon (AOR, 1.90; 95% CI 1.83, 1.97) than civilian children in the past 12 months. This is consistent with Reinhart et al. [30] (AOR 1.69; 95% CI 1.27, 2.25). Differences in physical fighting and carrying a weapon were largely non-significant between younger children with a civilian parent and military-connected (both deployed and non-deployed) children in Reed et al. study [26]. The only exception to this was in 8th grade males with a deployed parent who reported significantly more physical fighting compared to children with civilian parents (AOR 1.57; 95% CI 1.00, 2.47). In older children (10th/12th grade), those with deployed and military (non-deployed) parents were significantly more likely than civilian children to engage in physical fighting (see Table 2). However, significant differences between groups in terms of weapon carrying were only observed in older males with deployed parents (AOR, 2.27; 95% CI 1.48, 3.47) and females with non-deployed military parents (AOR 2.03; 95% CI 1.15, 3.59).

Four studies reported child substance use, including tobacco, alcohol consumption, marijuana, and other drug use (Table 3).

Children with civilian parents were found to have lower rates of alcohol and drug consumption compared to military-connected youth as reported by Sullivan et al. [27] and Acion et al. [28]. While Reed et al. [26] found older children (10th/12th grade) in military connected families (both non-deployed and deployed parents) to report significantly greater drug and alcohol use than civilian children with no associated observations for younger children (8th grade), irrespective of gender. Although no significant differences in alcohol consumption between those with a military parent vs civilian parent were found, younger children (8th grade) with a deployed parent were statistically more likely to consume alcohol than civilian children (Male OR 1.87; 95% CI 1.15, 3.03; Female OR 1.93; 95% CI 1.15, 3.21). No statistically significant differences in alcohol and drug consumption between military (non-deployed) and deployed parental groups were observed, with the exception of significantly greater alcohol consumption in younger (8th grade) females with deployed parents (OR 1.98; 95% CI 1.01, 3.88 [data not shown in table]).

Gilreath et al. [15] found no significant differences in alcohol, marijuana and tobacco consumption between children with civilian and military parents. The only significant association found was in terms of illicit drug use (e.g. crack/cocaine, inhalants, methamphetamine, LSD, etc.) and children with a parent in the military were significantly more likely to report consumption than children with a civilian parent (OR 1.28; 95% CI 1.04, 1.57). Children with a sibling in the military were significantly more likely to consume alcohol than children with a parent in the military (OR 1.30; 95% CI 1.04, 1.64, [data not shown in table]), although those with a sibling in the military were not significantly more likely to consume alcohol compared to children with civilian parents (OR 1.18; 95% CI 0.98, 1.43; [15]). However, it should be noted that this effect is approaching significance.

Three studies examined child self-report of suicidal ideation over the last 12-months [14, 29, 31], with two of these studies also reporting data on child self-reported depression [14, 31]. One study examined child PTSD [24]. Three studies examined child wellbeing more generally (perceptions of stress, [14]; positive affect, [31]; poor quality of life, [14]).

A relationship was found between military family membership and higher rates of externalising behaviour compared to children from civilian families, including engagement in physical fights and carrying a weapon [26, 27, 30]. However, this effect appeared to be moderated by child demographic characteristics (i.e. older age, male gender) and parental deployment [26].

Externalising behaviours are an important component to consider in examining child psychological wellbeing given the relationship between externalising problems and issues later in child and adulthood (e.g. crime, low education attainment, etc.) [2]. It is possible that the elevated rates of violent behaviour found in some children from military families is due to these youths being disproportionately influenced by portrayals of the military and war in the media, which emphasises physical fighting and weapons [26]. Furthermore, US military connected children may have increased access to weapons as, compared to civilians, military personnel are more likely to have a firearm in their home [37]. These results tentatively suggest that additional support, such as violence prevention programs, for some children within military families could be beneficial.

Substance use in childhood represents a concern as 90% of US adults with chronic substance abuse problems report starting drinking or using drugs before the age of 18 years [38]. Three studies found higher rates of substance use in youth who had experienced the deployment of a family member [15, 26, 28] compared to those with non-deployed or civilian parents. The deployment of a family member can be a substantial stressor for some children and alcohol and drugs may be utilised as a coping strategy [39]. Lower levels of parental monitoring during adolescence are associated with higher levels of child drug and alcohol use [40, 41] and parental deployment may also reduce the availability of the non-deployed parent due to increased household responsibilities [42]. As older children in military families were more likely to consume alcohol and drugs than those in civilian families [26], this could potentially suggest that children in military families are vulnerable to substance use problems at a certain age. Children with a sibling in the military were also significantly more likely to consume alcohol than children with a parent in the military [15]. This indicates that it may be beneficial for existing services for youth from military families, such as Families Overcoming Under Stress (FOCUS) and Military OneSource [43, 44], to offer advice and support for issues, such as child substance misuse, as an adjunct to the familial resilience intervention.

There was a lack of consistent evidence for the relationship between military connectedness and poorer mental health (i.e. suicidal ideation, depression, and PTSD) and low quality of life (i.e. perceived stress, positive affect, quality of life). Contributing to this was the notable lack of research regarding the relationship between military connectedness and mental health outcomes, for example only one study examined child PTSD [14]. Nonetheless, rates of mental health problems in military and non-military connected youth were generally consistent with nationally representative studies of US adolescents [13]. However, some evidence for elevated rates of mental health difficulties in youth with deployed parents compared to civilian parents was found (e.g. [14, 31]). This effect was more pronounced for males, with male children reporting significantly more depressive symptoms, poorer quality of life, and suicidal ideation compared to those with civilian parents [14]. This is notable as previous research has found such mental health problems to be generally more common in adolescent females than males in the general population (e.g. [45, 46]). Moreover, the deleterious impact of deployment on child psychological adjustment was not restricted to parental deployment and Cederbaum et al. [31] found an increased likelihood of depressive symptoms in youth who experienced the deployment of either a parent or a sibling compared to those who had not experienced familial deployment.

While most military connected children cope well, it is possible that the deployment of a family member can contribute towards the development of mental health difficulties in some children. This could be due to the stress experienced by young people when a family member deploys, such as the disruption of family routines, fears for the potential injury of the service member and uncertainty about the service member’s return [1, 47]. Familial reintegration following deployment can also be distressing for children due to the renegotiation of familial roles and psychological and/or physical injuries of the service member [48, 49].

This review was limited by several factors. First, no included study collected retrospective data on the child’s mental health. Second, most studies were cross-sectional (n = 8), school-based and reliant on child self-report. As a result, no data was collected regarding sample socio-economic status, deployment features (e.g. length of deployment, date of deployment) or family characteristics (e.g. family history of mental health problems). Therefore, it is unclear whether child wellbeing worsened immediately following parental deployment but later stabilised, if the elevated rates of child internalising and externalising problems are due to parental deployment or a comparatively challenging environment. Moreover, the collection of data regarding parental military status and wellbeing outcomes via child self-report may be subject to bias and future studies should include parent and/or teacher report. Third, several studies used data from the same large-scale public-school surveys and must be taken into consideration when interpreting the findings. Finally, all included studies were conducted in the US. Rates of mental health difficulties for civilian and military personnel children differ between the US and other nations, and the duration of deployment of US. AF is often longer than in other countries [13, 50, 51]. Therefore, the impact of parent or sibling military service on child wellbeing may be greater compared to non-US contexts, although this relationship requires further exploration.

Nonetheless, this review has several strengths, including the use of a thorough, systematic search strategy and the direct comparison of child wellbeing in military and non-military families. This review utilised a broad definition of what constitutes a “military family,” including a parent and/or a sibling in the AF. In doing so, we found having a sibling in the military to be significantly associated with higher rates of alcohol consumption compared to children with a military parent [15] as well as an increased likelihood of depressive symptoms in youth whose parent or sibling had deployed compared to those who had not experienced familial deployment [31]. This highlights the need for the impact of military family membership on child wellbeing to be examined more broadly, including not only parents but also siblings and other close relatives.