bs_bs_banner APPLIED PSYCHOLOGY: HEALTH AND WELL-BEING, 2019 doi:10.1111/aphw.12168 InDependent but not Alone: A Web-Based Intervention to Promote Physical and Mental Health among Military Spouses Emily L. Mailey* , Brandon C. Irwin and Wei-Wen Hsu Kansas State University, Manhattan, KS, USA Jillian M. Joyce Oklahoma State University, Stillwater, OK, USA Background: Military spouses must cope with multiple threats to their physical and mental health, yet few interventions have been developed to promote health in this population. Methods: For this quasi-experimental study, military spouses (N = 231) received a standard educational intervention or an interactive, theorybased intervention; both were delivered online and lasted 10 weeks. The educational intervention directed participants to content on the existing website, Operation Live Well. The interactive intervention was based on Self-Determination Theory, delivered weekly content via podcasts, and encouraged participants to complete weekly challenges to improve physical activity, diet, and mental health. Linear mixed effects models were used to examine self-reported changes in stress, anxiety, depression, loneliness, self-esteem, physical activity, and diet from pre- to post-intervention. Results: Significant improvements were observed for all mental health outcomes, total physical activity, and sugar consumption. However, there were no significant group by time interaction effects. Conclusions: Web-based interventions may promote positive changes in mental health and health behaviours among military spouses. In this study, an interactive theory-based intervention was no more effective than an information-based intervention. Future studies should aim to determine the minimum “dose” needed to elicit meaningful changes in this population. Keywords: health promotion, military spouses, online, stress INTRODUCTION There are currently over 700,000 military spouses in the United States (Department of Defense, 2015). Compared to civilian populations, military spouses face *Address for correspondence: Emily L. Mailey, 8 Natatorium, Manhattan, KS 66506, USA. Email: emailey@ksu.edu © 2019 The International Association of Applied Psychology 2 MAILEY ET AL. unique stressors, including frequent moves, isolation from family and friends, and bearing sole responsibility for household and childcare responsibilities while their active duty spouse is deployed or on an extended separation (Blakely, Hennessy, Chung, & Skirton, 2014). These factors likely contribute to elevated levels of stress, anxiety, and depression frequently reported among military spouses (Dimiceli, Steinhardt, & Smith, 2010; Verdeli et al., 2011). Furthermore, the military lifestyle may present numerous barriers to engaging in health behaviours, as several studies have demonstrated that military spouses report low participation in health behaviours such as engaging in physical activity and maintaining a healthy diet (Suarez, 2011), and high levels of obesity (Fish, 2013; Tenconi, 2011). Thus, military spouses must cope with multiple ongoing threats to their physical and mental health. In spite of these concerns, only a handful of health promotion interventions have been developed specifically for military spouses. Kees and Rosenblum (2015) developed the HomeFront Strong intervention to promote psychological health and resilience among military spouses. The intervention was grounded in McCubbin and McCubbin’s Resiliency Theory, and employed evidence-based strategies targeting constructs in the theory (i.e. Grounding, Build Community, Manage Stress, Allow Emotion, Rethink Thinking, Cultivate Optimism). The content was delivered in eight weekly face-to-face group sessions, and a pilot test among ten participants aged 22–50 revealed significant reductions in stress (Hedges’ g = 1.50) and anxiety (Hedges’ g = 0.91) following the 8-week intervention. Cole and Horacek (2010) developed the “My Body Knows When” programme to promote healthy eating patterns among military spouses. The intervention was developed systematically using the Precede-Proceed Model and taught intuitive eating principles in ten weekly group sessions. Compared to control participants (n = 14, Mage = 37.0), intervention participants (n = 18, Mage = 37.5) demonstrated significant improvements in diet mentality (d = 1.02), but no significant changes in eating behaviours. Niederhauser, Maddock, LeDoux, and Arnold (2005) implemented a community-based wellness programme (“Building Strong and Ready Families”) targeting multiple health behaviours (e.g. physical activity, stress management, alcohol and tobacco use, safe sex) for both soldiers and their spouses (n = 245, Mage = 25.8). The conceptual framework of the intervention was the Transtheoretial Model of change, and content of 2-day-long sessions focused on education about the adverse effects of unhealthy behaviours, and introduction of skills and resources for healthy behaviour changes. Compared to a control condition, the intervention group demonstrated significant improvements in stress management, seat belt use, and tobacco cessation. Collectively, these interventions show some promise for improving a variety of health behaviours among military spouses. However, the authors also discussed multiple challenges. In particular, the face-to-face delivery methods created difficulties with participant recruitment and engagement. Because many © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 3 military spouses are sole childcare providers, carving out time to attend a health promotion programme in person may be difficult or impossible. One strategy for overcoming barriers to access associated with face-to-face programmes is to deliver interventions online, so participants can access the content when and where it is convenient for them (Murray, 2012). On the other hand, there may be great value in the social interaction that comes with face-to-face programmes, especially for military spouses, who commonly report feelings of loneliness and isolation (SteelFisher, Zaslavsky, & Blendon, 2008). Thus, an optimal approach may be to integrate online delivery with social interaction by providing opportunities for peer support among participants (Richardson et al., 2010). Many previous online interventions have incorporated a peer support component (e.g. through discussion boards), but the use of these components is often limited to a small percentage of participants (Van Mierlo, 2014). Furthermore, web-based interventions generally see steep declines in engagement, often after only a few weeks (Wangberg, Bergmo, & Johnsen, 2008). One potential approach to facilitate social interaction and engagement is to incorporate group dynamics strategies designed to build cohesion among participants. These strategies, which target group environment (e.g. optimal group size), group processes (e.g. collective goal setting), and group structure (e.g. establish roles and norms), have been shown to increase the effectiveness of physical activity interventions (Estabrooks, Harden, & Burke, 2012), but have rarely been used in online settings. If group dynamics strategies are successful at increasing online peer-topeer interaction, they may also increase engagement by keeping content fresh, providing ongoing participant support and encouragement, and facilitating regular prompts or “pushes” without much added burden for the research team, all of which have been associated with enhanced effectiveness of web-based interventions (Murray, 2012). Previous research has also found that health promoting interventions are more effective when they are theory based (Webb, Joseph, Yardley, & Michie, 2010). One theory that may be particularly relevant to military spouses is Self-Determination Theory (SDT) (Deci & Ryan, 2008). SDT identifies three core psychological needs that must be fulfilled for optimal functioning: autonomy, competence, and relatedness. A primary goal of SDT-based interventions is to facilitate autonomous motivation by fulfilling the three core needs (Ryan, Patrick, Deci, & Williams, 2008). For military spouses, who frequently report loss of autonomy due to their “dependent” status (Blakely, Hennessy, Chung, & Skirton, 2012), an intervention that empowers them to purposefully prioritise their own health and well-being may be very influential. Given the feelings of isolation and numerous other barriers to engaging in healthy behaviours that military spouses face, the effectiveness of the intervention would likely be enhanced by recommending small, manageable changes to promote feelings of competence, and providing a platform to facilitate meaningful connections among participants (i.e. relatedness). © 2019 The International Association of Applied Psychology 4 MAILEY ET AL. Thus, the purpose of this study was to examine the effects of an interactive, theory-based web-delivered intervention relative to a more generic, educational web-delivered intervention on mental health outcomes (stress, depression, anxiety, loneliness, self-esteem) and health behaviours (physical activity, diet) among military spouses. We hypothesised that both groups would report improvements in mental health, physical activity, and diet, but that there would be significant group by time interaction effects such that the magnitude of the improvements would be greater for participants who received the interactive intervention. Additionally, we hypothesised that website use would be higher in the interactive condition throughout the 10-week intervention. METHODS This study used a quasi-experimental design to compare an interactive, theorybased web-delivered intervention (InDependent but not Alone: A Journey to Finding Balance for Military Spouses; IBNA) to a more generic, educational web-delivered intervention (Operation Live Well; Control). Because the funding for the study was specifically designated to promote health in the local community, all local spouses who enrolled in the study were assigned to the IBNA condition. Spouses stationed elsewhere in the United States were simultaneously recruited for the control condition. Participants were not aware that there were two different interventions. Individuals were eligible to participate in the study if they were currently married to an active duty military person, had regular internet access, and were willing to commit to the 10-week programme. Additionally, all participants were required to pass the Physical Activity Readiness Questionnaire or sign a waiver indicating that they could safely engage in physical activity. All study procedures were approved by an Institutional Review Board in February 2016 (#8065). Recruitment began immediately and the intervention started in April 2016. Procedures Participants for the control condition were recruited primarily via social media. A brief blurb about the study was posted on the spouses/wives Facebook pages of various military installations throughout the United States. The posts included a link to a brief screening survey that described the study in more detail and assessed eligibility. A similar strategy was utilised for the IBNA condition; however, participants were also recruited locally by distributing flyers on post at the local Army base, attending monthly Family Readiness Group (FRG) roundup meetings, and sending information via FRG listservs. Snowball sampling was also used for both groups, whereby spouses who had enrolled in the programme were encouraged to invite their friends to participate. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 5 Individuals who were deemed eligible after completing the screening survey received a link to complete the baseline questionnaires via email, and an informed consent document along with a self-addressed envelope to return the document via US mail. Participants had to return the signed informed consent and complete the baseline questionnaires before they were officially enrolled in the study. Enrolled participants received an email informing them of the upcoming study start date. The interventions lasted 10 weeks and are described in detail below. At the beginning of each week, participants received an email with a link to the website, where all content was housed in weekly modules. The websites were accessible via computer, tablet, or smartphone. At the end of week 10, participants received a link to complete the follow-up questionnaires via email. Participants from both groups received a $20 gift card upon full completion of the follow-up questionnaires. Interventions Control. The control condition was based on the existing website, Operation Live Well. This website, developed and maintained by the United States Military Health System, includes articles and resources related to six focus areas: integrative wellness, mental wellness, nutrition, physical activity, sleep, and tobacco-free living. For the present study, our research team developed a separate website that housed weekly modules consisting of 3–4 direct links to Operation Live Well articles pertaining to one of the topics, as well as a vague prompt to “think about how you can apply this information in your own life”. Table S1 provides an outline of the weekly topics covered and examples of weekly modules for several of the topics. Each Monday, a new module was published and participants received an email introducing the topic for the week and reminding them to visit the website and view the content. The control website also included a discussion board where participants could discuss the material with other spouses in the programme. During week 1, the discussion board prompted participants to introduce themselves to the group; thereafter, participants were encouraged to “share your thoughts on this week’s content with the group”. Intervention (IBNA). The overarching goal of the IBNA intervention was to develop an engaging, interactive programme that would empower local military spouses to prioritise their own physical, emotional, and social health and well-being. The IBNA intervention primarily delivered content via podcasts. Each weekly module consisted of 3–5 podcasts that were specifically tailored to military spouses: one related to physical activity, one related to diet, and 1–3 related to personal growth or emotional well-being. The substantial focus on emotional wellness was based on our formative research (focus groups), during which military spouses indicated that improving mental well-being was their top © 2019 The International Association of Applied Psychology 6 MAILEY ET AL. priority. Thus, the other health behaviours were also discussed in the context of mental and emotional benefits. Each podcast had a corresponding workbook activity designed to stimulate participants to actively engage with the content and think about and apply the content to their own lives. The workbook activities could be accessed and downloaded via the website each week. Table S2 provides a detailed outline of intervention content, including references to the specific behaviour change techniques incorporated each week based on Michie’s taxonomy (Michie et al., 2013). In addition, sample workbook activities for each of the three main content areas are provided in Appendix S1. The intervention was based on Self-Determination Theory (SDT) and was designed to promote competence, autonomy, and relatedness among participants. To promote competence, participants were encouraged to make small, feasible changes to improve their physical activity, diet, social connection, and emotional well-being. The website facilitated these changes by challenging participants to complete 3–5 “mini-tasks” related to these behaviours each week (e.g. replace one sugar sweetened beverage with water each day [diet], go for a 10-minute walk [physical activity], call a friend [social connection], write down ten things you are grateful for [emotional well-being]). To promote relatedness, the intervention incorporated group dynamics strategies to facilitate peer-to-peer support and cohesion among participants (Estabrooks et al., 2012). Specifically, each participant was placed on a team with 4–6 other spouses. Participants had the option to request teammates if they were enrolling in the intervention with friends; otherwise, teams were created based on common characteristics (e.g. location of residence, age, children’s ages, etc.). Each team had a captain, whose role was to motivate and encourage the other teammates, and to collect weekly reports of the number of challenges each individual had completed. Captains were volunteers who received a brief (10-minute) training prior to the start of the intervention. Teams were encouraged to set weekly group goals, and to meet up in person throughout the 10-week intervention to provide each other with additional support. Participants also had access to a website discussion board where they could connect with all participating spouses. To facilitate autonomy, the intervention prompted participants to identify their core values, and to align health behaviours such as diet and physical activity with those values, as opposed to reasons they think they “should” engage in these behaviours, such as weight loss. Throughout the intervention, participants had substantial autonomy regarding the behaviour changes on which they wanted to focus, and what goals they wanted to set and pursue. The IBNA intervention began with a live kick-off event. All participants were invited to attend this in-person event to learn about the 10-week programme and meet their teammates and other programme participants. During the kick-off event, participants engaged in several team-building activities with their teammates, and developed a plan for communicating with each other throughout the 10-week programme. Team captains received contact information for any team © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 7 members who were unable to attend the kick-off. Following this event, all intervention content was delivered via the website. Measures Demographics and Baseline Motivation. Demographic information reported included gender, age, race, education, employment status, number of children, years married, and spouse’s rank and deployment status. Participants also reported their interest in changing the targeted health behaviours (physical activity, diet, and stress) on a scale from 1 (not at all interested) to 3 (very interested). Mental Health Outcomes. Mental health outcomes assessed included stress, anxiety, depression, loneliness, and self-esteem. Stress was assessed using the 10-item version of the Perceived Stress Scale (Cohen, Kamarck, & Mermelstein, 1983), which measures the degree to which an individual appraises her life circumstances as stressful. The Hospital Anxiety and Depression Scale (Zigmond & Snaith, 1983) was used to assess anxiety and depression. Each subscale consists of seven items that reflect common symptoms of anxiety and depression in general non-psychiatric populations. The 20-item UCLA Loneliness Scale assessed the extent to which individuals experienced feelings of loneliness (Russell, Peplau, & Cutrona, 1980). Finally, overall self-esteem was assessed using the 10-item Rosenberg Self-Esteem Scale (Rosenberg, 1965). All measures have been widely used and validated in adult populations, and all are scored using Likert scales that ask participants to report the extent to which they agree or disagree with the statements. Internal consistency of all scales was acceptable at each time point and ranged from a = 0.77 (anxiety) to a = 0.94 (loneliness). Physical Activity. Physical activity was assessed using the long version of the International Physical Activity Questionnaire (IPAQ; Hagstromer, Oja, & Sjostrom, 2006). Participants reported the frequency and duration of moderate activity, vigorous activity, and walking in different domains including work-related physical activity, transportation-related physical activity, domestic and gardening activities, and leisure-time physical activity. Activities were multiplied by metabolic equivalent (MET) values to determine MET-minutes/week, then summed across domains and intensities. A total physical activity score encompassing all activities was also calculated. To address outliers, data were carefully screened, and truncation rules were applied such that no single domain/intensity variable (e.g. work-related vigorous activity, leisure-time walking) exceeded 4 hours per day. Reported durations that exceeded this threshold were recoded to equal 240 minutes (4 hours) prior to multiplying the duration by frequency and intensity values. © 2019 The International Association of Applied Psychology 8 MAILEY ET AL. Diet. Participants’ typical dietary intake was assessed using the Dietary Screener Questionnaire (DSQ; Thompson, Midthune, Kahle, & Dodd, 2017). This 30-item questionnaire was developed by the National Cancer Institute, Division of Cancer Control & Population Sciences to capture consumption of specific dietary components. Participants reported the frequency of consuming various foods and drinks during the past month. Estimated intake of fruits and vegetables, added sugar, added sugar from sugar sweetened beverages (SSBs), fibre, whole grains, calcium, and dairy were compiled from the DSQ using SAS version 9.4 software code provided by the NCI: http://epi.grants.cancer.gov/nhanes/ dietscreen/questionnaires.html. Engagement. Two indicators of engagement were obtained directly from the websites: number of page views per week and number of discussion board posts per week. Page views were summed across all 10 weeks to yield a “Total Page Views” variable. Programme Evaluation. At the beginning of the post-intervention survey, participants completed a programme evaluation. Participants in both groups responded to a set of identical questions rating their overall perceptions of the programme (e.g. the online format, time commitment, programme staff, etc.) on a scale from 1 (strongly disagree) to 5 (strongly agree). Participants in the IBNA condition responded to additional questions about specific aspects of the intervention, including the website, podcasts, workbook, team component, and three content areas. They provided both quantitative ratings of these components and qualitative comments about what they liked and what they would change about each component. Data Analysis Distributions of all outcomes were examined for normality. Because all variables were skewed, square root transformation was conducted. This resulted in a normal distribution for the mental health and diet outcomes. Independent samples ttests were used to determine whether there were any significant baseline differences between the intervention and control groups, as well as between participants who dropped out vs. those who completed the post-intervention measures. To evaluate changes in the mental health and diet outcomes, linear mixed effects models that can account for the subject-specific association in repeated measurements were used to examine the effects of the intervention as well as other explanatory variables/covariates on these outcomes. Coupled with the use of a sandwich estimator, the estimated standard errors were then robust against any misspecification of the association structure assumed in the working models, which therefore yielded more reliable statistical inferences (Fitzmaurice, Ware, & Laird, 2004). In our analyses, the explanatory variables were Wave, Group, © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 9 Wave*Group and Total Page Views, where Wave reflected changes over time (from pre- to post-intervention) and Wave*Group reflected interaction effects which were used to evaluate any differential effects of the intervention between groups over time. Total Page Views was incorporated into the models as an indicator of engagement with a proper transformation. All models were run with complete cases only and with all available observations by imputing missing post-intervention data in the linear mixed effects models (intent-to-treat analysis). To further evaluate the associations between engagement and outcomes in the intervention group, a simple dose–response analysis was performed using Total Page Views as the dose variable and change (post–pre) in the outcome variable as the response. For the physical activity outcomes, due to the large number of zeros present in the data, square root transformation did not produce a normal distribution for most outcomes. Thus, zero-inflated mixed models were used to accommodate not only the extra zeros, but also the correlations among the pre- and post-test observations. First, non-zero observations were transformed using a log transformation to establish the baseline distribution in the zero-inflated model. Second, the mixture probability parameter was assumed to represent the probability of having an extra zero. The models assumed that the overall average of the nonzero observations was dependent upon these covariates: Wave, Group, and Wave*Group, and Total Page Views. To account for multiple outcomes within each category (i.e. mental health, diet, and physical activity), we used the Hochberg and Benjamini (1990) adaptive step-up Bonferroni method to adjust for the inflated type I error rate in the multiple testing procedure. All p-values reported in the tables for the linear mixed effects models and the zero-inflated models are Hochberg-Benjamini adjusted p-values. All analyses were conducted in SAS 9.4. RESULTS Participant Characteristics and Retention Participant flow through the study is depicted in Figure 1. A total of 119 spouses completed the baseline questionnaires and were enrolled in the IBNA programme; 112 spouses enrolled in the control programme. The demographic characteristics of the two groups are presented in Table 1. More IBNA participants had a spouse that was currently deployed; no other demographic characteristics differed significantly between groups. Although male spouses were not excluded from participating, all enrolled participants were female. Participants in both groups were primarily White and college-educated. A majority of participants (72.3%) had children, and a minority (26.8%) were employed full-time. Just over half of participants in each group (54%) were married to an officer; the © 2019 The International Association of Applied Psychology 10 MAILEY ET AL. remainder were spouses of enlisted soldiers. Reponses to the baseline motivation questions revealed that participants in both groups were highly motivated to change their health behaviours; means ranged from 2.59 to 2.79 on the 1–3 scale. Baseline motivation to increase physical activity was significantly higher among IBNA participants; motivation to change diet or stress did not differ between groups. The percentages of participants who completed the follow-up measures were 67.2 per cent and 67.0 per cent in the intervention and control groups, respectively. Participants did not provide reasons for not completing the follow-up measures. Within the intervention group, participants who dropped out of the study (n = 39) reported significantly higher levels of anxiety and depression, and significantly lower self-esteem than those who completed the follow-up measures (n = 80) at baseline. There were no significant baseline differences between completers (n = 75) and dropouts (n = 37) in the control group. Intervention Effects on Health Outcomes and Behaviours The baseline scores of the outcome measures did not differ significantly between the intervention and control groups (all p > .05). For mental health outcomes, the results are given in Table 2. The covariate Wave was significantly associated with the outcomes (all adj. p < .05), indicating that the outcomes differed FIGURE 1. Participant flow diagram. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 11 TABLE 1 Demographic Characteristics of the Sample Mean (SD)/Freq (%) Variable Gender: Female Age Years married Spouse rank Enlisted Officer Spouse deployed Employment status Full-time Part-time Full-time homemaker Other Number of children 0 1 2 3 or more Race: White Education High school graduate College graduate Advanced degree Interest in changing physical activity* Interest in changing diet* Interest in reducing stress* IBNA group (n = 119) Control group (n = 112) 119 (100%) 31.89 (6.11) 7.74 (5.63) 112 (100%) 33.06 (6.33) 8.31 (6.18) 51 (43%) 64 (54%) 53 (45%) 45 (40%) 64 (54%) 31 (28%) 36 11 52 20 (30%) (9%) (44%) (17%) 26 26 44 16 (23%) (23%) (39%) (14%) 33 20 31 35 108 (28%) (17%) (26%) (29%) (91%) 31 27 33 21 104 (28%) (24%) (29%) (19%) (93%) 26 65 28 2.79 2.66 2.72 (22%) (55%) (23%) (0.45) (0.49) (0.50) 19 54 39 2.65 2.59 2.68 (17%) (48%) (35%) (0.57) (0.53) (0.52) *Response options included (1) not at all interested, (2) somewhat interested, and (3) very interested. between baseline and follow-up, such that both groups reported increases in selfesteem and reductions in stress, anxiety, depression, and loneliness. However, examination of the Wave*Group interaction revealed that there were no differences in the mental health outcomes between the intervention and control groups over time (all adj. p > .05). Total page views was not significant for all mental health outcomes, suggesting that engagement was not associated with these outcomes. Results did not differ between the complete case analysis and the intentto-treat analysis. For physical activity (Table S3), the mixture probabilities were significant for all outcomes, which suggests that the existence of zero inflation in the data and the use of zero-inflated models was appropriate. Significant time effects (i.e. Wave) were observed for work physical activity, household physical activity, moderate physical activity, vigorous physical activity, and total physical activity, © 2019 The International Association of Applied Psychology 12 MAILEY ET AL. such that these types of activity increased from pre- to post-intervention. However, the coefficients of the interaction term (i.e. Wave*Group) were not significant, suggesting that the changes did not differ between the IBNA and control groups. No significant changes in walking, leisure-time activity, or physical activity for transport were reported. Total page views were negatively associated with work physical activity, physical activity for transport, household physical activity, moderate physical activity, and walking, suggesting that participants who visited the website more reported larger decreases in these types of physical activity. For diet (Table S4), participants reported significant decreases in consumption of sugar sweetened beverages and total sugar post-intervention compared to baseline. No significant changes in consumption of fruits and vegetables, dairy, whole grains, or fibre were reported. Additionally, none of the Wave*Time interaction effects were significant. When missing values were imputed, total page TABLE 2 Mental Health Outcomes from Linear Mixed Effects Models: The Estimate (SE), Adjusted p-value† Given Underneath Outcome Stress Anxiety Depression Loneliness Selfesteem Dataset Complete cases Data with imputation Complete cases Data with imputation Complete cases Data with imputation Complete cases Data with imputation Complete cases Data with imputation # of subjects 154 230 154 230 154 230 154 230 154 230 Group0: Control1: IBNA Wave0: Pre1: Post Group*Wave Total page views** 0.22 (0.14) 0.096 0.18 (0.11) 0.491 0.16 (0.12) 0.202 0.01 (0.10) 0.908 0.18 (0.17) 0.305 0.03 (0.13) 0.908 0.31 (0.16) 0.060 0.14 (0.13) 0.905
0.20 (0.10) 0.044
0.10 (0.08) 0.905
0.57 (0.10) <0.001
0.59 (0.07) <0.001
0.46 (0.07) <0.001
0.50 (0.05) <0.001
0.46 (0.08) <0.001
0.51 (0.06) <0.001
0.53 (0.07) <0.001
0.57 (0.05) <0.001 0.20 (0.04) <0.001 0.23 (0.03) <0.001 0.07 (0.14) 0.830 0.10 (0.10) 0.734 0.03 (0.12) 0.830 0.10 (0.08) 0.734
0.11 (0.14) 0.830
0.04 (0.09) 0.734 0.03 (0.12) 0.830 0.07 (0.08) 0.734 0.06 (0.07) 0.830 0.02 (0.05) 0.734
0.03 (0.04) 0.943
0.02 (0.02) 0.601
0.01 (0.04) 0.943
0.04 (0.02) 0.281 0.004 (0.05) 0.943
0.03 (0.01) 0.601 0.04 (0.05) 0.943 0.02 (0.03) 0.601 0.01 (0.03) 0.943 0.01 (0.02) 0.601 Boldface text indicates statistical significance; SE, standard error; IBNA, Independent but not Alone. †Hochberg and Benjamini (1990) adaptive step-up Bonferroni method. pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi **3 Total Page Views. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 13 views was negatively associated with sugar intake, indicating that participants who exhibited higher levels of engagement reported larger decreases in sugar consumption. The raw (non-transformed) means, standard deviations, and effect sizes for all outcomes (complete cases only) are included in Table S5. Effect size calculations revealed that improvements in mental health outcomes were moderate, with control group changes ranging from d = 0.30 (stress) to d = 0.64 (depression), and intervention group changes ranging from d = 0.41 (self-esteem) to d = 0.62 (stress). Increases in physical activity were generally small. Overall, the greatest improvements were observed for household physical activity, moderate physical activity, and total physical activity, with effect sizes ranging from d = 0.29 to d = 0.36 for these outcomes in both groups. Dietary changes were also small and unlikely to be of clinical significance, with the exception of reductions in consumption of sugar and sugar sweetened beverages, which exhibited moderate effects (d = 0.42–0.48). Website Use Of the 119 IBNA participants, 106 (89.1%) accessed the website at least once during the 10-week intervention, and 63 (52.9%) posted on the discussion board at least once, although only 28 participants (23.5%) had more than three discussion board posts. Of the 112 control participants, 77 (68.9%) accessed the website at least once, and 50 (44.6%) posted on the discussion board at least once, although only 29 (25.9%) posted more than once during the 10-week intervention period. The total number of page views per participant was significantly higher in the IBNA group (72.6) than the control group (21.3) across the 10week intervention. However, as seen in Figure S1, website use declined significantly as the intervention progressed, such that each group averaged less than two page views per participant by week 10. The dose–response analyses conducted for the IBNA group only did not yield any significant results, which suggests that there was no statistical evidence to support the existence of relationships between website use and the outcomes. Programme Evaluation Table S6.a displays the mean ratings from the programme evaluation completed by both intervention and control participants. Overall, evaluations were moderately positive, and not significantly different between groups, with one exception: participants in the control condition reported the time commitment of the programme to be more manageable than participants in the IBNA condition. Table S6.b presents the results of the more detailed evaluation completed by IBNA participants. The table includes a sample of participant quotes that represent the most common comments on the evaluation. Overall, participants © 2019 The International Association of Applied Psychology 14 MAILEY ET AL. identified many positive aspects of the intervention, but also many areas for improvement. It was apparent from the evaluations that due to the breadth of topics covered by the intervention, participants felt that the time commitment was extensive, and they would have preferred an in-depth focus on a few topics, rather than more shallow, “basic” coverage of many topics. Most participants reported that the website was easy to use and appreciated that they could access the podcasts at their convenience. The online format of the workbook appeared to be less convenient and thus less widely utilised; many participants expressed a preference for a printed workbook and reported that completing written assignments was time-consuming and felt like “homework”. Feedback on the team aspect of the intervention was mixed and clearly dependent upon whether participants connected with their teammates or not. Participants recommended providing more training for team captains and placing more emphasis on common characteristics (e.g. age of children, employment status) when forming teams to better facilitate cohesion among members. Some participants also suggested starting with larger teams so a core group of engaged members would remain even if a few members dropped out or were not actively engaged. In general, participants universally expressed appreciation for the development of a programme specifically for military spouses, but their comments highlighted the difficulty of meeting every individual’s needs without more extensive tailoring. DISCUSSION This study examined the effectiveness of an interactive, theory-based intervention to promote health among military spouses relative to the existing standard of care: a more generic educational website with links to health-promoting articles and resources. The interactive intervention underwent substantial formative research to design content that was highly tailored to the unique needs and preferences of military spouses. The website aimed to keep participants engaged by incorporating interactive features such as podcasts that could be easily accessed anywhere and anytime, weekly challenges focused on small, manageable changes, and teams to facilitate peer-to-peer support. Following the 10-week intervention, participants reported improved mental health and some modest increases in physical activity and reductions in sugar consumption; however, the magnitude of these changes was comparable for participants in the comparison programme. Additionally, although intervention participants accessed their website more than control participants overall, website engagement declined substantially over the course of the programme. Thus, the results of this study suggest that web-based interventions may elicit short-term improvements in mental health among military spouses. The extent to which an interactive, theory-based intervention is necessary to produce such changes in questionable, as improvements were also observed from a more minimal intervention. These findings are discussed in detail below. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 15 To the best of our knowledge, this is the first intervention among military spouses to use an “active” control condition as opposed to a mere-measurement control group. Thus, it builds upon previous health behaviour interventions for military spouses by comparing the theory-based intervention to a programme that delivers content that health-motivated individuals may seek out and access on their own. In so doing, however, it makes it difficult to determine the extent to which any observed changes were due to social desirability. The improvements in the control condition were unexpected, because although control participants did receive content related to health behaviours, previous research has shown that interventions primarily focused on provision of information are insufficient to produce substantial changes in behaviour (Rothert et al., 2006; Silva et al., 2010). One potential explanation for the unexpected improvements in the control group is that individuals who enrolled in the study were highly motivated to invest in improving their health. The control intervention was advertised as a health promotion programme, so it may have attracted individuals who were already interested in changing their diet and/or physical activity behaviours, and provided the “spark” they needed to initiate those changes. Participants unanimously expressed appreciation for having something “just for them”, so it may be that as a group, military spouses feel so neglected that simply having something designed specifically for them provided a boost in self-care behaviours and mental health outcomes. Furthermore, fewer control participants’ spouses were deployed when they enrolled in the intervention, so they may have had fewer barriers to improving health behaviours than intervention participants. In spite of the improvements observed, this study encountered many of the challenges that are typical of web-based interventions. In particular, participant engagement declined rapidly in both groups, many participants “lurked” but did not actively contribute to discussions, and the self-guided nature of the intervention likely contributed to substantial heterogeneity in website use and behavioural outcomes. Achieving the optimal balance between automation and individualised support is an ongoing challenge of web-based interventions. Although there is some evidence that providing guidance (e.g. via an e-coach) enhances intervention effectiveness (Baumeister, Teichler, Munzinger, & Lin, 2014), it is important to weigh the magnitude of these improvements against the time and money required to provide tailored support. In the present study, incorporating some degree of tailoring or stage-matched content may have been beneficial, particularly for the physical activity and nutrition content, which was purposely designed to target “beginners” and promote very small changes (e.g. taking a 10-minute walk). However, because many participants were already motivated and health-conscious, such recommendations were likely too low in intensity to meet their needs. Additionally, the workbook was a primary tool for delivering the behaviour change content and engaging participants in strategies such as goal setting, planning, and self-monitoring. Unfortunately, results from the programme evaluation suggest that the workbook was sparsely used, perhaps © 2019 The International Association of Applied Psychology 16 MAILEY ET AL. because of the online format or because there was no mechanism for getting “credit” or receiving feedback from these activities. Generally, research has supported the notion that participants who are more engaged in the intervention exhibit greater improvements in behavioural and health outcomes (Glasgow et al., 2011; Richardson et al., 2013). In this study, website use was largely unrelated to study outcomes, with the exception of some unexpected negative relationships between page views and physical activity when both groups were combined for the analyses. One potential explanation for these findings is that individuals who were inclined to be more sedentary (with greater screen time) had more time to access the website. However, the manner in which engagement is conceptualised and measured is important. Simply using website use/access as an indicator may be problematic, as it does not provide insight into offline engagement with intervention content (Yardley et al., 2016). It is possible that using workbook completion as an indicator of engagement may have yielded different findings because many of the “active ingredients” of the intervention were delivered through this mechanism. The drop in engagement observed in this study may be at least partially attributable to the substantial number of participant requirements associated with the IBNA intervention (Glasgow et al., 2007). The intervention was designed to be comprehensive and incorporate a variety of topics and delivery strategies to meet the needs of a variety of spouses. Although some participants viewed this as a strength because they were able to find something with which they could connect, others reported that the amount of content was overwhelming to them. Each week consisted of 3–5 podcasts, corresponding workbook activities, 3–5 challenges, and a recommended team meet-up. Thus, many participants were unable to commit the necessary time to complete the programme, despite their best intentions to do so when they enrolled. Although frequent content updates and interactive features such as peer and/or counsellor support have been associated with increased adherence to web-based interventions (Kelders, Kok, Ossebaard, & Van Gemert-Pijnen, 2012), adding complexity to interventions may be overwhelming to some participants, particularly those who have many other competing demands (Danaher & Seeley, 2009). In the present study, almost half of the IBNA participants had a spouse who was currently deployed, so they were already managing work, childcare, and household responsibilities on their own. Military spouses also frequently encounter unexpected challenges and demands (Dimiceli et al., 2010), and the IBNA programme was perhaps not flexible enough to accommodate these challenges. Although participants were encouraged to go at their own pace and revisit past modules as time allowed, many reported that once they fell behind, they felt powerless to catch up, and in some cases discontinued their participation completely. A key strategy to promote sustained engagement in the IBNA programme was the team component. We incorporated a variety of group dynamics strategies such as limiting group size, assigning team captains, creating team © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 17 names, facilitating group goal setting, and promoting friendly competition (Estabrooks et al., 2012). For some teams, these strategies provided the framework for a positive group experience, and group members interacted frequently and developed lasting friendships. Other teams, however, never connected, either because they felt they had little in common, the logistics of meeting up proved to be too difficult, or their captain was not an effective leader. Previous research has shown that group leader characteristics can significantly impact group members’ sense of empowerment and efficacy, and ultimately the cohesiveness and performance of the group (Jung & Sosik, 2002). In the present study, captains’ engagement and effectiveness appeared to be integrally tied to the group’s success. While some individuals flourished in the captain role, others felt that it was an added burden that they were not prepared to take on. Future studies employing group dynamics strategies should consider enrolling and training captains in advance to ensure that they have the motivation and resources to lead their teams effectively (Kelloway, Barling, & Helleur, 2000). Alternatively, the research team could appoint an individual to provide tailored support to take the burden off participants and ensure that all groups receive similar support and leadership (Inauen et al., 2017). Future studies might also explore whether assigned versus self-selected teams are optimal for promoting cohesion. Facilitating a positive team experience may be a key determinant of the ultimate success of the intervention. It is important to acknowledge several key limitations of this study. First, the nature of the funding mechanism necessitated a quasi-experimental design, whereby participants were assigned to the intervention or control condition based on their current location of residence. Although there were few significant differences between the groups at baseline, it is not possible to rule out the influence of systematic differences between the two study groups on intervention outcomes. Furthermore, the extent to which different recruitment strategies (i.e. in person vs. online) contributed to the differences in engagement observed cannot be determined. Second, all outcomes were self-reported and participants were highly motivated to change the target behaviours, and thus social desirability bias may have contributed to the improvements observed in both groups. Although all measures have been validated and widely used, the IPAQ is notoriously subject to over-reporting and the dietary screener is less accurate than diary methods. Furthermore, the DSQ outcomes may not have aligned well with some of the changes recommended in the intervention; thus, some behaviour changes may not have been captured by this measure. Outcomes were only assessed at baseline and immediately post-intervention, so we cannot draw any conclusions about whether the short-term improvements observed were maintained over time. Additionally, although the intervention was guided by SDT, we did not incorporate measures of autonomy, competence, and relatedness. Thus, it is not possible to determine whether the lack of added effects of the © 2019 The International Association of Applied Psychology 18 MAILEY ET AL. IBNA intervention are attributable to a failure to impact the core theoretical constructs that were hypothesised to drive changes. Finally, these results cannot be generalised to all military spouses. Despite substantial efforts to recruit spouses of lower enlisted soldiers and spouses with limited motivation and/or resources to improve their health, the final sample consisted of primarily white, highly educated, highly motivated spouses. Furthermore, military spouses experiencing high levels of stress, depression, and loneliness may not have taken the initiative to enrol in a health promotion programme. Future studies should continue to explore strategies for engaging unmotivated or ambivalent individuals in webbased health promotion programmes. This study also had a number of strengths. Although not randomly assigned, the inclusion of a “usual care” control group is considered superior to a no-treatment or waitlist control group, as it helped account for expectancy effects and allowed us to determine the relative effectiveness of the active ingredients of the interactive, theory-based intervention (Danaher & Seeley, 2009). Indeed, the results of this study differed from a number of previous interventions for military spouses, which demonstrated some significant improvements relative to no-treatment control groups. Thus, the extent to which theory-based content adds to the effectiveness of interventions for this population remains unclear, and future studies should aim to identify optimal website features, delivery strategies, and behaviour change techniques for maximising the effectiveness of health-promoting interventions for military spouses. The intervention content was based on substantial formative research with the target population, and military spouses were part of the team that developed and delivered the intervention. This allowed the content to be highly tailored to the target population, and participants indicated that this was critical to them feeling understood. The use of podcasts to deliver intervention content was a novel strategy that enhanced accessibility and has rarely been used in health promoting interventions. Future studies should continue to explore optimal features of web-based intervention to promote sustained engagement and interaction. Overall, the results of this study suggest that web-based health promotion programmes may help support positive changes in health behaviours and mental health among military spouses. In spite of documented high levels of stress, anxiety, depression, and obesity in this population, few health promotion programmes have been developed specifically for military spouses. Web-based interventions may be a promising avenue to provide easily accessible content and support to this population, but researchers must be mindful of the challenges associated with online programmes. In the present study, an in-depth, interactive, theory-based intervention was no more effective than an information-based intervention that required a more modest time commitment. Future studies should aim to determine the minimum “dose” needed to elicit meaningful changes in this population, and examine the extent to which any changes are sustained over time. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 19 ACKNOWLEDGEMENTS This study was funded by the Center for Engagement and Community Development at Kansas State University. The authors thank Michele Bradfield and Danielle Corenchuk for their contributions to the study. CONFLICT OF INTEREST There are no conflicts of interest to disclose. REFERENCES Baumeister, H., Teichler, L., Munzinger, M., & Lin, J. (2014). The impact of guidance on internet-based mental health interventions: A systematic review. Internet Interventions, 1, 205–215. Blakely, G., Hennessy, C., Chung, M.C., & Skirton, H. (2012). A systematic review of the impact of foreign postings on accompanying spouses of military personnel. Nursing & Health Sciences, 14, 121–132. Blakely, G., Hennessy, C., Chung, M.C., & Skirton, H. (2014). Adaption and adjustment of military spouses to overseas postings: An online forum study. Nursing & Health Sciences, 16, 387–394. Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. Cole, R.E., & Horacek, T. (2010). Effectiveness of the “My Body Knows When” intuitive-eating pilot program. American Journal of Health Behavior, 34, 286–297. Danaher, B.G., & Seeley, J.R. (2009). Methodological issues in research on web-based behavioral interventions. Annals of Behavioral Medicine, 38, 28–39. Deci, E.L., & Ryan, R.M. (2008). Self-determination theory: A macrotheory of human motivation, development, and health. Canadian Psychology, 49, 182–185. Department of Defense (2015). 2015 demographics report: Profile of the military community. Retrieved from: http://download.militaryonesource.mil/12038/MOS/Reports/ 2015-Demographics-Report.pdf Dimiceli, E.E., Steinhardt, M.A., & Smith, S.E. (2010). Stressful experiences, coping strategies, and predictors of health-related outcomes among wives of deployed military servicemen. Armed Forces & Society, 36, 351–373. Estabrooks, P.A., Harden, S.M., & Burke, S.M. (2012). Group dynamics in physical activity promotion: What works? Social and Personality Psychology Compass, 6, 18–40. Fish, T.L. (2013). Risk of being overweight or obese among Army spouses: The impact of deployment, distress, and perceived social support. Unpublished doctoral dissertation. Baltimore, MD: University of Maryland. Fitzmaurice, G.M., Ware, J.H., & Laird, N. (2004). Applied longitudinal analysis. Hoboken, NJ: John Wiley and Sons. Glasgow, R.E., Christiansen, S.M., Kurz, D., King, D.K., Woolley, T., Faber, A.J., . . . Dickman, J. (2011). Engagement in a diabetes self-management website: Usage patterns and generalizability of program use. Journal of Medical Internet Research, 13, e9. © 2019 The International Association of Applied Psychology 20 MAILEY ET AL. Glasgow, R.E., Nelson, C.C., Kearney, K.A., Reid, R., Ritzwoller, D.P., Strecher, V.J., . . . Wildenhaus, K. (2007). Reach, engagement, and retention in an Internet-based weight loss program in a multi-site randomized controlled trial. Journal of Medical Internet Research, 9, e11. Hagstromer, M., Oja, P., & Sjostrom, M. (2006). The International Physical Activity Questionnaire (IPAQ): A study of concurrent and construct validity. Public Health Nutrition, 9, 755–762. Hochberg, Y., & Benjamini, Y. (1990). More powerful procedures for multiple significance testing. Statistics in Medicine, 9, 811–818. Inauen, J., Bolger, N., Shrout, P.E., Stadler, G., Amrein, M., Rackow, P., & Scholz, U. (2017). Using support group messaging to promote healthy eating: A randomized trial in daily life. Applied Psychology: Health and Well-Being, 9, 303–323. Jung, D.I., & Sosik, J.J. (2002). Transformational leadership in work groups: The role of empowerment, cohesiveness, and collective-efficacy on perceived group performance. Small Group Research, 33, 313–336. Kees, M., & Rosenblum, K. (2015). Evaluation of a psychological health and resilience intervention for military spouses: A pilot study. Psychological Services, 12, 222–230. Kelders, S.M., Kok, R.N., Ossebaard, H.C., & Van Gemert-Pijnen, J.E.W.C. (2012). Persuasive system design does matter: A systematic review of adherence to web-based interventions. Journal of Medical Internet Research, 14, e152. Kelloway, E.K., Barling, J., & Helleur, J. (2000). Enhancing transformational leadership: The roles of training and feedback. Leadership & Organization Development Journal, 21, 145– 149. Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W., . . . Wood, C.E. (2013). The behavior change technique taxonomy (v1) of 93 hierarchically-clustered techniques: Building an international consensus for the reporting of behavior change interventions. Annals of Behavioral Medicine, 46, 81–95. Murray, E. (2012). Web-based interventions for behavior change and self-management: Potential, pitfalls, and progress. Medicine 2.0, 1 , e3. Niederhauser, V.P., Maddock, J., LeDoux, F., & Arnold, M. (2005). Building strong and ready Army families: A multirisk reduction health promotion pilot study. Military Medicine, 170, 227–233. Richardson, C.R., Buis, L.R., Janney, A.W., Goodrich, D.E., Sen, A., Hess, M.L., . . . Piette, J.D. (2010). An online community improves adherence in an internet-mediated walking program. Part 1: Results of a randomized controlled trial. Journal of Medical Internet Research, 12, e71. Richardson, A., Graham, A.L., Cobb, N., Xiao, H., Mushro, A., Abrams, D., & Vallone, D. (2013). Engagement promotes abstinence in a web-based cessation intervention: Cohort study. Journal of Medical Internet Research, 15, e14. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Rothert, K., Strecher, V.J., Doyle, L.A., Caplan, W.M., Joyce, J.S., Jimison, H.B., & Roth, M.A. (2006). Web-based weight management programs in an integrated health care setting: A randomized, controlled trial. Obesity, 14, 266–272. © 2019 The International Association of Applied Psychology WEB-BASED INTERVENTION FOR MILITARY SPOUSES 21 Russell, D., Peplau, L.A., & Cutrona, C.E. (1980). The revised UCLA loneliness scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472–480. Ryan, R.M., Patrick, H., Deci, E.L., & Williams, G.C. (2008). Facilitating health behavior change and its maintenance: Interventions based on Self-Determination Theory. European Psychologist, 10, 2–5. Silva, M.N., Vieira, P.N., Coutinho, S.R., Claudia, M.S., Matos, M.G., Sardinha, L.B., & Teixeira, P.J. (2010). Using self-determination theory to promote physical activity and weight control: A randomized controlled trial in women. Journal of Behavioral Medicine, 33, 110–122. SteelFisher, G.K., Zaslavsky, A.M., & Blendon, R.J. (2008). Health-related impact of deployment extensions on spouses of active duty Army personnel. Military Medicine, 173, 221–229. Suarez, N.J. (2011). Exploring healthy lifestyle behaviors, deployment factors, and adjustment among military spouses. Unpublished doctoral dissertation. Manoa, HI: University of Hawaii. Tenconi, D. (2011). Obesity and US military spouses: An examination of risk perception and health behavior. Unpublished master's thesis. Manhattan, KS: Kansas State University. Thompson, F.E., Midthune, D., Kahle, L., & Dodd, K.W. (2017). Development and evaluation of the National Cancer Institute's Dietary Screener Questionnaire scoring algorithms. Journal of Nutrition, 147, 1226–1233. Van Mierlo, T. (2014). The 1% rule in four digital health social networks: An observational study. Journal of Medical Internet Research, 16, e33. Verdeli, H., Baily, C., Vousoura, E., Belser, A., Singla, D., & Manos, G. (2011). The case for treating depression in military spouses. Journal of Family Psychology, 25, 488– 496. Wangberg, S.C., Bergmo, T.S., & Johnsen, J.K. (2008). Adherence in internet-based interventions. Patient Preference and Adherence, 2, 57–65. Webb, T.L., Joseph, J., Yardley, L., & Michie, S. (2010). Using the internet to promote health behavior change: A systematic review and meta-analysis of the impact of theoretical basis, use of behavior change techniques, and mode of delivery on efficacy. Journal of Medical Internet Research, 12, e4. Yardley, L., Spring, B.J., Riper, H., Morrison, L.G., Crane, D.H., Curtis, K., . . . Blandford, A. (2016). Understanding and promoting effective engagement with digital behavior change interventions. American Journal of Preventive Medicine, 51, 833–842. Zigmond, A.S., & Snaith, R.P. (1983). The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica, 67, 361–370. SUPPORTING INFORMATION Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. (a) Weekly intervention content: Control condition. (b) Examples of weekly modules received by the control condition © 2019 The International Association of Applied Psychology 22 MAILEY ET AL. Table S2. (a) Detailed weekly intervention content: personal growth. (b) Detailed weekly intervention content: nutrition. (c) Detailed weekly intervention content: physical activity Appendix S1. (A) Sample Personal Growth Workbook Pages. (B) Sample Physical Activity Workbook Pages. (C) Sample Nutrition Workbook Pages Table S3. Physical activity outcomes from zero inflated models (with all available data) Table S4. Diet outcomes from linear mixed effects models: the estimate (SE), adjusted p-value† given underneath Table S5. Raw means, standard deviations, and effect sizes for all outcomes (complete cases only) Figure S1. Website use (expressed as average number of page views per week by group) across the10-week intervention Table S6. Programme Evaluation Results © 2019 The International Association of Applied Psychology