The Relationship Between Perceived Barriers for Physical Activity and Depression in Low Active Community-Dwelling Women


 Background: Women are less physically active, report greater perceived barriers for exercise and show higher levels of depression. This contributes to high global disability. The relationship between perceived barriers for physical activity and depressive symptoms in women remains largely unexplored. The aims of this cross-sectional analysis were to examine the association between physical activity barriers and depressive symptoms, and identify types of barriers in low active community dwelling women. Methods: 318 physically inactive women aged 25-65 years completed the Barriers to Being Active Quiz (BBAQ) developed by the Centers for Disease Control and Prevention, and the Center for Epidemiological Studies Depression Scale (CES-D) at the baseline visit of the mobile phone based physical activity education (mPED) trial. The BBAQ consists of 6 subscales (lack of time, social influence, lack of energy, lack of willpower, fear of injury, lack of skill, and lack of resources). We used multivariate regression analyses, correcting for sociodemographics. Results: Higher physical activity barriers were associated with greater depression scores (linear effect, estimate=0.75, 95% CI: 0.39 – 1.12, p<0.001). This effect appeared to taper off for the higher barrier scores (quadratic effect, estimate: -0.02, 95% CI: -0.03 – -0.01, p=0.002). Exploratory analyses indicated that score differences were most driven by the social influence (p=0.027) and lack of energy subscales (p=0.017).Conclusions: Higher depression scores were associated with higher physical activity barriers. Social influence and lack of energy were particularly important barriers. Thus, addressing these barriers may improve the efficacy of physical activity interventions in women with higher depressive symptoms. However, a randomized controlled trial is warranted. Trial Registration ClinicalTrials.gov#: NCTO1280812 registered January 21, 2011


Introduction
Despite the physical and mental health bene ts of regular physical activity, women in every age group are less likely to meet the recommended levels of physical activity than men(1) (2) (3). According to both self-report and accelerometry measures, moderate and vigorous physical activity (MVPA) levels were higher in men than women (1). This may be in part because women face greater barriers to barriers to exercise than men(4) (5). Common self-reported barriers in women include lack of time and motivation (3), the failure to see themselves as athletes (4), and health problems(6) (7).
Women also have a risk up to two-fold compared to men of developing depression (8). Evidence from Mendelian randomization studies suggests that physical inactivity and depression have a bidirectional relationship (9). Those diagnosed with depression are less likely to maintain an exercise program (10). In turn, increasing physical activity can reduce depressive symptoms (11), and prevent depressive episodes (12). Therefore, focusing on promoting physical activity in women can substantially bene t both women's mental (e.g. reduce depressive symptoms) and physical health (e.g. reduce the risk of diabetes, cardiovascular disease and cancer) (13).
In depression, a scoping review found that emotional constraints, which included "low mood", "lack of energy" and "being too tired" were among the most important barriers for physical activity (14). However, less is known about the relationship between physical activity barriers and depression in physically inactive women. This information would be highly bene cial, as barriers to physical activity may moderate the effect of physical activity interventions (15). Physical activity interventions have been designed both speci cally for women (16) and for depression (17). Though these interventions have shown medium to high effect sizes (standardized mean difference of 0.28 after 6 months(18)), they may suffer from reduced effectiveness if they don't tailor speci cally to individual barriers and mental health symptoms.
It's imperative to understand if women with higher depressive symptoms report higher barriers, and which unique barriers are of particular importance. This will help us to better tailor physical activity interventions to this important subgroup. This paper examines the relationship between self-report physical activity barriers and depressive symptoms in community dwelling physically inactive women who enrolled in a mobile physical activity intervention, the mobile phone based physical activity education (mPED).
The aims of this study were to: 1) Examine if depressive symptoms predict the total barriers to physical activity score, correcting for demographics, clinical variables and social support.
2) Assess if depressive symptoms predict the barrier subscale scores correcting for demographics, clinical variables and social support in an exploratory manner.

Study design and sample
In this cross-sectional analysis of 318 women in the mPED trial, we analyzed the sociodemographic, clinical, and self-reported questionnaire data collected at the screening/baseline study visit. Detailed descriptions of the study design and outcomes have been previously published (19) , (20) , (21) (20,22,23). In short, eligibility criteria were female sex, age from 25 to 65 years, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) of 18.5 to 43.0, physically inactive at work and/or during leisure time based on the Stanford Brief Activity Survey, intent to be physically active, access to a home telephone or mobile phone, ability to speak and read English, no medical conditions or physical problems that required special attention in an exercise program, no current participation in other lifestyle modi cation programs, and no mild cognitive impairment as determined by the Mini-Cog test.
During the screening/baseline visit, sociodemographics, medical and lifestyle history, the Center for Epidemiological Studies Depression Scale (CES-D) (24), body mass index (BMI) and the Barriers to Being Physically Active Quiz (BBAQ) were assessed by a trained research staff. All methods were conducted in accordance with the Declaration of Helsinki.

Depressive symptoms
The Center for Epidemiological Studies Depression Scale (CES-D)(24) was used to assess self-reported depressive symptoms. The CES-D is a valid and reliable instrument that is widely used to assess depressive symptoms in a research context. The CESD ranges from (0-60), with a cutoff score of 16 indicating risk for clinical depression.
Physical activity barriers scale: The Barriers to Being Physically Active Quiz developed by the Centers for Disease Control and Prevention (CDC)(25) is a 21-item measure assessing the following barriers to physical activity: 1) lack of time, 2) social in uence, 3) lack of energy, 4) lack of willpower, 5) fear of injury, 6) lack of skill, and 7) lack of resources (e.g., recreational facilities, exercise equipment). Each domain contains 3 items, with a total score range of 0 to 63. Respondents rate the degree of activity interference on a 4-point scale, ranging from 0 = "very unlikely" to 3 = "very likely."

Other measures
A research staff asked the participants to ll out the sociodemographic and medical questionnaires immediately after obtaining the written consent form. The emotional support question, "How many people can you count on to provide you with emotional support?" was developed by the research team. To calculate body mass index (BMI), weight was measured with a Tanita WB-110 digital electronic scale, and height was measured at baseline with a standard stadiometer.

Statistical analysis
We assessed the relationship between depressive symptoms and the total Barriers to Being Physically Active Quiz scale score correcting for demographics, clinical variables and social support. We additionally conducted a posthoc analysis to examine the relationship between depression and the BBAQ subscales.
We used descriptive statistics to summarize sample characteristics and linear and logistic regression to examine the association between the total barriers and subscale scores with depression scores. For subscale scores with a non-normal distribution we computed a binary variable based on the median scores. We included self-reported age, BMI, employment (paid work, yes or no), whether participants have children at home (yes/no), whether they have driven a car in the past week (yes/no) and marital status as covariates in the model, based on previous evidence (19) , (20) , (21) (26). We examined non-linear effects of depression by including the quadratic effect of depression in the models and used Likelihood Ratio (LR) tests to assess the need for including these non-linear effects. We removed in uential observations using Cook's d (see results; we also conduct a sensitivity test when the in uential observations are not removed) and checked further model assumptions by visual inspections of residual plots. Tables are presented in the supplementary analysis. Analyses were carried out in R studio V. 1.1.423.

Results
Baseline characteristics of the 318 participants and the mean or median scores of the total barriers and the subscales are shown in Table 1. For the overall sample, based on the mean and median scores, the greatest barriers were lack of time, lack of energy and lack of willpower. Fear of injury was the least common reported barrier. In the supplementary material (Supplementary Table 1), we also show differences in the total scale and subscales between groups with high depression (CES-D ≥ 16) for illustrative purposes (in analyses we used the continuous scores). Results of the regression model are shown in Table 2 and a plot of this relationship is shown in Fig. 1.
Other factors associated with a higher barriers score were age and employment (Table 3). To assess this relationship using more exible modeling, we also examined a spline regression model. Spline regression is a non-parametric technique that divides the datasets into intervals with different ts, which may yield better results depending on the nature of the data. Results for spline regression were relatively similar but with a slower taper off at higher depression scores (see supplementary material). Likelihood ratio tests showed that a model including a quadratic term was a better t than a spline regression model.

Discussion
We showed that higher depression scores were associated with higher physical activity barriers when adjusting for sociodemographic and clinical variables. There was a quadratic effect of depression on physical activity barriers, so that as CES-D scores increased, the positive relationship with physical activity barriers tapered off. These results indicate that depression level is a particularly important factor to consider when designing physical activity interventions for physically inactive women.
Our study shows that physical activity interventions for inactive women with depression diagnoses or symptoms may bene t from taking into account their greater perceived barriers to engaging in exercise. Many physical activity interventions still use a 'one size ts all' approach. For instance, a review paper illustrated that unique barriers to physical activity in mental health are usually not accounted for within behavior change theories used to design physical activity interventions (27). This may be one of the reasons that individuals with depression are less likely to engage in exercise programs, especially when they have a low baseline level of tness (10). Previous work shows that physical activity interventions are more effective when they are tailored to subgroups like women and ethnic minorities (28). These ndings suggest interventions may further increase their potential by tailoring depression levels as well.

Most important barriers for physical activity
For the sample as a whole, we identi ed that willpower, lack of time and lack of energy were the most signi cant barriers whereas injury and lack of skill were less often reported. This is consistent with previous research in physically inactive pregnant women and a community sample of Brazilian women, also using the same barriers scale (29,30).
The Barriers to Being Physically Active Quiz subscales that were most associated with depression in a posthoc analysis were the Social In uence and Lack of Energy subscales. Although we assessed these relationships in an exploratory analysis, these barriers may be the driving factors behind the differences in total subscale scores.
The Social In uence subscale included statements such as: "None of my family members or friends like to do anything active, so I don't have a chance to exercise." and "I'm embarrassed about how I will look when I exercise with others". A lack of social support has been suggested as a risk factor for physical inactivity before (31), and our results show that this may be even more important in women with high depression scores. A review on physical activity barriers in depression (not speci c to women) also identi ed other people's behaviors and a lack of social encouragement as important barriers, and social support as a facilitator (14). Other research found that having a family member who exercises, or who encourages exercise, motivates engaging in healthy behaviors (32). Our ndings also suggest that selfconsciousness in social exercise-related situations (e.g. appearance towards others when exercising) may be an important factor discouraging women with higher depressive symptoms from physical activity.
The Lack of Energy subscale includes statements like: "I'm too tired after work to catch up on exercise", and "I need the weekend to catch up on sleep". A lack of energy and increased fatigue are symptoms of depression. Low energy was also identi ed before as one of the most important barriers in men and women with depression (14). Of note, changes in physical activity may improve energy, and reduce the magnitude of perceived energy as a barrier (29).
Our posthoc results, though they need to be con rmed in future work, suggests the need for an emphasis on social in uence and boosting energy to increase the effectiveness of physical activity promotion. Physical activity interventions could, for instance, integrate social support from family or friends, utilize peer-support or use community-based structures. Furthermore, exercise interventions can build-in graded exercise, personalized to a women's individual tness levels to help slowly overcome feelings of fatigue (33). Future work should also quantify and integrate facilitators to exercise in women with high depressive symptoms. In mixed gender populations with depression, facilitators included having a reason for exercising, being able to identify the psychological bene ts of exercise, having positive social support and integrating cognitive behavior change strategies (34).
Physical activity interventions are increasingly delivered in digital formats and via smartphones, using apps, and text-messaging and conversational agents. There is a growing interest in adaptive interventions, which alter their content based on the day-to-day behavior of individuals (35). We argue that physical activity interventions should both adapt to individuals' daily changing circumstances, and tailor their content to overcoming barriers of user subgroups.
The quadratic effect of depression After a depression score (CES-D) of around 20, past the clinical cut-off for identifying individuals at risk for clinical depression (≥ 16), physical activity barriers no longer increased with higher scores. We can't make nite conclusions about why this effect tapers off but there are multiple possible explanations.
First, for participants with higher depression scores, physical activity may not be a priority and therefore they are less aware of their physical activity barriers. Another potential explanation is that the Barriers to Being Physically Active Quiz scale doesn't capture all barriers that are relevant when women reach more severe levels of depression. For instance, previous research (36) showed that in severe mental illness, low mood and stress are perceived as the most signi cant barriers for physical activity, followed by social support. In addition, in outpatients with depression, physical exertion was the most common reported PA barrier (37). The Barriers to Being Physically Active Quiz scale used in this study assess lack of energy, but it does not capture whether low mood, high stress or physical exhaustion prevent women from exercising.
We recommend these questions be included in future versions.
One caveat here is that our data were sparse for very high depression scores, making these estimates less precise. Further, when we retained in uential observations in the model the quadratic effect lost signi cance. Future work should assess differences in barriers between women with slightly elevated symptoms and clinically diagnosed depression.
Age, employment and physical activity Another interesting nding was the inverse relationship between age and perceived physical activity barriers. In our posthoc analyses higher age was associated with lower barrier subscale scores, except for the injury and skills scales. These ndings complement previous work in community samples showing that younger adults, both men and women, (25-44 years) report most physical activity barriers, and older adults (> 65 years) least (38). In line with our ndings, earlier work has also identi ed lack of time and energy due to family and household responsibilities among the top barriers to physical activity for women (39). Older women may be less impacted by these responsibilities.
Finally, full-time or part-time employment was associated with higher barriers opposed to unemployment.
Past work revealed that working women perceived lack of time and energy as most frequently reported barriers to physical activity (26). Future interventions may want to incorporate strategies for promoting physical activity into working hours to overcome these barriers for working women.

Strengths & Limitations
We included a relatively large sample of diverse women. To our knowledge, this is one of the rst studies to systematically examine the relationship between physical activity barriers and depressive symptoms in women using the instrument that was developed and validated by the CDC (14).
A limitation is that ndings may be speci c to our sample of female adults aged 25-65 years with relatively high levels of education from the San Francisco Bay Area. Our sample included fewer women with high depression scores, i.e. CESD > 20, therefore the estimate of the relationship may be less reliable after these higher scores. In addition, our analyses are cross-sectional, thus don't allow us to understand the causal relationship between perceived barriers and depression. This should be explored in future work. Another limitation may be selection bias. Since we included women who signed up to participate in a physical activity intervention, the study participants might be more motivated to engage in physical activity. Further, women with higher depressive symptoms are less likely to participate in the study. This may in part explain the tapering off effects we observe for high depression scores.

Conclusion
In community dwelling women who enrolled in the mPED trial, higher depression scores were associated with higher physical activity barriers when correcting known confounding factors. In an exploratory posthoc analysis we identi ed social in uence and lack of energy as particularly important physical activity barriers. Thus, addressing these barriers may improve an e cacy of physical activity interventions in women with high depressive symptoms. Further research, such as an RCT, is warranted.
Abbreviations BBAQ = the Barriers to Being Active Quiz, BMI = Mody Mass Index, CES-D = Center for Epidemiological Studies Depression Scale.

Declarations
Ethics approval: The study protocol was approved by the University of California, San Francisco (UCSF) Committee on Human Research and the mPED Data and Safety Monitoring Board (DSMB).
Consent to participate: All participants provided written consent prior to study enrollment.

Consent for publication: not applicable
Availability of data and materials: All data generated or analysed during this study are included in this published article [and its supplementary information les].
Competing of Interests: the authors report no competing interests.
Author contributions: YF designed and implemented the study and collected the data. CAF drafted the rst version of the manuscript and conducted the quantitative analysis. TJH supervised the statistical analysis. AA assisted with interpreting the results of the analyses. All authors contributed to the writing of the nal manuscript.
Funding: The project described was supported by the Award Number R01HL104147 from the National Heart, Lung, and Blood Institute, by the American Heart Association, and by a grant (K24NR015812) from the National Institute of Nursing Research