Body image in older women: A mediator of BMI and wellness behaviors

Abstract

Negative body image is prevalent among mid- and late-life women. In younger women, negative body image is associated with reduced quality of life (QOL) when controlling for body mass index (BMI), and mediates the relationship between obesity and emotional wellbeing. Yet, much remains unknown about body image in older populations. In our sample of women aged 50–86 (N=181), negative body image mediated the relationship between BMI and sleep, all four domains of QOL, negative affect, nutritious food consumption, and psychosocial impairment, but not enjoyment of physical activity. Findings suggest negative body image impacts the wellbeing of older women.

As women age, biological and lifestyle changes constitute risk for weight gain, which is particularly relevant during midlife (Karvonen-Gutierrez & Kim, 2016). For instance, reductions in estrogen during the menopausal transition leads to increase in total body fat mass, and body shape changes involve redistribution of adipose tissue to the waist/hips (Karvonen-Gutierrez & Kim, 2016). Furthermore, experiences common with aging such as medical conditions, chronic medication use, sleep disturbance, and mood problems all can contribute to weight gain as women age (Kapoor, Collazo-Clavell & Faubion, 2017).

Moreover, elevated body mass index (BMI), commonly described as overweight or obesity, is regularly decried as a major public health problem (e.g., Williams, Mesidor, Winters, Dubbert, & Wyatt, 2015; Wyatt, Winters, & Dubbert, 2006). When making the case for classifying elevated BMI as a public health concern, researchers, clinicians, and policy makers commonly point to the association between elevated BMI and medical conditions such as type 2 diabetes, hypertension, cardiovascular disease, hyperlipidemia, and sleep apnea (Hofmann, 2016; Williams et al., 2015; Wyatt et al., 2006). This is especially the case for obese postmenopausal women (Kapoor et al., 2017). Furthermore, many also note that elevated BMI is associated with reduced health related quality of life (QOL; WHO, 2003), mental health concerns, and reduced physical activity (Hofmann, 2016; Williams et al., 2015).

Increasingly, researchers are recognizing that elevated BMI is a highly stigmatized condition (Pomeranz, 2008; Puhl & Heuer, 2009; Williams et al., 2015) and that being the recipient of ongoing stigma may impact health related outcomes both directly and indirectly (Annis, Cash, & Hrabosky, 2004; Friedman et al., 2005; Mond & Baune, 2009). For instance, evidence suggests that many medical providers hold stereotyped notions and negative attitudes about elevated BMI (Phelan et al., 2015; Puhl & Brownell, 2006), which may in turn influence the care they provide to those with elevated BMI. As noted by Williams et al. (2015), the internalized experience of weight discrimination and stigma also may mediate the association of elevated BMI and health concerns/health-related QOL. Indeed, the authors explicitly call for additional research in this area.

One mechanism by which weight stigma may operate is negative body image. Body image refers to a complex, multi-dimensional construct comprised of thoughts, emotions, perceptions, and evaluations relative to one’s own body (Cash & Pruzinsky, 1990). Negative body image is typically operationalized as preoccupation with, negative evaluation of, and/or dissatisfaction with weight or shape. Extensive research, conducted primarily with adolescent girls and young adult women, supports an association between negative body image and both mental and physical health concerns such as depression (Paxton, Neumark-Sztainer, Hannan, & Eisenberg, 2006; Stice, Hayward, Cameron, Killen, & Taylor, 2000), eating disorders (Barker & Galambos, 2007), unhealthy weight control behaviors (Paxton et al., 2006), decreased physical activity (Jensen & Steele, 2008), smoking (Clark et al., 2005; Stice & Shaw, 2003), and worsened health-related QOL (Wilson, Latner, & Hayashi, 2013). Importantly, these relationships often remain when controlling for BMI (e.g., Neumark-Sztainer, Paxton, Hannan, Haines, & Story, 2006). Research in younger adult women also found that negative body image and has been associated with reduced QOL when controlling for BMI (Mond et al., 2013). Additionally, negative body image mediated the relationship between obesity and emotional wellbeing in female adolescents (Mond, Van den Berg, Boutelle, Hannan, & Neumark-Sztainer, 2011). Yet, because negative body image is commonly conceptualized as a concern of youth, older women have been largely neglected from research examining the consequences associated with this phenomenon (see review by Kilpela, Becker, Wesley, & Stewart, 2015).

The limited research investigating body image concerns and eating disorders among older women has largely been descriptive, though it demonstrates that common perceptions about these being problem of youth are, in fact, incorrect. More specifically, negative body image appears to be prevalent throughout mid- (Deeks & McCabe, 2001; Jackson et al., 2014) and late-life (Kilpela et al., 2015; Slevec & Tiggemann, 2011). For instance, Mangweth‐Matzek et al. (2006) found that over 60% of women aged 60–70 endorsed body dissatisfaction, and Becker, Verzijl, Kilpela, Wilfred, and Stewart (2017) found that 24% of women aged 61–86 endorsed marked body image concerns. In a sample of women aged 50 and over, 62% of women endorsed negative body image that had a negative impact on life at least occasionally (Gagne et al., 2012).

Of note, negative body image in older adult women may be more complex than negative body image in younger women and girls, secondary to the body-related changes that occur with aging (e.g., weight gain and adipose tissue redistribution during menopause; Kilpela et al., 2015). Given that negative body image is linked with poorer health and wellness behaviors in younger women, investigation of these relations in older women is critical. Although a small extant body of research supports an association between negative body image and health behaviors, health-related QOL and functional impairment in women throughout adulthood (Becker et al., 2017; Mond et al., 2013; Van Zutven, Mond, Latner, & Rodgers, 2015), virtually all of this limited literature has included both younger and older adult women in study samples.

Therefore, given the high prevalence of negative body image in older women and the link with poorer health and wellness behaviors observed in younger women, it is important to investigate potential health correlates of negative body image in older samples. As such, the aim of the current study is to evaluate negative body image as a potential mediator of the relationship between BMI and QOL, negative affect, and health/wellness behaviors in a sample of women aged 50 and over. We hypothesized that negative body image would mediate the relations between BMI and each outcome variable for women aged 50 and over.

METHOD

Participants

Participants included 200 women recruited via snowball sampling. Of the 200 women, 19 participants who did not report data to calculate BMI were excluded; 18 of the excluded women did not disclose height, and one stated, “don’t know” for weight. The mean BMI of the sample analyzed (N = 181) was 26.65 (SD = 5.74), with a range from 18.40 to 43.25. Participants’ ages ranged from 50 to 86 (M age = 58.26, SD = 6.48). Regarding race/ethnicity, participants had the option to select all that applied; 93.0% endorsed Caucasian, 7.0% Hispanic, 1.5% Native American/Alaska Native, 1.5% African American, 1.0% Asian, and 2.0% endorsed multiple races. The sample was highly educated, with 80.5% reporting at least a bachelor’s degree. Most of the sample was currently married (75.5%), while 5.0% were never married, 10.0% were divorced, and 4.5% were widowed.

Procedure

After obtaining Institutional Review Board approval, we recruited women to participate in a survey about women’s body image and health/wellness. We distributed recruitment scripts via personal and professional networks through email, social networking sites (e.g., Facebook), word of mouth, and social media sites dedicated to women’s fitness or wellness. All communications asked that women forward the survey invitation to their own networks. Upon survey completion, participants could provide email addresses to enter a raffle for a $200 Amazon gift card.

This study represents a secondary analysis of a larger evaluation of body image and wellness behaviors among women across the lifespan (Becker et al., 2017). Only women aged 50 and older were included in the current study; after providing informed consent, 200 women aged 50 and over completed the self-report questionnaires online.

Measures

Demographic information.

Participants provided self-reported age, race/ethnicity, as well as highest level of education.

Body mass index (BMI).

BMI was calculated using participants’ self-reported height and weight by using the equation BMI = Weight (kg)/Height (m)². Although research indicates a high agreement between self-reported and objectively measured BMI (Himes, Hannan, Wall, & Neumark-Sztainer, 2005), self-report is not considered the optimal method for obtaining weight and height data; however, it was the only option for this study.

Negative Body Image.

We used the Body Shape Questionnaire (BSQ; Cooper, Taylor, Cooper, & Fairbum, 1987), which uses a 6-point Likert scale (1 = Never, 6 = Always), to assess negative body image over the past 4 weeks. Research supports the reliability and validity of the BSQ (Rosen, Jones, Ramirez, & Waxman, 1996). A sample question is “have you been so worried about your shape that you have been feeling you ought to diet?” To reduce participant burden, we utilized an 8-item version (BSQ-8D: Evans & Dolan, 1993); shortened versions have similar convergent and discriminant validity to the original BSQ (Evans & Dolan, 1993). Scores are summed with higher scores indicating greater body image concerns. Two questions from BSQ-8D (e.g., have you imagined cutting off fleshy areas of your body) were replaced with items from the BSQ-8B (e.g., have you avoided wearing clothes which make you particularly aware of the shape of your body?) to focus on less pathological aspects of negative body image. Internal consistency for this modified BSQ sample was excellent (Cronbach’s α = .91).

QOL and Psychosocial Impairment.

We assessed QOL with the 26-item short version of the World Health Organization Quality of Life Scale (WHOQOL-BREF; Skevington, Lofty, & O’Connell, 2004), which yields four subscales covering the past month: physical health (e.g., activities of living), psychological health (e.g., self-esteem), social relationships (e.g., social support), and environment (e.g., financial resources); this measure uses a 5-point Likert scale. Subscales are summed for a raw subscale score; higher scores indicate better QOL. Research supports the internal consistency of the subscales (Skevington et al., 2004); internal consistency for each subscale in this sample was good (α: physical health = 0.83, psychological = 0.88, environment = 0.86, social relationships = 0.76).

We evaluated psychosocial impairment in four domains of life (mood/self-perception, cognitive functioning, interpersonal functioning, work performance) secondary to concerns with weight and shape over the past 28 days using the 16-item Clinical Impairment Assessment Questionnaire (CIAQ; Bohn et al., 2008). For example: “over the past 28 days, to what extent have your eating habits, exercising, or feelings about your eating, shape or weight stopped you from going out with others?”). Items are rated on a 4-point Likert scale (1 = not at all, 4 = a lot) and are summed; higher scores indicate greater impairment. We removed “eating” and “exercise” from the question stem to emphasize shape/weight concerns (i.e., “…to what extent have your feelings about your shape and weight). Past research (Bohn et al., 2008) supports CIAQ construct validity and internal consistency, and the modified version demonstrated high internal consistency (α = .96).

Wellness Behaviors.

We used the 8-item Physical Activity Enjoyment Scale (PACES; Mullen et al., 2011) to assess enjoyment associated with physical activity (e.g., “when I am physically active I find it pleasurable”). The PACES uses a 5-point scale (1 = Disagree a lot, 5 = Agree a lot), and scores are summed. Higher scores indicate greater physical activity positive affect. Internal consistency in this sample was excellent (α = .97). The Pittsburgh Sleep Quality Index (PSQI; Buysse, Reynolds III, Monk, Berman, & Kupfer, 1989) is a 19-item measure of self-reported sleep over the past month. Seven component sleep scores comprise the PSQI: quality, latency, duration, efficiency, disturbances, use of sleep medication, and daytime dysfunction. The seven scores are summed to indicate a global score, in which lower scores indicate better sleep. The PSQI has demonstrated internal consistency and reliability (α = 0.83; Smyth, 1999). We found good internal consistency (α = .67).

We used two items to assess self-reported consumption of fruits, vegetables, and nutrient-dense foods over the past week: 1) asked “how often did you consciously try to increase the nutrient density of your meal (choose nutrient rich foods instead of foods with empty calories)?” and 2) how often did you eat fresh fruits and vegetables?” Items rated on a 5-point Likert scale (1 = consume at every meal, 5 = never), were summed. Higher scores indicate fewer attempts to eat nutritionally dense foods. Internal consistency in this sample was adequate (α = .77).

Negative affect.

Seventeen items comprising the fear, guilt and sadness subscales of the Positive and Negative Affect Schedule Expanded (PANAS; Watson & Clark, 1999; Watson, Clark, & Tellegen, 1988) assessed negative affect. Items are rated on a 5-point Likert scale (1 = very slightly or not at all, 5 = extremely) over the past 3 weeks, with higher scores indicating greater negative affect. Research supports construct validity (Watson & Clark, 1999). Internal consistency was excellent in the current sample (α = .96).

Analytic Strategy

All analyses were conducted in SPSS (Version 24; IBM, 2016). Means, standard deviations, and correlations for all measures are presented in Table 1. Cross-sectional mediation models were tested using PROCESS (Model 4) for SPSS (Version 24), with BMI as the independent variable and negative body image (BSQ) score as the mediator for all models.¹ Outcome variables included the four domains of QOL (psychological, physical, social, and environmental), psychosocial impairment (CIAQ), enjoyment of physical activity (PACES), global sleep index (PSQI), consumption of nutritious foods (EBQ), and negative affect (PANAS-X). See Figure 1 for the conceptual model.

Table 1.

Means, standard deviations, and zero order correlations for primary variables

	M	SD	BMI	BSQ	QOL-Phys#	QOL-Psy#	QOL-Soc#	QOL-Env#	CIAQ	PACES#	PSQI	EBQ	PANAS
BMI	26.65	5.74	-
BSQ	26.88	9.14	.43**	-
QOL-Phys#	28.64	4.63	−.34**	−.43**	-
QOL-Psy#	23.23	4.48	−.34**	−.53**	.63**	-
QOL-Soc#	11.28	2.51	−.29**	−.42**	.53**	.69**	-
QOL-Env#	33.94	4.77	−.35**	−.33**	.54**	.60**	.61**	-
CIAQ	24.38	8.68	.39**	.63**	−.44**	−.65**	−.39**	−.34**	-
PACES#	32.87	6.40	−.31**	−.23**	.40**	.43**	.30**	.37**	−.33**	-
PSQI	8.39	3.21	.28**	.40**	−.59**	−.53**	−.39**	−.38**	.42**	−.18*	-
EBQ	3.98	1.73	.26**	.30**	−.26**	−.31**	−.17*	−.26**	.30**	−.26**	.13	-
PANAS	28.12	12.96	.23**	.49**	−.50**	−.73**	−.58**	−.42**	.68**	−.23**	.47**	.20**	-

Note. M = mean score; SD = standard deviation; BMI = Body Mass Index; BSQ = Body Shape Questionnaire; QOL-Phys = Quality of Life Physical Health; QOL-Psy = Quality of Life Psychological; QOL-Soc = Quality of Life Social Relationships; QOL-Env = Quality of Life Environment; CIAQ = Clinical Impairment Assessment Questionnaire; PACES = Physical Activity Enjoyment Scale; PSQI = Pittsburgh Sleep Quality Index; EBQ = nutrient density consumption; PANAS = Negative Affect;

^#= higher scores indicate healthier/less problematic (for all other measures, higher scores indicate more problematic); All significance tests were two-tailed (*p < .05; **p < .01).

Figure 1. Conceptual model of negative body image as a mediator.

Note: QOL = Quality of Life; PA = Physical Activity; c’ = indirect effect.

RESULTS

Analyses revealed a significant association between BMI and negative body image, such that higher BMI was associated with higher negative body image. Additionally, both BMI and negative body image were associated with all outcome variables (see Table 1). Specifically, both higher BMI and higher negative body image were each associated with lower QOL (all four domains), greater psychosocial impairment, less enjoyment of physical activity, poorer sleep, less frequent consumption of nutritious foods, and greater negative affect.

Mediation models revealed a significant indirect effect of BMI through negative body image for: negative affect (b = .43, Boot SE = .09, z = 4.40; 95% CI [.27, .64]); sleep (b = .09, Boot SE = .02, z = 3.52; 95% CI [.04, .14]); consumption of nutritious foods (b = .03, Boot SE = .01, z = 2.66; 95% CI: [.01, .05]), and psychosocial impairment (b = .35, Boot SE = .07, z = 5.09; 95% CI [.25, .52]). Analyses also indicated significant indirect effects between BMI and all domains of QOL through negative body image: physical QOL (b = −.48, Boot SE = .13, z = −3.79; 95% CI [−.79, −.26]), psychological QOL (b = −.66, Boot SE = .14, z = −4.59; 95% CI [−.97, −.40]), social QOL (b = −.62, Boot SE = .16, z = −4.05; 95% CI [−.99, −.36]), and environmental QOL (b = −.28, Boot SE = .11, z = −2.83; 95% CI [−.48, −.09]). The indirect effect between BMI and enjoyment of physical activity (PACES) indicated that this mediation model was not significant (b = −.05, Boot SE = .04, z = −1.16; 95% CI [ −.14, .03]).

DISCUSSION

The current study sought to extend the existing literature on body image and health in older women by evaluating the relations between BMI, negative body image, and various health and quality of life outcomes (QOL) in women aged 50 and over. In our sample, BMI was positively associated with negative body image, such that higher BMI was related to greater body shape dissatisfaction. BMI was also related to all health/wellness and QOL outcome variables in the predicted direction; higher BMI was associated with poorer health/wellness and QOL indices. Finally, negative body image was related to all outcomes, such that greater negative body image was associated with poorer outcomes. Contrary to our hypothesis, the mediation model for the relation between BMI and enjoyment of physical activity was nonsignificant. Consistent with our hypotheses, mediation models were significant for the relations between BMI and negative mood, sleep, consumption of nutritious foods, psychosocial impairment, and all domains of QOL (psychological QOL, social QOL, physical QOL, and environmental QOL). Therefore, our results indicate that negative body image plays a significant role in older women’s wellness and different facets of QOL.

These findings are important in understanding the various factors that influence health and wellness behaviors, QOL, and psychosocial functioning in older women. Specifically, with the recent public health emphasis on obesity and weight status, these results point to an individual’s cognitive and affective perception of her own body as influential in health and wellness behaviors (e.g., sleep, consumption of nutritious foods, physical QOL), as well as psychological health (e.g., negative affect, psychosocial impairment, and psychological and social QOL). For instance, negative body image mediated the relations between BMI and negative mood and sleep quality, both of which are associated with numerous health and wellness consequences among older populations (Black, O’reilly, Olmstead, Breen, & Irwin, 2015; Haigh, Bogucki, Sigmon, & Blazer, 2018). In younger samples, negative mood and poor sleep are associated with weight gain over time (e.g., Owens & Group, 2014). Moreover, negative body image was found to be a significant shared risk factor for depressive symptoms, disordered eating, and overweight prospectively in younger women (Goldschmidt, Wall, Choo, Becker, & Neumark-Sztainer, 2016). Therefore, a better understanding of the role of negative body image in promoting and/or maintaining negative mood among older women has the potential to influence psychological treatment targets for geriatric depression or weight management.

Furthermore, our results are consistent with research in younger female samples suggesting that negative body image plays a notable role in the relation between BMI and various health-related QOL and wellness outcomes. For instance, in younger adult women and female adolescents, negative body image has been linked with reduced QOL even when controlling for BMI (Mond et al., 2013). Mond and colleagues (2011) also found that negative body image mediated the relationship between obesity and emotional wellbeing among female adolescents over time. Thus, although there are likely various psychological, biological, physical, and social factors that operate differently in older women as compared to younger in terms of the manifestation and correlates of negative body image, our results suggest that negative body image: a) similarly affects QOL, health behaviors, and emotional, psychological, and physical wellbeing in older women, and b) may be a viable target for interventions designed to improve older women’s health and wellness, beyond BMI.

Our findings also highlight the importance of assessing negative body image in older women in clinical settings, given its role in health behaviors and psychological wellbeing. Although historically considered to be a concern of younger women, the mediational role of negative body image in older women suggests that assessment of negative body image in older women is important in the consideration of health behaviors. Results from the current study also point to need to expand research and consideration of harmful outcomes and behaviors related to negative body image, beyond eating disorders and weight status, across the lifespan.

Notably, psychological interventions targeting negative body image, such as dissonance-based interventions, have demonstrated strong results (e.g., Stice, Becker, & Yokum, 2013) and indicate that negative body image is a malleable construct. Furthermore, body shape and appearance changes associated with the menopausal transition (e.g., redistribution of adipose tissue to the trunk, reduction in lean muscle mass, reduced skin firmness) conflict with Western societal beauty ideals of thinness and perpetual youth (Kilpela et al., 2015). Such physical changes may comprise a period of increased risk for new onset or resurfacing of negative body image for women in midlife (Mangweth-Matzek et al., 2013). Thus, interventions incorporating strategies to improve body image would likely offer a benefit to older women in improving health behaviors that are important for healthy aging and longevity.

Finally, our findings highlight the need for more research investigating the psychological construct of body image, as well as its correlates and consequences, in older populations in order to better understand the interplay of aging and body image. Consistent with work encouraging the use of a life course framework (Kuh & Hardy, 2002; McLaren & Wardle, 2002), the current study benefited from novel contributions by extending the existing literature on the relations between BMI, negative body image, and various health and quality of life outcomes in a large sample of women aged 50 and over. More research is needed, however, to investigate aging-related physical, social, environmental, and emotional experiences and the effect on self-perception of one’s own body. For instance, older individuals are more likely to have chronic illness, medication use, or physical limitations than are younger populations (Kilpela et al., 2015). Therefore, negative body image may be relevant to older women not only in terms of body appearance, but also functional limitations of the body associated with the aging process. Of note, body image theory and research has traditionally emphasized the appearance aspect of negative body image over functional aspects (e.g., dissatisfaction with how body function has changed over time) (Tiggemann, 2004); however, it is important to incorporate aging and associated factors into the evaluation and conceptualization of negative body image in women across the lifespan.

Despite the strengths of the study, a number of limitations deserve consideration. First, we used exclusively self-report data that were completed online and not in a controlled environment. Second, our sampling method resulted in minimal racial and ethnic diversity, and we did not collect demographic information on country of residence, thus limiting our ability to generalize findings to women of diverse racial and ethnic backgrounds or draw conclusions regarding cultural differences. Additionally, our models were cross-sectional; longitudinal research to investigate these phenomena over time among older women is warranted. Future research would also benefit from including more diverse samples, as well as more comprehensive measures for evaluation (e.g., clinical interview, medical/biological data).

Overall, results indicate that body dissatisfaction plays a notable role in the relationship between BMI and QOL outcomes in older women, and this relation appears to be stronger for psychological and social domains of QOL, as well as negative mood and sleep. Findings suggest the need for additional research investigating the complex issue of body image and its consequences among older women, as well as the potential need for interventions tailored for this population.