Associations Between Body Mass Index, Body Image Satisfaction, and Self‐Weighing During a Behavioral Weight Loss Program

ABSTRACT

Background

Greater adherence to daily self‐weighing has been associated with improved outcomes within behavioral weight‐loss programs. Identifying factors that can predict adherence to self‐weighing may support the development of novel tailored interventions.

Methods

The current study examined whether baseline body mass index (BMI) and body image satisfaction (assessed via the Body Image States Scale [BISS]) predicted adherence to self‐weighing during a 16‐week behavioral weight‐loss program in 449 adults with obesity (mean ± SD age = 49.47 ± 11.37 years; BMI = 35.73 ± 4.03 kg/m²; 83.52% female, 74.16% White). Participants were provided with e‐scales and encouraged to self‐weigh daily during the program. Linear regressions were used to examine associations between BMI and BISS scores (and their interaction) and the proportion of program days that participants self‐weighed.

Results

At baseline, average (mean ± SD) BISS scores were 3.57 ± 1.28; higher BMI was associated with lower BISS scores, r = −0.11. Participants self‐weighed an average of 80.92 ± 23.35% of days during the intervention; greater self‐weighing was associated with greater weight loss, r = −0.56. Neither BMI nor BISS at baseline predicted self‐weighing adherence; moreover, there was not an interaction between BMI and BISS scores in relation to self‐weighing adherence.

Conclusions

Although results demonstrating negative associations between BMI and body image satisfaction and between self‐weighing adherence and weight loss were consistent with prior literature, the lack of an association between BMI, BISS, and self‐weighing was not consistent with hypotheses or the prior cross‐sectional literature. Given the role self‐weighing plays in behavioral weight‐loss programs, future research should aim to identify other predictors of self‐weighing adherence.

1. Introduction

Regular self‐monitoring of weight (i.e., “self‐weighing”) is a standard component in comprehensive behavioral weight‐loss programs [1]. According to self‐regulation theory, self‐weighing gives feedback regarding how an individuals' changes in their eating and activity habits affect their weight, providing reinforcement for goals that are met and guiding future goal setting [2]. In line with this theory, higher adherence to self‐weighing goals has been consistently associated with greater success at weight loss over the course of weight‐loss interventions [3, 4, 5, 6, 7]. More frequent self‐weighing has also been associated with weight loss on a proximal (i.e., within‐week) level, such that a greater number of days of self‐weighing within a given week has been associated with greater weight loss during that same week [8].

Given this key role, researchers should identify which factors predict adherence to self‐weighing to serve as key targets for future intervention tailoring. The process of self‐weighing may be aversive or cause discomfort, especially for individuals dissatisfied with their current weight [9]. Thus, a plausible hypothesis is that pre‐intervention BMI and body image satisfaction may serve as important predictors of self‐weighing during a weight management intervention. Indeed, cross‐sectional associations have been observed between body image satisfaction and self‐weighing, such that people with higher body image satisfaction engage in self‐weighing more frequently [10]. The directionality of this association, however, remains unknown. Cross‐sectional associations have also been observed between BMI and self‐weighing such that, among women with overweight, those with lower BMIs reported engaging in self‐weighing more frequently [11]; moreover, higher BMI was associated with less self‐reported self‐weighing at baseline in adults with overweight or obesity enrolling in a weight‐loss program [12]. Although no studies have investigated whether baseline BMI predicts self‐weighing during a comprehensive weight‐loss program, a study by VanWormer and colleagues [13] found no association between BMI at baseline and self‐weighing during a 2‐year worksite weight‐gain prevention program and, conversely, a study by Brindal and colleagues [14] found that higher BMI was associated with greater use of an online weight‐tracking tool during a 12‐week Internet‐based weight loss program that included no professional contact.

The current study aimed to build upon this previous literature by investigating whether baseline body image satisfaction and baseline BMI predicted adherence to daily self‐weighing in adults with obesity who enrolled in a 16‐week comprehensive behavioral weight‐loss program. The primary study hypothesis was that higher baseline BMI and lower body image satisfaction would predict lower self‐weighing adherence during the intervention. Given prior literature demonstrating an inverse association between weight status and body image satisfaction [15], a secondary hypothesis was that a higher baseline BMI would be associated with lower body image satisfaction at baseline. Finally, exploratory analyses examined: (1) whether body image moderated the association between baseline BMI and adherence to daily self‐weighing during the intervention, such that the association between BMI and self‐weighing was stronger in participants with lower body image satisfaction, and (2) associations between baseline BMI, baseline body image satisfaction, self‐weighing adherence, and weight loss during the 16‐week intervention.

2. Methods

The current study conducted secondary analysis of data from the Project STAR trial (ClinicalTrials.gov NCT04116853). This parent study was a randomized clinical trial examining the impact of an adaptive intervention on weight regain following initial weight loss [16]. In this trial, adults with obesity were provided with a 16‐week behavioral weight‐loss program; individuals who lost at least 5% from their baseline weight at the end of this initial program were then randomized into the trial. The current study used data from the initial weight‐loss program prior to trial randomization.

2.1. Participants

A detailed account of the recruitment protocol and participant eligibility criteria for Project STAR has been published previously [17]. Briefly, recruitment strategies included flyers, community outreach, and e‐mails sent to individuals from a registry of UF Health patients who previously provided consent to be contacted for research. To be eligible for the weight‐loss program, potential participants needed to be adults (age 18–70 years) with obesity (BMI of 30.0–45.0 kg/m²) but with weight < 180 kg due to limitations of study‐provided e‐scales. Potential participants also had to report owning a smartphone with a cellular data plan and having no medical conditions that contraindicated weight‐loss program participation. Four‐hundred and forty‐nine participants enrolled in the initial weight‐loss program; at baseline, average age (mean ± SD) was 49.47 ± 11.37 years and BMI was 35.73 ± 4.03 kg/m² [16]. In terms of gender, 83.5% of the sample were female. Regarding race/ethnicity, 74.2% identified as white, 23.4% as Black or African American, 9.8% as Hispanic/Latino, 3.1% as Asian, 3.2% reported other race categories (more than one race/ethnicity could be selected by participants).

2.2. Intervention

Participants were enrolled in a 16‐week comprehensive weight‐loss intervention adapted from the Diabetes Prevention Program [18]. A detailed description of this intervention (including a list of sessions and their topics/content) has been published [17]. In brief, participants were asked to attend weekly group sessions including 9‐18 participants and two trained interventionists. Participants were provided with self‐monitoring tools (a smartphone app/website to self‐monitor dietary intake and physical activity along with an e‐scale to self‐monitor weight) and asked to self‐monitor dietary intake, physical activity, and weight each day. Daily self‐weighing was introduced as a way for individuals to track their trends in weight change over time; participants were encouraged to view weights as data versus as a judgment of self‐worth, and to look for overall patterns in weight trajectory (vs. being caught up in day‐to‐day variability in weight). Relevant to the current study, one session in the second half of the program focused specifically on body image, including discussions on how a negative body image can affect an individual and their health behaviors and the importance of body acceptance coupled with an activity aimed at improving body image. Participants experienced an average (mean ± SD) of 6.4 ± 5.0 kg weight loss during the 16‐week program, representing a 6.4 ± 4.9% decrease from baseline weight [16].

2.3. Measures

Assessments were conducted at baseline and after completion of the weight‐loss program (Month 4). Participants self‐reported their height during screening. Weight was obtained using study‐provided Bodytrace e‐scales which used a cellular connection to send data to research servers. Weights assessed using these e‐scales have shown high concordance with those measured in‐person [19, 20]. Following a standardized protocol [21], participants were asked to self‐weigh first thing each morning after voiding and before eating or drinking, wearing light indoor clothing with emptied pockets, and shoes removed. For each assessment, participants were asked to step on and off the scale three times, with the two closest weights averaged to calculate the assessment weight. Adherence to daily self‐weighing was also assessed using e‐scale data and operationalized as the proportion of days during the 16‐week intervention that each participant self‐weighed at least once using the study‐provided e‐scale (after removal of outliers, e.g. weights indicating other scale users). A proportion was used rather than a raw frequency count because of differences in intervention session scheduling across cohorts. Participants in three of the five cohorts of the study had 111 possible days of self‐weighing (representing 16 weeks, with the first week only having 6 days as participants were encouraged to start self‐monitoring on the day after their first intervention session); participants in the two remaining cohorts had a total of 118 possible days (17 weeks, again with the first week limited to 6 days) due to a one‐week winter‐holiday closure of the institution during the intervention period. During this extra week between scheduled sessions (occurring between sessions 12 and 13 in one cohort and between sessions 11 and 12 in the other), participants were encouraged to continue self‐monitoring each day, but no intervention was provided. Finally, body image satisfaction was evaluated using the 6‐item Body Image States Scale (BISS) [22]. The BISS is a 6‐item self‐report measure which asks participants to rate their satisfaction with their physical appearance on a 1‐9 point scale, with higher ratings indicating higher body image satisfaction [22]; a final score was developed by averaging ratings across the 6 items.

The University of Florida Institutional Review Board approved all study procedures; data and safety monitoring was overseen by a trial Safety Officer.

2.4. Statistical Analyses

R (version 4.4.1) was used to conduct analyses. Pearson correlations were used to investigate the associations between BMI and BISS scores at baseline and between self‐weighing adherence and percent weight loss. Linear regressions were used to examine the associations between baseline BMI and baseline BISS scores and self‐weighing adherence during the 16‐week intervention, with an interaction term used to investigate whether the association between BMI and adherence to self‐weighing was moderated by baseline BISS scores. The same approach was used to investigate associations between baseline BMI and BISS and their interaction and percent weight loss during the intervention.

3. Results

Average (mean ± SD) baseline BMI was 35.73 ± 4.03 kg/m², and average baseline BISS scores were 3.57 ± 1.28; higher BMI was associated with lower BISS scores, r = −0.11, p = 0.019. Participants self‐weighed an average of 80.92 ± 23.35% of possible days during the intervention. Neither baseline BMI nor baseline BISS scores predicted self‐weighing adherence during the intervention (ps = 0.758 and 0.570, respectively). Moreover, BISS scores did not moderate the association between baseline BMI and self‐weighing during the intervention, p = 0.501.

Greater self‐weighing adherence was associated with greater percent weight loss during the intervention, r = −0.56, p < 0.0001. Higher BMI at baseline predicted less percent weight loss during the intervention, B = 0.12, SE = 0.06, F(1,447) = 4.58, p = 0.033, R ² = 0.01. Moreover, higher BISS at baseline predicted less weight loss during the intervention, B = 0.06, SE = 0.03, F(1,447) = 4.15, p = 0.042, R ² = 0.01. BISS did not, however, moderate the association between baseline BMI and percent weight change during the intervention, p = 0.775.

4. Discussion

The current study explored associations between baseline BMI, baseline body image satisfaction, and adherence to self‐weighing in adults with obesity enrolled in a behavioral weight‐loss program. Overall, adherence to self‐weighing was high during the program, with participants self‐weighing an average of 81% of possible days (compared to rates of 43%–75% reported in a recent review of digital health self‐monitoring tools [5]). Consistent with prior literature, greater self‐weighing adherence was associated with greater success at weight loss (explaining 32% of the variance in weight change observed during the intervention) and higher BMI was associated with lower ratings of body‐image satisfaction at baseline [3, 4, 5, 6, 7, 15]. In contrast to hypotheses and the study by Brindal and colleagues [14], but consistent with results from VanWormer and colleagues [13], neither BMI nor body image satisfaction predicted adherence to self‐weighing during the intervention; moreover, body‐image satisfaction did not significantly moderate the association between BMI and self‐weighing.

Interestingly, even though there was no association between baseline BMI, BISS, and self‐weighing adherence, higher BMI at baseline predicted less weight loss; previous studies found that higher BMI predicts greater total weight loss during behavioral interventions, but that there is less consistency in this pattern when percent weight change is used as an outcome [23]. Moreover, greater BISS also predicted less weight loss during the 16‐week program; prior studies have demonstrated inconsistency regarding whether BISS predicts weight‐loss program outcomes [23]. It is important to note, however, that effect sizes for each of these associations were small (R ² = 0.1 for each, suggesting that these variables explained only 1% of the variance in percent weight change during the intervention); thus, caution should be taken when interpreting these results.

Taken together, the results suggest that baseline BMI and BISS scores may have little utility as targets for the development of novel tailored interventions to support self‐weighing adherence. Explanations for these results include that there are not associations between BMI, BISS, and self‐weighing adherence, or that existing comprehensive weight‐loss program components may buffer any potential impact of BMI or body image satisfaction on self‐weighing adherence. For example, when self‐weighing was introduced in the current program, emphasis was placed on not viewing weight as a judgment of self‐worth; instead, participants were encouraged to try to objectively view self‐weighing results as “data” providing feedback on the effectiveness of their changes in dietary intake and physical activity. Moreover, content in a later intervention session specifically focused on building a positive body image. As a final possible explanation, support from interventionists and other group members may have promoted self‐weighing adherence.

Unfortunately, as the current study did not have a control group which did not receive behavioral intervention, these possibilities cannot be directly investigated in the current study. This study also examined only short‐term outcomes (over the 16‐week intervention); there may be different patterns of association observed in the longer‐term, for example in relation to weight loss maintenance. Results also may be limited due to use of the BISS to assess body image satisfaction; this measure was used due to its brevity (only 6 items vs. the 34–69 items on other measures of body image satisfaction [24, 25, 26]) as it was included as part of a larger questionnaire battery for the parent weight‐loss maintenance trial. Longer, more detailed measures of body image satisfaction may lead to greater measurement precision and improve rigor for assessing associations between body image satisfaction, self‐weighing, and weight loss.

Generalizability of the results may also be limited given parent study demographics; participants were predominantly female and non‐Hispanic White, and over half of the sample reported having at least a college degree and household incomes ≥ $75,000/year. Thus, the results may not generalize to men or individuals who have lower education or lower household income. Furthermore, adults who found self‐weighing aversive may also have been less likely to enroll in our comprehensive weight‐loss program, potentially limiting the range of variability in self‐weighing adherence (and/or contributing to the high adherence to self‐weighing observed in this study). Future studies should investigate whether results are replicated in broader populations, using samples with a greater proportion of men and individuals from racial/ethnic backgrounds historically‐underrepresented in research, as body image and beauty standards may vary by gender and across cultures [27, 28]. Strengths of the current study included the use of data from 449 adults enrolled in a gold‐standard comprehensive behavioral weight‐loss program, and the use of objective e‐scale data to assess self‐monitoring adherence.

This study was the first to investigate whether baseline BMI or body image satisfaction predicted adherence to self‐weighing during a behavioral weight‐loss program. Taken together, the results affirmed the role of self‐weighing in promoting weight loss but suggested that neither BMI nor body image satisfaction may be useful tailoring variables for future intervention development. Future studies should identify other key predictors of self‐weighing adherence to support the development of novel intervention strategies.

Conflicts of Interest

The authors declare no conflicts of interest.