Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 10.
Published in final edited form as: Appetite. 2014 May 21;80:204–211. doi: 10.1016/j.appet.2014.05.017

Effect of glycemic load on eating behavior self-efficacy during weight loss

J Philip Karl a,#, Rachel A Cheatham a,#, Sai Krupa Das a, Raymond R Hyatt a, Cheryl H Gilhooly a, Anastassios G Pittas b, Harris R Lieberman c, Debra Lerner d, Susan B Roberts a, Edward Saltzman a,*
PMCID: PMC11891877  NIHMSID: NIHMS603257  PMID: 24859114

Abstract

High eating behavior self-efficacy may contribute to successful weight loss. Diet interventions that maximize eating behavior self-efficacy may therefore improve weight loss outcomes. However, data on the effect of diet composition on eating behavior self-efficacy are sparse. To determine the effects of dietary glycemic load (GL) on eating behavior self-efficacy during weight loss, body weight and eating behavior self-efficacy were measured every six months in overweight adults participating in a 12-mo randomized trial testing energy-restricted diets differing in GL. All food was provided during the first six months and self-selected thereafter. Total mean weight loss did not differ between groups, and GL-level had no significant effect on eating behavior self-efficacy. In the combined cohort, individuals losing the most weight reported improvements in eating behavior self-efficacy, whereas those achieving less weight loss reported decrements in eating behavior self-efficacy. Decrements in eating behavior self-efficacy were associated with subsequent weight regain when diets were self-selected. While GL does not appear to influence eating behavior self-efficacy, lesser amounts of weight loss on provided-food energy restricted diets may deter successful maintenance of weight loss by attenuating improvements in eating behavior self-efficacy.

Keywords: energy restriction, glycemic index, weight maintenance, weight regain, weight self-efficacy

Introduction

Two-thirds of American adults report wanting to lose weight, and more than half of overweight and obese American adults report making behavioral changes in an attempt to lose weight (Yaemsiri, Slining, & Agarwal, 2011), yet recidivism after weight loss is generally high. Though a number of behavioral factors have been associated with successful long-term weight loss maintenance (Wing & Phelan, 2005), the factors that facilitate developing and sustaining these behaviors is less clear (Stubbs et al., 2011). Identifying these factors and the variables that influence them is crucial to developing practical, individualized approaches to healthy weight management.

Self-efficacy is postulated to be an important psychosocial factor contributing to body weight management (Elfhag & Rossner, 2005; Teixeira, Going, Sardinha, & Lohman, 2005; U.S. Department of Health and Human Services et al., 2000), and is defined as a personal belief in one’s ability to successfully execute behaviors required to produce a desired outcome (Bandura, 1977). Self-efficacy motivates behavior (Baranowski, Cullen, Nicklas, Thompson, & Baranowski, 2003), subsequently influencing health outcomes by determining the activities in which one participates, and the amount of effort one is willing to expend and sustain to successfully achieve a desired outcome (Bandura, 1977, 2004). Mastering behaviors that are required to achieve the desired outcome and successfully achieving that outcome are primary contributors to heightened self-efficacy (Bandura, 1977). Eating behavior self-efficacy can be conceptualized as one’s belief in their ability to adopt eating behaviors necessary to achieve and maintain a desired body weight. Eating behavior self-efficacy is thought to be a reliable predictor of weight loss (U.S. Department of Health and Human Services et al., 2000) and often improves during weight loss (Bas & Donmez, 2009; Clark, Abrams, Niaura, Eaton, & Rossi, 1991; Clark, Cargill, Medeiros, & Pera, 1996; Karlsson et al., 1994; Palmeira et al., 2007; Pinto, Clark, Cruess, Szymanski, & Pera, 1999; Presnell, Pells, Stout, & Musante, 2008; Teixeira et al., 2010; Warziski, Sereika, Styn, Music, & Burke, 2008). Therefore, diet interventions that maximize eating behavior self-efficacy could conceivably improve weight loss outcomes.

The effect of diet composition on eating behavior self-efficacy is undetermined. However, it can be hypothesized that the easier a dietary intervention is for individuals to adopt (i.e., mastering behaviors), and the greater the magnitude of weight loss reached (i.e., achieving the desired outcome), the larger the expected increase in eating behavior self-efficacy. Hunger is an often-cited reason for abandoning weight control programs (McGuire, Wing, Klem, Lang, & Hill, 1999; Pasman, Saris, & Westerterp-Plantenga, 1999), and studies have suggested that minimizing hunger may be important for successfully adopting and maintaining a dietary weight loss intervention (Batra et al., 2013; Cuntz, Leibbrand, Ehrig, Shaw, & Fichter, 2001). Beneficial effects of low-glycemic load (GL) diets on weight loss (Livesey, Taylor, Hulshof, & Howlett, 2008; Thomas, Elliott, & Baur, 2007) are thought to be attributable in part to greater hunger suppression on low-GL relative to high-GL energy-restricted diets (Ludwig, 2002; Roberts, 2000). Therefore, low-GL energy-restricted diets may improve eating behavior self-efficacy to a greater extent than high-GL energy restricted diets.

This hypothesis was explored in a secondary analysis of data from the Tufts University cohort of the multicenter Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (CALERIE) trial- phase I. The study was a year-long randomized controlled trial testing the effects of GL on adherence to energy restriction (Das et al., 2007). In addition to providing data related to GL, the study provided an opportunity to assess eating behavior self-efficacy under two different conditions: an initial 6-mo phase during which all food was provided and a second 6-mo phase that allowed participants to manage their diets independently. This design provided a unique, highly controlled environment for examining interrelationships between GL, eating behavior self-efficacy, and weight loss. The primary aim of the present analysis was to determine the effect of GL on eating behavior self-efficacy during weight loss.

Self-efficacy theory postulates that enhancements in specific dimensions of self-efficacy resulting from treatment strategies targeting that specific dimension can impact separate but related dimensions of self-efficacy (Bandura, 1977). Thus, enhancing eating behavior self-efficacy may promote a generalized increase in self-efficacy that influences self-efficacy for other behaviors beneficial to weight loss such as increased physical activity (Annesi & Whitaker, 2010; Wingo et al., 2013). As such, a secondary aim was to test whether changes in eating behavior self-efficacy are associated with changes in physical activity. We hypothesized that a provided low-GL diet compared to a provided high-GL diet would enhance improvements in eating behavior self-efficacy during weight loss, and that improvements in eating behavior self-efficacy would be positively associated with changes in physical activity.

Material and methods

Study population

The study cohort consisted of 46 overweight (BMI 25.0–29.9 kg/m2) but otherwise healthy adults age 20–42 y participating in the Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy trial- phase I at Tufts University. The full aims and methods of the Tufts University-arm of the trial have been previously described (Das et al., 2007). Participants were free of major diseases and were on no medications that could have influenced outcomes. Additional exclusions included very high physical activity levels (> 12 hr/wk), weight fluctuation of > 6.8 kg during the prior year, and inability to complete an accurate 7-day dietary record (accuracy defined as a reported intake of 70–130% of predicted energy needs for weight maintenance). The study was conducted at the Metabolic Research Unit of the Jean Mayer U.S. Department of Agriculture Human Nutrition Research Center on Aging at Tufts University with approval by the Tufts Medical Center and Tufts University Health Sciences Institutional Review Board, and in accordance with the Declaration of Helsinki. All participants gave written, informed consent prior to participating, and were provided with a stipend.

Study design

The study consisted of a 7-wk pre-intervention period followed by a 12-mo intervention involving random allocation to one of four diet groups (Table 1). During the pre-intervention period, participants were instructed to maintain body weight by following their habitual diet and activity patterns. Usual energy requirements were determined during this time by the doubly labeled water method (Roberts, 1989; Schoeller, 1988). After the 7 wk period, participants were randomized using a block randomization scheme stratified by BMI and sex to receive one of four diets differing by level of energy restriction (10% or 30% less than weight maintenance energy needs), and dietary GL (high-GL: 116 g/ 1000 kcal or low-GL: 45 g/ 1000 kcal). As such, four separate diet groups were created: 10%+low-GL (n=6), 10%+high-GL (n=6), 30%+low-GL (n=17), and 30%+high-GL (n=17). During the 6 mo following the pre-intervention phase, all food was provided to participants. During the subsequent 6 mo participants consumed a self-selected diet, but were instructed to maintain their respective energy restriction-level and diet composition as randomized. During the self-selected diet phase no food was provided; participants assumed the responsibilities of planning, purchasing and preparing their meals. Physical activity level was not prescribed at any time. Participants visited the facility at least weekly throughout the study for a variety of activities including behavioral support groups, individual meetings with study dietitians, food pick-up, safety monitoring, and outcome testing. All outcome measures described below were conducted at the end of the pre-intervention period and every 6 mo thereafter.

Table 1.

Study design

Time period Diet Intervention componentsa Behavior change techniques employedb
Pre-intervention:
−7 to 0 weeks		 Habitual diet Outcome assessments Randomization
Phase 1:
0 to 6 months		 All food prepared and provided; 10% or 30% energy restriction + low- or high-glycemic load Daily to weekly:
 Recording of body weight, food intake, appetite, activity, behaviors, and emotions
 Visit metabolic unit to pick up food
Biweekly:
 Behavioral support groups
 Individual sessions with dietitian
Pre and post phase:
 Outcome assessments		 Personal-level:
 Repetition and substitutionc, associationsd, natural consequencese, feedback and monitoringf, goals and planningg, identityh
Group sessions:
 Repetition and substitutionc, goals and planningg, identityh, reward and threati, antecedentsj, social supportk, self-beliefl, comparison of outcomesm, shaping knowledgen
Individual sessions:
 Repetition and substitutionc, feedback and monitoringf, goals and planningg, identityh, antecedentsj, self-beliefl, comparison of outcomesm
Phase 2:
6 to 12 months		 No food provided, diet prepared by volunteer;
10% or 30% energy restriction + low- or high-glycemic load		 Daily to weekly:
 Recording of body weight, food intake, appetite, activity, behaviors, and emotions
Biweekly:
 Behavioral support groups
 Individual sessions with dietitian
Pre and post phase:
 Outcome assessments		 Personal-level:
 Repetition and substitutionc, natural consequencese, feedback and monitoringf, goals and planningg, identityh
Group sessions:
 Repetition and substitutionc, goals and planningg, identityh, reward and threati, antecedentsj, social supportk, self-beliefl, comparison of outcomesm, shaping knowledgen, comparison of behavioro
Individual sessions:
 Repetition and substitutionc, feedback and monitoringf, goals and planningg, identityh, antecedentsj, self-beliefl, comparison of outcomesm
Open in a new tab
a

Intervention components were the same for all study groups.

b

Characterized using cluster labels of the Behavior Change Techniques Taxonomy v.1 (Michie et al., 2013). Component behavior change techniques within each cluster are indicated with superscript letters. Techniques employed at the personal-level (P), in group sessions (G), or in individual sessions (I) as indicated.

c

Behavioral practice (P), behavior substitution (G), habit reversal (G), habit formation (G, I), generalization of a target behavior (G, I).

d

Satiation (P), exposure (P), and classical conditioning (P).

e

Health consequences (P), and self-assessment of affective consequences (P).

f

Feedback on behavior (P, I), biofeedback (P, I), other monitoring with awareness (P, I), self-monitoring of outcome behavior (P, I), and self-monitoring of behavior (P, I).

g

Discrepancy from gold standard (P, I), action planning (G), problem solving (G, I), commitment (G), goal setting (outcome/behavior) (G, I), review behavior/outcome goals (G, I).

h

Identification of self as role-model (P, G), self-affirmation (P, G), and identity associated with change behavior (P, G).

i

Incentive (G).

j

Restructuring physical/social environment (G, I), changing exposure to cues for the behavior (G, I).

k

Social support (practical/general/emotional) (G).

l

Focus on past success (G), and verbal persuasion (G, I).

m

Pros and cons (G, I).

n

Instruction on how to perform a behavior (G).

o

Modeling of the behavior (G).

Study Diets

Diets differed in GL and macronutrient composition (high-GL: 60% carbohydrate, 20% fat, 20% protein vs. low-GL: 40% carbohydrate, 30% fat, 30% protein), but were matched for fiber, palatability and energy density (Das et al., 2007). Both diets emphasized low energy density foods and limited liquid energy. During the provided-food phase, snacks and meals were picked up twice weekly. Participants were asked to consume provided food completely and record any deviations from prescribed intake in a daily diary.

Behavioral support

The behavioral component of the intervention was adapted from the LEARN Program for Weight Management (Brownell, 2000). This program stresses behavioral skills including goal setting, self-monitoring, stimulus control, identifying positive feedback, and relapse training.

Program implementation was achieved through group support meetings, individual counseling sessions, and self-monitoring throughout the study (Table 1). Bi-weekly behavioral support groups were attended by participants irrespective of diet randomization. The function of the support groups was to reinforce and promote adherence to study goals for dietary and behavioral self-monitoring, provide group support and positive feedback, and prepare volunteers to maintain their randomization following the provided-food phase. Discussion topics included self-monitoring, menu design, meal planning, cooking, grocery shopping, differentiating between hunger and non-hunger stimuli, relapse control, social and family support, practical strategies for social situations, and eating outside the home. Each participant also met individually with a dietitian every two weeks throughout the entire study. Individual meetings provided personalized support through discussing and troubleshooting individual challenges, and provided additional support for maintaining the intervention during the self-selected diet phase. Homebased self-monitoring was also required throughout the study. Required activities included daily recording of body weight, food intake, appetite and physical activity. Behavior and emotion inventories were also completed. Though the behavioral component of the intervention was designed to promote adherence to energy restriction, some of the behavior change techniques employed such as goal setting, planning, and reinforcement are reported to increase self-efficacy (Olander et al., 2013; Williams & French, 2011). The full complement of behavior change techniques employed during the intervention are categorized in table 1 according to the Behavior Change Technique Taxonomy v1 (Michie et al., 2013).

Anthropometrics

Height was measured by trained study staff to the nearest 0.1 cm at the beginning of the study. Body weight was measured by study staff to the nearest 0.1 kg at the beginning of each 6-mo diet period and at study conclusion.

Compliance with energy restriction and physical activity level

Total energy expenditure was measured by the doubly-labeled water method over a period of 14 consecutive days during each of the pre-intervention, provided-food and self-selected diet phases, as described previously (Das et al., 2007; Das et al., 2009). Energy intake was calculated from total energy expenditure plus estimated change in body energy stores based on body weight change (Bathalon et al., 2000; Das et al., 2009). Mean energy restriction achieved during the provided-food and self-selected diet phases, and during the full 12 mo intervention was calculated as time weighted averages of the individual estimates at separate timepoints (Das et al., 2009). Resting energy expenditure was measured during each 14-d measurement period after an overnight fast by indirect calorimetry (Deltatrac, Sensor Medics Corp, Yorba Linda, CA), and divided into total energy expenditure to calculate physical activity level.

Eating behavior self-efficacy

Eating behavior self-efficacy was measured at 0 mo, 6 mo and 12 mo using the Weight Efficacy Lifestyle Questionnaire (WEL). The WEL is a validated, 20-item self-report questionnaire that measures five situational dimensions of eating behavior self-efficacy (Clark et al., 1991). These dimensions, hereafter referred to as subscales, include availability (i.e., ability to resist eating in environments where food is readily available), negative emotions (i.e., ability to resist eating when feeling nervous, depressed, irritable, or experiencing failure), social pressure (i.e., ability to resist eating when others are encouraging or expecting food consumption), physical discomfort (i.e., ability to resist eating when tired, having a headache or experiencing other bodily pain) and positive activities (i.e., ability to resist eating when watching television, reading or before bed). Responses to each of four questions within each subscale are scored on a 0 to 9 Likert scale with higher scores indicating greater self-efficacy. The total score for each subscale therefore ranges from 0 to 36. Scores for each dimension are summed to derive a total eating behavior self-efficacy score ranging from 0 to 180.

The WEL demonstrated good internal consistency in the present sample (Cronbach’s alpha = 0.71), which is consistent with the internal consistency reported for WEL subscales (Cronbach’s alpha = 0.70 to 0.90) (Clark et al., 1991). External validity of the WEL has been demonstrated in multiple diet intervention studies documenting changes in WEL scores concurrent to weight loss (Clark et al., 1991; Clark et al., 1996; Palmeira et al., 2007; Pinto et al., 1999; Teixeira et al., 2010; Warziski et al., 2008).

Statistical analysis

Eating behavior self-efficacy was a secondary outcome and not considered in a priori power calculations. Rather, sample sizes were derived from power calculations completed using actual energy intake and total energy expenditure as the outcomes of interest, and were estimated to be sufficient for detecting between-group differences of 400 kcal/d in change in energy intake and 350 kcal/d in change in total energy expenditure at a power of 80% and an adjusted alpha of 0.05.

Changes in weight, WEL scores and other variables are expressed as final minus initial. WEL scores were analyzed using both WEL total and subscale scores. We previously reported that mean energy restriction achieved and weight loss did not differ between the 10% and 30% energy restriction groups (Das et al., 2009). These two groups were therefore combined in all analyses in this report, and a term for assigned energy restriction-level included as a covariate in all models.

Pre-intervention variables were compared between GL groups using independent samples t-tests. Mixed models repeated measures analysis of covariance was used to examine changes over time, between-group differences, and time-by-GL interactions in compliance with energy restriction prescription, anthropometric variables, physical activity level, and WEL scores. Pearson’s correlation was used to assess relationships between weight change and WEL scores. Multivariate regression was used to examine pre-intervention eating behavior self-efficacy, and change in eating behavior self-efficacy during the provided-food phase as predictors of weight change during both the provided-food and self-selected diet phases. Stepwise regression with sex, energy restriction-level, and GL-level entered as forced covariates was used to determine the WEL subscales contributing to weight change during both the provided-food and self-selected diet phases.

Primary analyses were completed with only study completers included, but repeated using all enrolled volunteers. Results were statistically similar, and the completers-only analyses are reported herein. Four study completers were missing WEL score or physical activity level data at a single timepoint. These missing data points were imputed in regression analyses using multiple imputation. SPSS version 21.0 was used for all analyses. All tests were two-sided and considered statistically significant at p ≤ 0.05.

Results

Thirty-nine participants, 20 in the high-GL group and 19 in the low-GL group, completed the study (Table 2). Study drop-outs did not differ from study completers by GL-assignment (χ2(1) = 0.17, p = 0.68), sex (χ2(1) = 1.20, p = 0.27), age (t(44) = 0.55, p = 0.58), baseline BMI (t(44) = 0.11, p = 0.91), or pre-intervention WEL total score (t(43) = 0.50, p = 0.62).

Table 2.

Pre-intervention volunteer characteristics.a

High-GL Low-GL
Sex (n; M / F) 4 / 16 5 / 14
Energy restriction-level (n; 10% / 30%) 5 / 13 4 / 14
Age (y) 35 ± 4 35 ± 5
Height (cm) 169 ± 10 169 ± 11
Weight (kg) 79.7 ± 11.5 80.7 ± 10.3
BMI (kg/m2) 27.9 ± 1.7 28.0 ± 1.5
Physical activity level 1.76 ± 0.17 1.71 ± 0.18
Open in a new tab
a

Mean ± SD. GL, dietary glycemic load. No between-group differences.

Measured energy restriction and weight loss

Measured energy restriction achieved did not differ by GL (F(1, 28) = 0.26, p = 0.62), decreasing in the full cohort from 39% of total energy expenditure at 1 mo to 9% at 12 mo (F(2, 60) = 22.4, p < 0.001). Though mean 0–12 mo total weight loss did not differ by GL (F(1, 36) = 0.20, p = 0.66), averaging 8% of initial body weight in the high-GL group and 6% in the low-GL group, the pattern of weight change over the 6 mo periods did differ by GL-level (F(2, 37) = 3.22, p-interaction = 0.05). At the completion of the provided-food period, body weight did not differ significantly between the low-GL (M = 73.1 kg, SD = 10.8 kg) and high-GL groups (M = 72.4 kg, SD = 10.1 kg, p = 0.84); however, during the self-selected diet phase the low-GL group experienced a mean weight regain (M = 2.8 kg, SD = 2.8 kg, p < 0.001) whereas the high-GL group maintained weight loss (M = 0.6 kg, SD = 2.8 kg, p = 1.00). Neither measured energy restriction achieved nor weight change differed by prescribed ER-level during the intervention (Das et al., 2009).

Eating behavior self-efficacy

No statistically significant differences in mean WEL total score were documented during the intervention (F(1, 37) = 2.58, p-interaction = 0.12; Table 3). A time-by-GL interaction (F(1, 37) = 5.48, p-interaction = 0.02) was documented for the negative emotions subscale of the WEL, with mean score at study completion lower in the low-GL relative to the high-GL group. However, no additional effects of GL on WEL scores were documented and the between-group effect sizes for WEL total score and subscale scores were generally small (Table 3).

Table 3.

Weight Efficacy Lifestyle Questionnaire total and subscale scores during intervention.a

0 months 6 months 12 months Cohen’s d
Total score
 High-GL 136 ± 14 136 ± 18 141 ± 19 −0.24
 Low-GL 152 ± 15ⱡ 143 ± 19 137 ± 18
Availability
 High-GL 23 ± 5 23 ± 5 23 ± 6 −0.31
 Low-GL 27 ± 5ⱡ 23 ± 6 23 ± 8
Negative emotionsb
 High-GL 26 ± 5 27 ± 5 28 ± 6 −0.65
 Low-GL 30 ± 6 29 ± 6 26 ± 6ⱡ
Social pressure
 High-GL 27 ± 6 27 ± 6 27 ± 6 0.11
 Low-GL 32 ± 3ⱡ 29 ± 6 29 ± 6
Physical discomfort
 High-GL 30 ± 5 29 ± 5 30 ± 6 0.004
 Low-GL 31 ± 4 31 ± 6 29 ± 4
Positive activities
 High-GL 30 ± 4 31 ± 4 32 ± 4 −0.43
 Low-GL 32 ± 3ⱡ 31 ± 4 31 ± 3
Open in a new tab
a

Mean ± SD, n = 39. GL, dietary glycemic load. Mixed models repeated measures ANCOVA with Bonferroni adjustments. Baseline score and energy restriction level included as a covariates in all models.

b

time-by-glycemic load interaction, p ≤ 0.05.

ⱡ

Significantly different from high-GL, p ≤ 0.05.

As no significant effects of GL on WEL total score were observed, data were combined for exploratory analyses investigating relationships between eating behavior self-efficacy and weight change. Pre-intervention WEL total score was positively associated with weight change during the provided-food phase (r = .40, p = 0.01). Of the subscales, only pre-intervention availability (r = .40, p = 0.01) and social pressure (r = .37, p = 0.02) scores were significantly correlated with weight change during the provided-food phase.

Although mean WEL total scores did not change during the provided-food phase despite significant weight loss, considerable variability was observed with 20 participants reporting a decrease and 19 reporting no change or an increase in eating behavior self-efficacy. Weight change during the provided food phase was inversely associated with concurrent changes in WEL total scores (Figure 1a). Change in WEL total score during the provided food phase was also inversely associated with subsequent weight change during the self-selected diet period (Figure 1b). Within the subscales, weight change during the provided-food phase was inversely associated with concurrent changes in availability (r = −.49, p = 0.001), social pressure (r = −.67, p < 0.001), and positive activities (r = −.40, p = 0.01) scores. However, only changes in availability (r = −.44, p = 0.005) and positive activities (r = −.39, p = 0.01) subscale scores during the provided food phase were significantly associated with weight change during the self-selected diet phase. Neither WEL total score nor any subscale score at the end of the provided-food phase was correlated with weight change during the self-selected diet phase (data not shown).

Figure 1.

Figure 1

Open in a new tab

Relationships between change in eating behavior self-efficacy measured by the Weight Efficacy Lifestyle Questionnaire (WEL) during the provided-food phase and weight loss during a) provided-food phase (0–6 mo), and b) the self-selected diet phase (6–12 mo).

In multivariate models, both pre-intervention WEL total score and change in WEL total score during the provided food phase remained significant positive and negative predictors, respectively, of weight change during the provided-food phase (Table 4). However, neither change in WEL total score during the provided-food phase nor WEL total score upon completing the provided-food phase were significantly associated with weight change during the self-selected diet phase (Table 4). In order to identify which components of eating behavior self-efficacy might have influenced weight change, stepwise multiple regression was conducted with pre-intervention WEL subscale scores, WEL subscale change scores during the provided-food phase, and WEL subscale scores at the end of the provided-food phase considered as predictors. The strongest prediction model for weight loss during the provided-food phase included pre-intervention availability score and change in social pressure score during the provided-food phase as significant positive and negative predictors, respectively (Table 4). In the model predicting weight loss during the self-selected diet phase, change in availability score during the provided-food phase and physical discomfort score at the end of the provided-food phase both entered as negative predictors (Table 4).

Table 4.

Eating behavior self-efficacy as a predictor of body weight at 6 and 12 months.a

Weight 6 mo (kg)
β ± SE P-value β ± SE P-value
Total score models Subscale score model b
Model 1b  Availability score, 0 mo 0.22 ± 0.10 0.02
 WEL total score, 0 mo 0.12 ± 0.04 0.005  Social pressure score, Δ0–6 mo −0.33 ± 0.07 < 0.001
 Adjusted R2 0.88 < 0.001  Adjusted R2 0.93 < 0.001
Model 2b
 WEL total score, Δ0–6 mo −0.12 ± 0.03 < 0.001
 Adjusted R2 0.90 < 0.001
Weight 12 mo (kg)
β ± SE P-value β ± SE P-value
Total score models Subscale score model c
Model 1c  Availability score, Δ0–6 mo −0.16 ± 0.07 0.02
 WEL total score, 6 mo −0.03 ± 0.02 0.24  Physical discomfort score, 6 mo −0.17 ± 0.08 0.04
 Adjusted R2 0.95 < 0.001  Adjusted R2 0.96 < 0.001
Model 2c
 WEL total score, Δ0–6 mo −0.03 ± 0.02 0.14
 Adjusted R2 0.95 < 0.001
Open in a new tab
a

n = 39. WEL, Weight Efficacy Lifestyle Questionnaire.

b

Covariates in model: glycemic load-level, energy restriction-level, sex, and body weight at 0 mo.

c

Covariates in model: glycemic load-level, energy restriction-level, sex, and body weight at 6 mo.

Physical activity and eating behavior self-efficacy

Pre-intervention physical activity level and WEL total score were not correlated (r = −.18, p = 0.28). Change in WEL total score during the provided-food phase did not correlate with changes in physical activity level during the same time period (r = .02, p = 0.93) or during the subsequent self-selected diet phase (r = −.08, p = 0.67). However, the change in WEL total score during the self-selected diet phase was correlated with concomitant changes in physical activity level (r = .50, p = 0.001).

Discussion

Identifying factors that influence psychosocial predictors and correlates of weight loss is critical for improving behavioral interventions for healthy weight management. This study examined interrelationships between eating behavior self-efficacy and weight loss during a 12-mo dietary intervention of low-GL and high-GL energy restricted diets in which food was provided for the first six months and self-selected thereafter. Our findings suggest that dietary GL has little effect on eating behavior self-efficacy during weight loss. Moreover, weight loss while consuming a provided energy restricted diet unexpectedly did not consistently improve eating behavior self-efficacy, suggesting that provided-food weight loss diets may not reliably facilitate improvements in eating behavior self-efficacy that could help sustain weight loss.

This study is the first to examine the effects of GL on eating behavior self-efficacy. Low-GL diets are thought to promote satiety (Roberts, 2000), and we have previously reported that within this cohort the desire to eat non-study foods increased during the first 3 mo in the 30% energy restriction + high-GL group but not the 30% energy restriction + low-GL group (Das et al., 2007). We speculated that these hunger- and temptation-suppressing effects of low-GL diets would lead to greater improvements in eating behavior self-efficacy in the low-GL group. However, with the exception of decreased self-efficacy to resist eating during emotional distress documented in the low-GL group, dietary GL had little effect on eating behavior self-efficacy. These findings do not support our hypothesis. However, the finding that GL does not significantly differentially affect eating behavior self-efficacy is nonetheless important and encouraging given that low-GL diets are often considered healthier than high-GL diets, but may be more challenging and require greater effort to plan, purchase and prepare (Brekke, Sunesson, Axelsen, & Lenner, 2004).

The finding that pre-intervention eating behavior self-efficacy was inversely associated with weight loss was unexpected and contrary to the concept of heightened eating behavior self-efficacy as a reliable predictor of weight loss (U.S. Department of Health and Human Services et al., 2000). This incongruence could be attributable to differences in measurement methodology as general measures of self-efficacy are postulated to be more predictive of subsequent weight loss than the specific dimensions of eating behavior self-efficacy measured by the WEL (Teixeira et al., 2005). However, a number of recent studies have also challenged the concept of heightened eating behavior self-efficacy as a reliable predictor of weight loss by failing to document associations between pre-treatment eating behavior self-efficacy and subsequent weight loss (Byrne, Barry, & Petry, 2012; Linde et al., 2004; Smith, Sondhaus, & Porzelius, 1995; Teixeira et al., 2002), and reporting that higher pre-treatment eating behavior self-efficacy was associated with less weight loss (Martin, Dutton, & Brantley, 2004; Wingo et al., 2013). The latter findings were attributed to overconfidence and underestimating the difficulties of behavior change. The same factors may also partly explain our results as the pre-intervention WEL scores were higher than those previously reported in overweight and obese adults (e.g., (Bas & Donmez, 2009; Clark et al., 1991; Richman, Loughnan, Droulers, Steinbeck, & Caterson, 2001; Teixeira et al., 2010; Warziski et al., 2008)), and more similar to those reported in normal-weight women (Richman et al., 2001) and following successful weight loss (Clark et al., 1996; Pinto et al., 1999).

Interestingly, during the provided-food phase greater weight loss was associated with improvements, but less weight loss with decrements in eating behavior self-efficacy resulting in no net effect on eating behavior self-efficacy. This finding is consistent with a recent combined analysis of participants from the PREMIER trial in whom dietary self-efficacy decreased on average despite substantial weight loss (Wingo et al., 2013). Nonetheless, our results were unexpected as increased self-efficacy concomitant to weight loss has frequently been reported in adults consuming self-selected diets (Bernier & Avard, 1986; Glynn & Ruderman, 1986; Palmeira et al., 2007; Teixeira et al., 2010; Warziski et al., 2008), a finding that is consistent with self-efficacy theory which states that personal achievements that contribute to realizing a desired outcome (i.e., weight loss) benefit self-efficacy (Bandura, 1977). A potential explanation for our unexpected result may be that more opportunities for personal achievements to increase self-efficacy are available when self-selecting diets. For example, weight loss on provided energy restricted-diets results primarily from avoidance of foods not provided whereas losing weight on self-selected diets requires additional skills such as those needed to identify and prepare healthy meals.

In univariate models, improvements in eating behavior self-efficacy were associated with less weight regain when food was self-selected. However, this association was attenuated and not significant when additional variables were included in the model. Likewise, Teixeira et al. (Teixeira et al., 2010) recently reported that improvements in eating behavior self-efficacy documented during a weight loss intervention were not associated with post-intervention weight loss maintenance. However, the behaviors required to maintain weight loss likely differ from those required to lose weight (Stubbs et al., 2011; Wing & Phelan, 2005), and the participants in our study were asked to continue energy restriction throughout the entire 12 mo period. Moreover, to the best of our knowledge, this is the first study to examine whether changes in eating behavior self-efficacy facilitated by providing energy restricted-diets predict subsequent weight change upon resuming a self-selected diet. The findings suggest that provided-food weight loss diets may not reliably facilitate improvements in eating behavior self-efficacy that could help sustain weight loss. Interestingly, analyses of WEL subscale scores suggest that long-term success of weight-loss interventions may be particularly influenced by beliefs captured by the availability subscale, specifically beliefs concerning one’s ability to resist eating in situations where food is readily available and often over-consumed.

While only correlational, the observed associations between WEL scores and physical activity levels provide some support for the theory that interventions affecting one specific dimension of self-efficacy (e.g., eating behavior) may similarly impact separate, but related, dimensions of self-efficacy (e.g., exercise). However, the finding that eating behavior self-efficacy was only correlated with physical activity level when diets were self-selected suggests that providing food may attenuate a generalized increase in weight loss self-efficacy. Possibly, as mentioned above, provided-energy restricted diets do not provide sufficient opportunities to experience the small mastery experiences required to cumulatively produce a general increase in weight management-related self-efficacy. Of note, we did not specifically measure exercise self-efficacy. Therefore, these results should be considered preliminary but warranting of further investigation as higher exercise self-efficacy (Teixeira et al., 2010) and physical activity appear to be important predictors of successful weight loss maintenance (Pronk & Wing, 1994).

Novel aspects of our study include the use of a unique design that provided all food for six months, measurement of adherence to the dietary regimen with doubly-labeled water, and serial measurements of eating behavior self-efficacy and body weight. By having all of their food provided, participants were repeatedly challenged within situational dimensions of eating behavior self-efficacy as measured by the WEL. A limitation to this approach is that other aspects of self-efficacy for weight loss such as a belief in one’s ability to choose or prepare healthy meals may not be captured. Moreover, definitive conclusions regarding the effects of providing food on eating behavior self-efficacy are precluded by the absence of a control group experiencing similar amounts of weight loss on a self-selected diet. The high pre-intervention WEL scores in this cohort are also a limitation. Though substantial variability in WEL change scores allowed us to assess relationships between weight change and eating behavior self-efficacy, a ceiling effect may have weakened our ability to detect these relationships. Finally, the intervention was designed and powered to study changes in metabolic and physiologic parameters associated with energy restriction, not psychosocial outcomes. Given the small sample size these findings should be considered preliminary, and verification in larger trials is needed.

Our findings suggest that GL does not influence eating behavior self-efficacy, and provided energy restricted-diets may not facilitate improvements in eating behavior self-efficacy unless substantial weight loss is achieved. With the growing popularity of weight loss treatment programs that provide full or partial diets, additional research is warranted to determine the impact of these diets on self-efficacy, approaches to optimize eating behavior self-efficacy while following these diets, and the implications for subsequent weight loss success.

Highlights.

  • High eating behavior self-efficacy may contribute to successful weight loss.

  • A provided low glycemic load weight loss diet did not affect eating self-efficacy.

  • Eating self-efficacy decreased in those losing less weight during food provision.

  • Self-efficacy did not consistently predict weight regain after food provision.

  • Provided-food weight loss diets may not reliably improve in self-efficacy.

Acknowledgements

The material is based upon work supported by the U.S. Department of Agriculture, Agricultural Research Service, under agreement No. 58-1950-7-707. Any opinions, findings, conclusion, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the U.S. Department of Agriculture, the U.S. Army, or the U.S. Department of Defense.

The Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy trial- phase I study was supported by National Institutes of Health grant U01-AG20480, the U.S. Department of Agriculture under agreement No. 58-1950-4-401, K23 DK61506 from the National Institutes of Diabetes and Digestive and Kidney Diseases, and Boston Nutrition Obesity Research Center (BONRC) H150001. JPK is supported by the Science, Mathematics, and Research Transformation Defense Education Program. RAC was supported by a NIH T32 grant (#DK62032-11).

We wish to thank the study volunteers and the staff of the Metabolic Research Unit for their dedication to this project.

U.S. National Institutes of Health clinicaltrials.gov identifier: NCT00099099

Abbreviations:

GL

glycemic load

WEL

Weight Efficacy Lifestyle Questionnaire

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of interest statement: No authors report a conflict of interest.

References

  1. Annesi JJ, & Whitaker AC (2010). Psychological factors associated with weight loss in obese and severely obese women in a behavioral physical activity intervention. Health Educ Behav, 37, 593–606. [DOI] [PubMed] [Google Scholar]
  2. Bandura A (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev, 84, 191–215. [DOI] [PubMed] [Google Scholar]
  3. Bandura A (2004). Health promotion by social cognitive means. Health Educ Behav, 31, 143–164. [DOI] [PubMed] [Google Scholar]
  4. Baranowski T, Cullen KW, Nicklas T, Thompson D, & Baranowski J (2003). Are current health behavioral change models helpful in guiding prevention of weight gain efforts? Obes Res, 11 Suppl, 23S–43S. [DOI] [PubMed] [Google Scholar]
  5. Bas M, & Donmez S (2009). Self-efficacy and restrained eating in relation to weight loss among overweight men and women in Turkey. Appetite, 52, 209–216. [DOI] [PubMed] [Google Scholar]
  6. Bathalon GP, Tucker KL, Hays NP, Vinken AG, Greenberg AS, McCrory MA, & Roberts SB (2000). Psychological measures of eating behavior and the accuracy of 3 common dietary assessment methods in healthy postmenopausal women. Am J Clin Nutr, 71, 739–745. [DOI] [PubMed] [Google Scholar]
  7. Batra P, Das SK, Salinardi T, Robinson L, Saltzman E, Scott T,Pittas AG, Roberts SB (2013). Eating behaviors as predictors of weight loss in a 6 month weight loss intervention. Obesity. doi: 10.1002/oby.20404 [DOI] [PubMed] [Google Scholar]
  8. Bernier M, & Avard J (1986). Self-efficacy, outcome, and attrition in a weight-reduction program. Cognit Ther Res, 10, 319–338. [Google Scholar]
  9. Brekke HK, Sunesson A, Axelsen M, & Lenner RA (2004). Attitudes and barriers to dietary advice aimed at reducing risk of type 2 diabetes in first-degree relatives of patients with type 2 diabetes. J Hum Nutr Diet, 17, 513–521. [DOI] [PubMed] [Google Scholar]
  10. Brownell KD (2000). The LEARN Program for weight management 2000. Dallas, Texas: American Health. [Google Scholar]
  11. Byrne S, Barry D, & Petry NM (2012). Predictors of weight loss success. Exercise vs. dietary self-efficacy and treatment attendance. Appetite, 58, 695–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Clark MM, Abrams DB, Niaura RS, Eaton CA, & Rossi JS (1991). Self-efficacy in weight management. J Consult Clin Psychol, 59, 739–744. [DOI] [PubMed] [Google Scholar]
  13. Clark MM, Cargill BR, Medeiros ML, & Pera V (1996). Changes in self-efficacy following obesity treatment. Obes Res, 4, 179–181. [DOI] [PubMed] [Google Scholar]
  14. Cuntz U, Leibbrand R, Ehrig C, Shaw R, & Fichter MM (2001). Predictors of posttreatment weight reduction after in-patient behavioral therapy. Int J Obes Relat Metab Disord, 25 Suppl 1, S99–S101. [DOI] [PubMed] [Google Scholar]
  15. Das SK, Gilhooly CH, Golden JK, Pittas AG, Fuss PJ, Cheatham RA, … Roberts SB (2007). Long-term effects of 2 energy-restricted diets differing in glycemic load on dietary adherence, body composition, and metabolism in CALERIE: a 1-y randomized controlled trial. Am J Clin Nutr, 85, 1023–1030. [DOI] [PubMed] [Google Scholar]
  16. Das SK, Saltzman E, Gilhooly CH, DeLany JP, Golden JK, Pittas AG, Roberts SB (2009). Low or moderate dietary energy restriction for long-term weight loss: what works best? Obesity, 17, 2019–2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Elfhag K, & Rossner S (2005). Who succeeds in maintaining weight loss? A conceptual review of factors associated with weight loss maintenance and weight regain. Obes Rev, 6, 67–85. [DOI] [PubMed] [Google Scholar]
  18. Glynn SM, & Ruderman AJ (1986). The development and validation of an eating self-efficacy scale. Cognit Ther Res, 10, 403–420. [Google Scholar]
  19. Karlsson J, Hallgren P, Kral J, Lindroos AK, Sjostrom L, & Sullivan M (1994). Predictors and effects of long-term dieting on mental well-being and weight loss in obese women. Appetite, 23, 15–26. [DOI] [PubMed] [Google Scholar]
  20. Linde JA, Jeffery RW, Levy RL, Sherwood NE, Utter J, Pronk NP, & Boyle RG (2004). Binge eating disorder, weight control self-efficacy, and depression in overweight men and women. Int J Obes Relat Metab Disord, 28, 418–425. [DOI] [PubMed] [Google Scholar]
  21. Livesey G, Taylor R, Hulshof T, & Howlett J (2008). Glycemic response and health--a systematic review and meta-analysis: relations between dietary glycemic properties and health outcomes. Am J Clin Nutr, 87, 258S–268S. [DOI] [PubMed] [Google Scholar]
  22. Ludwig DS (2002). The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease. JAMA, 287(18), 2414–2423. [DOI] [PubMed] [Google Scholar]
  23. Martin PD, Dutton GR, & Brantley PJ (2004). Self-efficacy as a predictor of weight change in African-American women. Obes Res, 12, 646–651. [DOI] [PubMed] [Google Scholar]
  24. McGuire MT, Wing RR, Klem ML, Lang W, & Hill JO (1999). What predicts weight regain in a group of successful weight losers? J Consult Clin Psychol, 67, 177–185. [DOI] [PubMed] [Google Scholar]
  25. Michie S, Richardson S, Johnston M, Abraham C, Francis J, Hardeman W, Eccles MP, Cane J, & Wood CE (2013) Ann Behav Med, 46, 81–95. [DOI] [PubMed] [Google Scholar]
  26. Olander EK, Fletcher H, Williams S, Atkinson L, Turner A, & French DP (2013). What are the most effective techniques in changing obese individuals’ physical activity self-efficacy and behaviour: a systematic review and meta-analysis. Int J Behav Nutr Phys Act, 10, 29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Palmeira AL, Teixeira PJ, Branco TL, Martins SS, Minderico CS, Barata JT, Sardinha LB (2007). Predicting short-term weight loss using four leading health behavior change theories. Int J Behav Nutr Phys Act, 4, 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Pasman WJ, Saris WH, & Westerterp-Plantenga MS (1999). Predictors of weight maintenance. Obes Res, 7, 43–50. [DOI] [PubMed] [Google Scholar]
  29. Pinto BM, Clark MM, Cruess DG, Szymanski L, & Pera V (1999). Changes in self-efficacy and decisional balance for exercise among obese women in a weight management program. Obes Res, 7, 288–292. [DOI] [PubMed] [Google Scholar]
  30. Presnell K, Pells J, Stout A, & Musante G (2008). Sex differences in the relation of weight loss self-efficacy, binge eating, and depressive symptoms to weight loss success in a residential obesity treatment program. Eating Behaviors, 9, 170–180. [DOI] [PubMed] [Google Scholar]
  31. Pronk NP, & Wing RR (1994). Physical activity and long-term maintenance of weight loss. Obes Res, 2, 587–599. [DOI] [PubMed] [Google Scholar]
  32. Richman RM, Loughnan GT, Droulers AM, Steinbeck KS, & Caterson ID (2001). Self-efficacy in relation to eating behaviour among obese and non-obese women. Int J Obes Relat Metab Disord, 25, 907–913. [DOI] [PubMed] [Google Scholar]
  33. Roberts SB (1989). Use of the doubly labeled water method for measurement of energy expenditure, total body water, water intake, and metabolizable energy intake in humans and small animals. Can J Physiol Pharmacol, 67, 1190–1198. [DOI] [PubMed] [Google Scholar]
  34. Roberts SB (2000). High-glycemic index foods, hunger, and obesity: is there a connection? Nutr Rev, 58, 163–169. [DOI] [PubMed] [Google Scholar]
  35. Schoeller DA (1988). Measurement of energy expenditure in free-living humans by using doubly labeled water. J Nutr, 118, 1278–1289. [DOI] [PubMed] [Google Scholar]
  36. Smith MC, Sondhaus E, & Porzelius LK (1995). Effect of binge eating on the prediction of weight loss in obese women. J Behav Med, 18, 161–168. [DOI] [PubMed] [Google Scholar]
  37. Stubbs J, Whybrow S, Teixeira P, Blundell J, Lawton C, Westenhoefer J, Raats M (2011). Problems in identifying predictors and correlates of weight loss and maintenance: implications for weight control therapies based on behaviour change. Obes Rev, 12, 688–708. [DOI] [PubMed] [Google Scholar]
  38. Teixeira PJ, Going SB, Houtkooper LB, Cussler EC, Martin CJ, Metcalfe LL, Lohman TG (2002). Weight loss readiness in middle-aged women: psychosocial predictors of success for behavioral weight reduction. J Behav Med, 25, 499–523. [DOI] [PubMed] [Google Scholar]
  39. Teixeira PJ, Going SB, Sardinha LB, & Lohman TG (2005). A review of psychosocial pre-treatment predictors of weight control. Obes Rev, 6, 43–65. [DOI] [PubMed] [Google Scholar]
  40. Teixeira PJ, Silva MN, Coutinho SR, Palmeira AL, Mata J, Vieira PN, Sardinha LB (2010). Mediators of weight loss and weight loss maintenance in middle-aged women. Obesity, 18, 725–735. [DOI] [PubMed] [Google Scholar]
  41. Thomas DE, Elliott EJ, & Baur L (2007). Low glycaemic index or low glycaemic load diets for overweight and obesity. Cochrane Database Syst Rev, 3, CD005105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, & National Heart Lung and Blood Institute. (2000). The practical guide. Identification, evaluation and treatment of overweight and obesity in adults.
  43. Warziski MT, Sereika SM, Styn MA, Music E, & Burke LE (2008). Changes in self-efficacy and dietary adherence: the impact on weight loss in the PREFER study. J Behav Med, 31, 81–92. [DOI] [PubMed] [Google Scholar]
  44. Williams SL, & French DP (2011). What are the most effective intervention techniques for changing physical activity self-efficacy and physical activity behaviour-and are they the same? Health Educ Res, 26, 308–322. [DOI] [PubMed] [Google Scholar]
  45. Wing RR, & Phelan S (2005). Long-term weight loss maintenance. Am J Clin Nutr, 82, 222S–225S. [DOI] [PubMed] [Google Scholar]
  46. Wingo BC, Desmond RA, Brantley P, Appel L, Svetkey L, Stevens VJ, & Ard JD (2013). Self-efficacy as a Predictor of Weight Change and Behavior Change in the PREMIER Trial. J Nutr Educ Behav, 45, 314–321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yaemsiri S, Slining MM, & Agarwal SK (2011). Perceived weight status, overweight diagnosis, and weight control among US adults: the NHANES 2003–2008 Study. Int J Obes, 35, 1063–1070. [DOI] [PubMed] [Google Scholar]

RESOURCES