Abstract
Context
Increased intake of fruits and vegetables (FV) may be useful for weight loss.
Objective
This study was designed to evaluate the effects of changes in FV intake on the overall diet and to determine if self-regulation affects the association between changes in FV intake and weight.
Methods
Women with obesity (N = 74, mean age = 47.7 years) participating in a year-long behavioral weight-loss treatment were assessed regarding changes in consumption of various food groups, physical activity, and eating self-regulation over 6, 12, and 24 months.
Results
FV intake change significantly predicted changes in weight and consumption of dairy and bread products and sweets. The only other notable relationships were among changes in sweets, bread, and dairy consumption over 24 months. Over 6, 12, and 24 months, changes in self-regulation significantly mediated the FV intake-weight change relationship. The overall mediation models were significant (R2 values = 0.19, 0.13, and 0.32, respectively). A reciprocal relationship between changes in FV intake and self-regulation also was found. Significant increases in physical activity outputs did not influence weight changes.
Conclusion
Findings supported a relationship between FV intake and weight loss occurring through self-regulatory skills. Associations between FV intake and reduced consumption of other food groups provided data useful for improving the architecture of behavioral weight-loss treatments and the foci of medical practitioners’ helping methods.
INTRODUCTION
Obesity is a medical condition with considerable effects on health risks and quality of life. Although the popular focus on targeting proportions of diet macronutrients (ie, proteins, fats, and carbohydrates) for weight loss has proved ineffective,1 fruit and vegetable (FV) intake has been associated with weight loss.2,3 Although FV intake can influence consumption of other food groups,1 this has been an understudied area. Targeting as few behavioral changes as possible may serve as a therapeutic advantage when designing weight-loss treatments.4
Because it takes considerable self-control to eat more healthy food and less unhealthy food such as sweets, self-regulation instruction has become a focus of behavioral weight-loss treatments that is based on accepted behavior-change theories such as social cognitive theory.5,6 It is possible that the increased use of self-regulation methods (eg, goal setting/progress tracking, restructuring unproductive self-talk, environmental [stimulus] control) mediates the FV-weight loss relationship and, further, has a symbiotic, mutually reinforcing relationship with FV intake.
Women with obesity who participated in a community-based behavioral weight-loss program were assessed over 6, 12, and 24 months. It was expected that increased FV intake would be significantly associated with reduced weight and reduced consumption of other food groups. Investigators also predicted that the relationship between increased FV intake and reduced weight would be significantly mediated by treatment-associated changes in dietary self-regulation, and that relationships between changes in FV intake and self-regulation would be reciprocal.
METHODS
Participants
Participant (N = 74) data were derived from a longitudinal test of treatment-associated changes in emotional eating.7 The volunteers were required to be at least age 21 years with a body mass index (BMI) of 30 kg/m2 to 40 kg/m2 and not be participating in a weight-loss program or taking a psychotropic medication. Women with physical contraindications to participation (such as pregnancy) also were excluded. The mean age was 47.7 years (standard deviation [SD] = 8.0), and the mean BMI was 34.9 kg/m2 (SD = 3.2). The racial/ethnic makeup was approximately two-thirds white and one-third African American; several participants represented other racial categories. Nearly all were from households in the middle family income range of $50,000 to $100,000 yearly. University institutional review board approval and written informed consent were obtained.
Measures
Eating Behaviors
Daily servings of vegetables (eg, 118 mL [4 ounces] of peas); fruits (eg, 118 mL [4 ounces] of an orange); dairy products (eg, 236 mL [8 ounces] of yogurt); meats, beans, nuts, and other proteins (eg, 59 mL [2 ounces] of boneless chicken); bread products (eg, 59 mL [2 small slices] of whole-grain bread); and sweets (eg, 118 mL of a small brownie) as indicated by US Department of Agriculture sources8 were summed using a brief survey. FV intake was then aggregated. Response scores demonstrated significant correspondences (β values = 0.45–0.83, all p values < 0.001) with the scores of more comprehensive and extensively validated food frequency recalls, and test-retest reliabilities over 3 weeks ranged between 0.77 and 0.83.9,10
Eating Self-Regulation
To assess the use of self-regulation skills related to eating, responses to 10 items (1 = never to 5 = often) such as “I say positive things to myself about eating well” and “I keep a record of my eating” were summed.11 The internal consistency was Cronbach α = 0.81, and test-retest reliability over 2 weeks was 0.74.11 For the data within this study, Cronbach α = 0.80.
Physical Activity
The number of physical activity sessions per week of at least 15 minutes’ duration was recalled using the Leisure-Time Physical Activity Questionnaire (Section 1).12 Activity types corresponded to metabolic equivalents (MET) equating O2/kg/hour ranging from 3 MET (eg, easy walking) to 9 MET (running) and were summed.12 Previous research13,14 found significant relationships (β values = 0.38–0.57, all p values < 0.001) between the Leisure-Time Physical Activity Questionnaire and accelerometer, weight change, and maximal oxygen uptake results. Test-retest reliability over 2 weeks was 0.74.14
Weight
Body weight was measured using a recently calibrated digital scale. The mean of two consecutive measurements was recorded (in kg) after removal of heavy outer clothing.
Procedure
The treatment consisted of components of physical activity support and eating-behavior change based on tenets of social cognitive theory,5 especially the emphasis on use of self-regulatory skills to effectively deal with lifestyle barriers.6 The treatment was administered by trained wellness staff members at a community wellness center.
The Coach Approach protocol15 supported physical activity via six 45-minute 1-on-1 meetings. Assimilation of self-regulation methods such as goal setting, thought-stopping/cognitive restructuring, relapse prevention, dissociation, and stimulus control was the central focus. After 10 weeks of physical activity support alone, instruction in food logging was added, and group nutrition sessions were held every 2 weeks using the Weight Loss For Life protocol.16 Sessions adapted self-regulatory skills learned in the context of supporting physical activity to control eating. For example, restructuring unproductive statements that could lead to missing a planned evening physical activity session would be adapted to address self-talk about circumventing dietary goals for the day. Along with self-regulation and maintaining regular physical activity, increasing FV consumption was a major emphasis of the 56-week treatment. Fidelity checks on approximately 15% of treatment sessions indicated few protocol deviations that were easily remedied.
Data Analyses
Because there was no systematic bias detected within the 8% of missing cases, the intention-to-treat approach incorporated the expectation-maximization algorithm17 for imputation. For the primary analyses, a moderate effect of f2 = 0.15 at the conservative statistical power level of 0.90 (α < 0.05) necessitated a minimum sample size of 71.18 SPSS Statistics version 22.0 (IBM, Armonk, NY) was used for the statistical analyses.
Linear bivariate relationships first were calculated across changes in daily consumption of food groups and weight, separately from baseline to months 6, 12, and 24, at which statistical significance was set at α < 0.05 (2-tailed). Considering previous research and suggestions19 and an absence of floor and ceiling effects, change (gain) scores were unadjusted for their baseline value. Mediation of the relationship of changes in FV intake and weight by change in self-regulation for eating over 6, 12, and 24 months was computed next using a bootstrapping method with 20,000 resamples.20 Because directionality within the incorporated relationships was already established, 1-tailed analyses (α < 0.05) were used to evaluate statistical significance. R2 was calculated to determine significance of the overall models, and path a (predictor→mediator), path b (mediator→outcome), and path c’ (predictor→outcome, controlling for the mediator) were also calculated as β values (Figure 1). Mediation significance was established when the bias-corrected and accelerated 95% confidence intervals did not include 0. Change in physical activity was tested for its significance as a covariate.
Figure 1.
Mediation analyses diagram.
On the basis of previous suggestions,21 the presence of a reciprocal (bidirectional) relationship between changes in FV intake and self-regulation also was tested by reversing the position of the predictor and mediator variables in paired equations. If significant mediation was found in both equations, then a reciprocal relationship had been detected.
RESULTS
Score Changes
Physical activity significantly increased (all p values < 0.001) from a baseline mean of 6.50 (SD = 7.89) to a mean of 35.16 (SD = 19.69) at month 6, a mean of 29.59 (SD = 16.63) at month 12, and a mean of 30.06 (SD = 17.41) at month 24. Self-regulation for eating at baseline (mean = 22.09, SD = 5.85) demonstrated a significant increase to month 6 (meanchange = 9.79, SD = 7.13, p < 0.001), month 12 (meanchange = 6.25, SD = 9.12, p < 0.001), and month 24 (meanchange = 8.03, SD = 8.49, p < 0.001). Table 1 details all other score changes.
Table 1.
Descriptive statistics, change scores, and intercorrelations between changes in food group consumption and weight over 6, 12, and 24 months (N = 74)
| Category | Baseline, mean (SD) | Assessment month, mean (SD) | ΔBaseline-assessment month, mean (SD) | Category intercorrelations | |||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | ||||
| Assessment month 6 | |||||||||
| 1. Fruit/vegetable intake/d | 3.73 (1.53) | 5.65 (1.93) | 1.92a (1.79) | −0.11 | 0.22 | −0.26b | −0.28b | −0.28b | |
| 2. Dairy product intake/d | 1.69 (1.02) | 1.69 (0.93) | 0.00 (1.09) | 0.31c | 0.22 | 0.21 | 0.06 | ||
| 3. Meat/protein intake/d | 2.93 (0.98) | 2.86 (1.16) | −0.33 (0.520) | 0.00 | −0.12 | −0.13 | |||
| 4. Bread product intake/d | 3.05 (1.50) | 2.12 (1.17) | −0.93a (1.50 | 0.52c | 0.18 | ||||
| 5. Sweets intake/d | 1.86 (1.14) | 1.01 (1.03) | −0.86a (1.20) | 0.17 | |||||
| 6. Weight (kg) | 94.77 (12.19) | 88.77 (12.26) | −6.00a (4.87) | ||||||
| Assessment month 12 | |||||||||
| 1. Fruit/vegetable intake/d | 3.73 (1.53) | 5.59 (2.01) | 1.86a (1.880 | −0.38c | 0.17 | −0.39c | −0.27b | −0.30c | |
| 2. Dairy product intake/d | 1.69 (1.02) | 1.65 (1.02) | −0.04 (1.10) | 0.01 | 0.27b | 0.21 | 0.17 | ||
| 3. Meat/protein intake/d | 2.93 (0.98) | 2.74 (0.93) | −0.18 (1.05) | 0.08 | 0.01 | −0.01 | |||
| 4. Bread product intake/d | 3.05 (1.50) | 2.01 (1.06) | −1.04a (1.53) | 0.33c | 0.13 | ||||
| 5. Sweets intake/d | 1.86 (1.14) | 0.92 (0.78) | −0.94a (1.36) | 0.09 | |||||
| 6. Weight (kg) | 94.77 (12.19) | 89.30 (12.79) | −5.48a (5.97) | ||||||
| Assessment month 24 | |||||||||
| 1. Fruit/vegetable intake/d | 3.73 (1.53) | 5.87 (2.07) | 2.14a (2.08) | −0.31c | 0.14 | −0.34c | −0.24b | −0.51c | |
| 2. Dairy product intake/d | 1.69 (1.02) | 1.64 (0.98) | −0.05 (1.20) | 0.03 | 0.23 | 0.27b | 0.22 | ||
| 3. Meat/protein intake/d | 2.93 (0.98) | 2.83 (0.99) | −0.09 (1.21) | 0.12 | −0.08 | 0.04 | |||
| 4. Bread product intake/d | 3.05 (1.50) | 2.04 (1.35) | −1.01a (1.95) | 0.34c | 0.04 | ||||
| 5. Sweets intake/d | 1.86 (1.14) | 1.09 (1.05) | −0.78a (1.51) | 0.15 | |||||
| 6. Weight (kg) | 94.77 (12.19) | 89.35 (12.6) | −5.43a (6.41) | ||||||
p < 0.001.
p < 0.05.
p < 0.01.
SD = standard deviation.
Bivariate Relationships
FV intake changes from baseline to months 6, 12, and 24 served as significant predictors of changes in weight and bread and sweets consumption over 12 and 24 months and dairy consumption at 24 months (Table 1).
Mediation Analyses
When eating self-regulation was entered as a mediator of the relationships of change in FV intake and weight over 6, 12, and 24 months, change in physical activity did not demonstrate a significant effect on the prediction of weight change (β = −0.02, SE = 0.03; β = −0.04, SE = 0.04; and β = −0.07, SE = 0.04, respectively). As a result, this variable was removed from further analyses. The overall mediation models were each significant (R2 = 0.19, p = 0.001; R2 = 0.13, p = 0.006; and R2 = 0.32, p < 0.001). Change in self-regulation demonstrated significant mediation in each of the 3 equations (Table 2).
Table 2.
Mediation and reciprocal effects analyses (N = 74)a
| Predictor | Mediator | Outcome | Path a | Path b | Path c’ | Indirect effect through the mediator | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| β (SE) | p value | β (SE) | p value | β (SE) | p value | β (SE) | 95% CI | |||
| Changes from baseline to month 6 | ||||||||||
| Fruit/vegetable intake | Self-regulation | Weight | 1.26 (0.45) | 0.003 | −0.24 (0.08) | 0.001 | −0.45 (0.31) | 0.073 | −0.31 (0.20) | −0.695, −0.052 |
| Self-regulation for eating | Fruit/vegetable intake | Weight | 0.08 (0.03) | 0.003 | −0.45 (0.31) | 0.073 | −0.24 (0.08) | 0.001 | −0.04 (0.03) | −0.116, −0.002 |
| Changes from baseline to month 12 | ||||||||||
| Fruit/vegetable intake | Self-regulation | Weight | 1.57 (0.44) | < 0.001 | −0.18 (0.09) | 0.031 | −0.67 (0.38) | 0.042 | −0.28 (0.21) | −0.753, −0.033 |
| Self-regulation for eating | Fruit/vegetable intake | Weight | 0.10 (0.03) | < 0.001 | −0.67 (0.38) | 0.042 | −0.18 (0.09) | 0.031 | −0.07 (0.04) | −0.136, −0.015 |
| Changes from baseline to month 24 | ||||||||||
| Fruit/vegetable intake | Self-regulation | Weight | 1.75 (0.41) | < 0.001 | −0.22 (0.09) | 0.008 | −1.19 (0.34) | < 0.001 | −0.38 (0.18) | −0.792, −0.153 |
| Self-regulation for eating | Fruit/vegetable intake | Weight | 0.12 (0.03) | < 0.001 | −1.19 (0.34) | < 0.001 | −0.22 (0.09) | 0.008 | −0.14 (0.06) | −0.262, −0.055 |
Analyses were based on a 1-tailed bootstrapping method to assess mediation that incorporated 20,000 resamples of the data (path a [predictor→mediator], path b [mediator→outcome], and path c’ [predictor→outcome, controlling for the mediator]).20
CI = confidence interval; SE = standard error.
Reciprocal Relationship Analyses
In models in which the predictor and mediator variables were reversed, there were significant mediations (Table 2). This indicated reciprocal, bidirectional relationships between changes in self-regulation and FV intake.
DISCUSSION
Findings supported relationships between both short- and long-term increases in FV intake and reduced weight loss and reductions in consumption of other food groups. This finding is consistent with previous suggestions that increased FV intake will reduce consumption of other food types and overall energy intake.22 However, additional analyses within the present research suggest that the significant relationship between increased FV consumption and weight loss was achieved through (ie, significantly mediated by) increased self-regulation. Increasing positive (targeted) behaviors associated with eating more FV likely required participants to use more of their self-regulatory skills. As these skills were bolstered and used more effectively, improved eating behaviors emerged. Duplicating this scenario could advance the mutually reinforcing relationship that was documented through the reciprocal interaction between increases in FV intake and self-regulation. Both theory and systematic reviews strongly suggest the importance of self-regulation as a key mediator of weight-loss treatment effects.6,23,24
Although this study was limited by a single-group design with a volunteer sample that may have been especially motivated, it provides new analyses and a basis for further research into the dynamics of treatment-associated changes in diet, self-regulation, and weight loss. Extensions of this research should strengthen measurement instrumentation by incorporating validated food recall surveys that differentiate within food groups (eg, starchy vs nonstarchy vegetables) and accelerometers that can more objectively measure physical activity outputs. After implementing these measures and completing replications involving more ethnicities and sample types, a continued focus on FV intake and self-regulatory skills development can help increase the effectiveness of behavioral weight-loss programs that primarily are unreliable beyond the very short term.25 Ultimately, medical professionals’ support of patients’ weight-management efforts will be enhanced.
Acknowledgment
Kathleen Louden, ELS, of Louden Health Communications provided editorial assistance.
Footnotes
Disclosure Statement
The author(s) have no conflicts of interest to disclose.
References
- 1.Sacks FM, Bray GA, Carey VJ, et al. Comparison of weight-loss diets with different compositions of fat, protein, and carbohydrates. N Engl J Med. 2009 Feb 26;360(9):859–73. doi: 10.1056/NEJMoa0804748. DOI: https://doi.org/10.1056/NEJMoa0804748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Rolls BJ, Ello-Martin JA, Tohill BC. What can intervention studies tell us about the relationship between fruit and vegetable consumption and weight management? Nutr Rev. 2004 Jan;62(1):1–17. doi: 10.1111/j.1753-4887.2004.tb00001.x. DOI: https://doi.org/10.1111/j.1753-4887.2004.tb00001.x. [DOI] [PubMed] [Google Scholar]
- 3.Epstein LH, Gordy CC, Raynor HA, Beddome M, Kilanowski CK, Paluch R. Increasing fruit and vegetable intake and decreasing fat and sugar intake in families at risk for childhood obesity. Obes Res. 2001 Mar;9(3):171–8. doi: 10.1038/oby.2001.18. DOI: https://doi.org/10.1038/oby.2001.18. [DOI] [PubMed] [Google Scholar]
- 4.Luyben PD. Applied behavior analysis: Understanding and changing behavior in the community—a representative review. J Prev Interv Community. 2009;37(3):230–53. doi: 10.1080/10852350902975884. DOI: https://doi.org/10.1080/10852350902975884. [DOI] [PubMed] [Google Scholar]
- 5.Bandura A. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall; 1986. [Google Scholar]
- 6.Bandura A. The primacy of self-regulation in health promotion. Appl Psychol. 2005 Apr;54(2):245–54. DOI: https//doi.org/10.1111/j.1464-0597.2005.00208.x. [Google Scholar]
- 7.Annesi JJ. Exercise predicts long-term weight loss in women with class 1 and class 2 obesity through effects on emotional eating and its correlates. J Phys Act Health. 2017 Nov;23:1–7. doi: 10.1123/jpah.2017-0170. DOI: https://doi.org/10.1123/jpah.2017-0170. [DOI] [PubMed] [Google Scholar]
- 8.Myplate and historical food pyramid resources [Internet] Washington, DC: US Department of Agriculture; updated 2015 Oct 23 [cited 2017 Oct 25]. Available from: http://fnic.nal.usda.gov/dietary-guidance/myplate-and-historical-food-pyramid-resources. [Google Scholar]
- 9.Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. A data-based approach to diet questionnaire design and testing. Am J Epidemiol. 1986 Sep;124(3):453–69. doi: 10.1093/oxfordjournals.aje.a114416. DOI: https://doi.org/10.1093/oxfordjournals.aje.a114416. [DOI] [PubMed] [Google Scholar]
- 10.Mares-Perlman JA, Klein BE, Klein R, Ritter LL, Fisher MR, Freudenheim JL. A diet history questionnaire ranks nutrient intakes in middle-aged and older men and women similarly to multiple food records. J Nutr. 1993 Mar;123(3):489–501. doi: 10.1093/jn/123.3.489. [DOI] [PubMed] [Google Scholar]
- 11.Annesi JJ, Marti CN. Path analysis of cognitive-behavioral exercise treatment-induced changes in psychological factors leading to weight loss. Psychol Health. 2011;26(8):1081–98. doi: 10.1080/08870446.2010.534167. DOI: https://doi.org/10.1080/08870446.2010.534167. [DOI] [PubMed] [Google Scholar]
- 12.Godin G. The Godin-Shephard Leisure-Time Physical Activity Questionnaire. Health & Fitness Journal of Canada. 2011 Mar 1;4(1):18–22. [Google Scholar]
- 13.Jacobs DR, Jr, Ainsworth BE, Hartman TJ, Leon AS. A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med Sci Sports Exerc. 1993 Jan;25(1):81–91. doi: 10.1249/00005768-199301000-00012. [DOI] [PubMed] [Google Scholar]
- 14.Godin G, Shephard RJ. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci. 1985 Sep;10(3):141–6. [PubMed] [Google Scholar]
- 15.Annesi JJ, Unruh JL, Marti CN, Gorjala S, Tennant G. Effects of The Coach Approach intervention on adherence to exercise in obese women: Assessing mediation of social cognitive theory factors. Res Q Exerc Sport. 2011 Mar;82(1):99–108. doi: 10.1080/02701367.2011.10599726. DOI: https://doi.org/10.1080/02701367.2011.10599726. [DOI] [PubMed] [Google Scholar]
- 16.Research-tested intervention programs (RTIPs): Weight Loss For Life [Internet] Bethesda, MD: National Cancer Institute; 2015. [cited 2017 Oct 25]. Available from: https://rtips.cancer.gov/rtips/programDetails.do?programId=24805736. [Google Scholar]
- 17.Schafer JL, Graham JW. Missing data: Our view of the state of the art. Psychol Methods. 2002 Jun;7(2):147–77. DOI: https://doi.org/10.1037/1082-989X.7.2.147. [PubMed] [Google Scholar]
- 18.Cohen J, Cohen P, West SG, Aiken LS. Applied multiple regression/correlation analysis for the behavioral sciences. 3rd ed. Mahwah, NJ: Erlbaum; 2003. [Google Scholar]
- 19.Glymor MM, Weuve J, Berkman LF, Kawachi I, Robins JM. When is baseline adjustment useful in analyses of change? An example with education and cognitive change. Am J Epidemiol. 2005 Aug 1;162(3):267–78. doi: 10.1093/aje/kwi187. DOI: https://doi.org/10.1093/aje/kwi187. [DOI] [PubMed] [Google Scholar]
- 20.Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods. 2008 Aug;40(3):879–91. doi: 10.3758/brm.40.3.879. DOI: https://doi.org/10.3758/BRM.40.3.879. [DOI] [PubMed] [Google Scholar]
- 21.Palmeira AL, Markland DA, Silva MN, et al. Reciprocal effects among changes in weight, body image, and other psychological factors during behavioral obesity treatment: A mediation analysis. In J Behav Nutr Phys Act. 2009 Feb 9;6:9. doi: 10.1186/1479-5868-6-9. DOI: https://doi.org/10.1186/1479-5868-6-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ledoux TA, Hingle MD, Baranowski T. Relationship of fruit and vegetable intake with adiposity: A systematic review. Obes Rev. 2011 May;12(5):e143–50. doi: 10.1111/j.1467-789X.2010.00786.x. DOI: https://doi.org/10.1111/j.1467-789X.2010.00786.x. [DOI] [PubMed] [Google Scholar]
- 23.Teixeira PJ, Carraça EV, Marques MM, et al. Successful behavior change in obesity interventions in adults: A systematic review of self-regulation mediators. BMC Med. 2015 Apr 16;13:84. doi: 10.1186/s12916-015-0323-6. DOI: https://doi.org/10.1186/s12916-015-0323-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Teixeira PJ, Silva MN, Coutinho SR, et al. Mediators of weight loss and weight loss maintenance in middle-aged women. Obesity (Silver Spring) 2010 Apr;18(4):725–35. doi: 10.1038/oby.2009.281. DOI: https://doi.org/10.1038/oby.2009.281. [DOI] [PubMed] [Google Scholar]
- 25.Mann T, Tomiyama AJ, Westling E, Lew AM, Samuels B, Chatman J. Medicare’s search for effective obesity treatments: Diets are not the answer. Am Psychol. 2007 Apr;62(3):220–33. doi: 10.1037/0003-066X.62.3.220. DOI: https://doi.org/10.1037/0003-066X.62.3.220. [DOI] [PubMed] [Google Scholar]