Strategies for eating and body change among Brazilian women and men

Objectives: Our study was conducted to adapt the Body Image and Body Change Inventory (BIBCI) for Portuguese; to evaluate the BIBCI’s psychometric properties in samples of university students; to calculate the prevalence of strategies for eating and body change among students; and to evaluate the impact of demographic, social, and anthropometric characteristics on the BIBCI subscales. Methods: 798 students (women = 63%) answered the Portuguese translation of the BIBCI and answered a demographic questionnaire. All analyses were performed separately for women and men. The BIBCI’s psychometric properties were estimated using confirmatory factor analysis. Mean scores were calculated for each BIBCI subscale. A multivariate regression model was tested to evaluate the impact of demographic, social, and anthropometric characteristics on mean BIBCI subscale scores. Results: The psychometric properties of the BIBCI were adequate in the samples analyzed. The BIBCI subscales scores did not differ according to sex. According to the cut-off points adopted, most of the students were classified in the very low category of the BIBCI subscales. For women, characteristics such as self-reported eating assessment, economic class, physical activity level, and work were significant. For men, only physical activity level was significant. Conclusion: The Portuguese translation of the BIBCI was presented and its psychometric properties were found to be adequate in the samples analyzed. The models identified significant characteristics that can be used in intervention protocols for preventing inappropriate behaviors in relation to body image and eating.


Introduction
Physical fitness has become one of the main reasons people effect changes in eating habits and physical activities. These changes are generally made for aesthetic reasons, without concern for physical and mental health. Researchers have sought to understand individuals' relationships with their bodies in order to develop health promotion actions. 1,2 To investigate these relationships, it is relevant to identify the mental representation that individuals construct in relation to their bodies, which has been conceptualized by Cash This instrument was developed in two parts, the first (BI) referring to body image 4 and the second (BCI), to eating and body change. 5 McCabe and Ricciardelli 5 presented the composition of both parts (BI and BCI) with different subscales for each.
The BIBCI was originally developed in the English language to evaluate satisfaction with body image, importance of body image and eating and body change strategies in adolescents. The items were developed for application in girls and boys, and the subscales have been widely used in international contexts, [6][7][8][9][10][11] but not in Portuguese-speaking countries. 12,13 In Brazil, Conti et al. 12 has presented a Portuguese version of the BIBCI, but only the second part of the instrument (BCI) was translated. Although the BIBCI was developed using a sample of adolescents, the items do not appear to be exclusive to this population. Another point that merits highlighting is related to evaluation of the psychometric properties of BIBCI, which is rarely reported in the literature, and the related lack of consensus among researchers on the analytical strategy for estimating BIBCI subscale scores (i.e., some studies use means, and others use sums). The authors of the BIBCI did not In previous studies, 11,[14][15][16][17]    Previous studies were consulted with regard to estimation of the final scores of each BIBCI subscale.
Most of these studies calculated scores using sums, which can be problematic (for example, if items are excluded, the classification ranges will be prejudiced).
In this scenario, the mean was used to calculate BIBCI scores for women and men. The mean scores of the subscales were compared between men and women using analysis of variance (ANOVA) and a 5% level of significance. We used response scale percentiles (P25, Next, the significance of hypothetical causal paths (β) was evaluated using a 5% level of significance, calculated by the z-test at the critical ratios. Models were refined using the stepwise method to identify significant characteristics. These analyses were conducted using MPLUS (v.7.2).         seek body change. 38 No prior study has reported results similar to ours in relation to the significant paths found between work and strategies for body change. However, we believe that the increased availability of spare time for students who do not work may have influenced the greater frequency of use of strategies to increase muscle tone. Meanwhile, the significant path between work and binge eating has been discussed in recent studies 39 that reported work as an activity that has relevance for people's eating patterns; students who work may have little time to prepare a more balanced diet, resulting in changes in food intake. Moreover, physical activity level was significant for BIBCI subscales in women and men, and this relationship is commonly reported in previous studies. 2,40 According to those studies, individuals who practice physical activities have a different perception of body image and often change their activities according to their interpretations with relation to the body. The results highlight the importance that these characteristics can have for eating and body change strategies. There is therefore a need to develop protocols that include these characteristics for preventive, educational, and/or curative treatments.

Results
In general, we hope to have increased the possibilities for use of the BIBCI in both scientific and clinical contexts by producing the Portuguese version and defining the rating ranges. This instrument showed satisfactory psychometric properties in women and men, revealing its suitability for university students, who comprise a population considered vulnerable to issues related to body image. The characteristics found to be significant can be used in protocols for intervention or prevention of problems related to eating disorders and body dissatisfaction.
Some limitations should be mentioned. The nonprobabilistic sample selection means that the results cannot be generalized to all Brazilian university students, but may help better target future studies and clinical protocols. We also highlight as a limitation the cross-sectional design, which does not allow inference of cause-and-effect relationships between the characteristics evaluated in the regression models.
However, the results found may help future studies in terms of the choice of characteristics to include in protocols. Therefore, to overcome the limitations of our study, we suggest that further studies be carried out with the BIBCI in different samples and contexts.