Evaluation of body shape, eating disorders and ...

Evaluation of body shape, eating disorders and weight management related parameters in black female students of rural and urban origins

Marjanne Senekal* Department of Consumer Science, University of Stellenbosch

Nelia Patricia Steyn Department of Human Nutrition, University of the North

Teresa-Ann B Mashego Department of Psychology, University of the North

Johanna Helena Nel Research Administration, University of the North This study examined body shape dtssatlsfactlon, eating disorder and weight management-related parameters as well as assimilation of Western cultural norms regarding body shape In black female students of urban and rural origins. Subjects (n=180; 20 : 4,4 years old) were weighed, their height was measured and they completed the Body Shape Questionnaire, EAT-26, Eating Inventory (restraint scale), Adolescent Self-Concept Scale and a questionnaire concerning weight management behaviours. Main findings Include lower prevalences of overestimation of body weight and body shape dissatisfaction, similar prevalences for dieting and the use of hazardous weight reduction methods, and higher prevalences for overweight, obesity, disordered eating attitudes and behaviours, and dietary restraint among SUbjects than among similar white groups. Furthermore, those with urban origins were more likely to be restrained eaters, to have attempted weight reduction, to aim for weight loss and to fear weight gain. These deta Indicate that there are signs of more realism concerning weight status among black female students. However, there are also signs of assimilation of Western cultural norms concerning body shape, eating attitudes and behaviours and weight management. This diversity In the black student population In South Africa needs to be recognized when planning Interventions to address eating related problems. *To whom correspondence should be addressed: Department of ConsumerScience,University of Stellenbosch,Private Bag X1, Matielend,7602, Stellenbosch. South Africa Tel: (021) 8083394 Fax: (021) 8084804 E-mail: MS30Matles.sun.ac.za

Body shape dissatisfaction, disordered eating. inappropriate weight management practices and clinical eating disorders represent significant health concerns among students on university campuses (Harris. 1995; Mintz & Betze, 1988; Schwitzer, Bergholz, Dore & Salimi, 1998; Senekal, 1994; Sheward. 1995). Although it is generally accepted that the etiology of eating-related disorders is multifactorial, there has been increasing interest in the role that socialand cultural norms and consequent pressures play in either protecting against or causing these problems. It has been argued that the process of normal gender-role socialization leads women to be particularly concerned about their physical appearance as a means of gaining approval from others. Therefore, in Western societies where thinness is indisputably a strived for beauty ideal, young women in the process of establishing their identity are especially vulnerable to dissatisfaction with their shape and the pursuit of thinness through dieting, and consequently the developmentof eating disorders (Cogan, Bhalla, Sefa-Dedeh & Rothblum, 1996;Cooper, 1995; Nasser, 1988). The situation at a university campus can intensify this process, as it presents a community context within which students live, it embeds the social values that shape their choices about weight control, and it provides a context within which weight-related behaviors are enacted (Schulken, Pincario, Sawyer,Jensen & Hoban, 1997). It has been proposed that, compared to white women, nonwesternised and some groups of westernised black women, including college/university-age students, adopt a larger ideal body size, report greater body image satisfaction, are more accepting of being overweight, experience less pressure especially from men - to be thin, do not necessarily equate over-fatness with being unattractive. and are therefore less likely to aspire to thinness (Abrams, Allen & Gray, 1993;

Cogan et al., 1996; Dolan, 1991; Greenberg & La Porte. 1996; Melnyk & Weinstein, 1994; Striegel-Moore, Schreiber, Pike, Wilfley & Rodin, 1995; Toriola, Dolan, Evans & Adetimole, 1996; Wilfley, Schreiber, Pike. Striegel-Moore, Wright & Rodin 1996). In South Africa it has also been reported that black women view obesity as a normal state of health (Kruger. van Aardt, Walker & Bosman. 1994). This situation can partly be explained by the fact that black women are not under the same cultural pressure as white women to value thinness (Dolan, 199I) and thus to diet for the sake of appearance (Melnyk & Weinstein, 1994). because in several non-Western cultures fatness and obesity symbolise beauty (Nasser. 1988) and wealth (Cogan et al., 1996). These obesity-tolerant attitudes could protect black female adolescents and women from developing eating disorders. However. Wiltley et al. (1996) have found that black women are not immune to mainstream body image dissatisfaction and eating disturbances. This is supported by the fact that when persons of different cultural backgrounds internalize the Western norms (acculturation) of thinness as the ideal, a greater degree of disordered eating is observed (Akan & Grillo, 1995; Cogan et al., 1996; Lee. Leung, Lee Yu & Yeung. 1996;O'Dea. 1995). This acculturation process can occur in ethnic minorities in Western countries and also during the westernisation aodlor urbanisation of total populations, especially through the increased exposure to Western advertising, marketing, electronic media. entertainmentand fashions (O'Dea, 1995). Schwitzer et al. (1998) contend that there is a need for intervention strategies that target the increasing incidence of eating-related problems in Westernised societies. In westernising groups it is especially important to identify and

S. Afr. J. Psycho!. 2001,31

(1) Downloaded from sap.sagepub.com at Stellenbosch University on March 30, 2016

45

strengthen identified protective factors, e.g, greater body shape satisfaction, in such interventions. There is a paucity of data on body shape perception. eating disorders and weight management-related issues among black South African women. Because many of these women have recent or current ties with their non-Western cultural heritage, which favors a larger body size, it is speculated that the current processof westernisation and urbanisation which is taking place in this country could contribute to a rise in body image and eating-related problems. An example of such a group in transition is the female students at the University of the North (UNIN), which is a historically black university situated in the Northern Province of South Africa. The province, which is the main catchment area for the UNIN, has a population of 4,1 million, of whom 76,3% are black and of whom 88% live in rural areas (Central Statistics Service. 1996). Therefore. a large percentageof the students at the UNIN can be expected to have recent rural origins. which poses a unique opportunity to explore body shape, eating disorder and weight management related questions among black female students with either rural or urban origins. The aims of this study. which formed part of the baseline evaluations for the longitudinal follow-up of the health and weight status of female students at the UNIN, were to determine I) the extent of body shape dissatisfaction. disordered eating attitudes and behaviors, and dietary restraint among black female students, 2) the relationship between these three measures as well as their relationship with weight history, weight management practices and self-concept. 3) the role of rural and urban origins in the determination thereof and 4) formulation of conclusions concerning the assimilation of Westerncultural norms regarding body shape by black students of rural and urban origins. Detailed data on the baseline health and weight status of the students who participated in this study have been published elsewhere (Steyn, Senekal, Brits. Alberts. Mashego & Nel. 2000).

Methods SUbjects and procedures No generally accepted standard definition for rural and urban has been applied in South Africa.Therefore. for the purposes of this study. students who came from farms and small villages were classified as having rural origins, and those from towns and cities as having urban origins. This classification was done in collaboration with the Department of Geography at the UNIN. A total of 772 female students registered as first-years at the UNIN in 1994. Four hundred and thirty one of these students attended the first-year orientation programme during their first week at the UNIN. During this week the students completed a battery of questionnaires, including those reported in this paper. Further appointments were made with each student for anthropometric and other measurements. As all students did not arrive for these appointments, a complete set of anthropometric and appropriatequestionnairedata was obtained for 180 students. Permissionto conduct the study was obtained from the UNIN's ethics committee, and informed consent was obtainedfrom each participant.

Instruments Concerns aboutbody shape The Body Shape Questionnaire (BSQ) developed by Cooper, Taylor, Cooper and Fairburn (1987) is a 34-item questionnaire which measures female concerns with body weight and shape. The BSQ has been validated in clinical and non-clinical populations varying in age and cultural backgrounds (Cooper et al., 1987; Lacey & Evans, 1990; Mumford, Whitehouse & Platts, 1991; Rosen, Jones, Ramirez & Waxman, 1996). Internal reliability (Cronbach's alpha) of the BSQ for our sample was very satisfactory at 0.93 (95% CI: 0.92 - 0.95), which compares well with the 0,97 found for Afro-Caribbean women (Dolan et al., 1990). Factor analysis showed that one main factor accounted for a large part (32,4%) of the sample variance, as was also reported by Dolan et at. (1990). Higher BSQ scores reflect greater body shape dissatisfaction, but no clinical cut-off scores have been described. For the purposes of this study a score of ~ 129 was used to identify probable cases of bulimia nervosaI obese dieters I body image therapy patients. This was based on the mean score of 129,3 ± 17 found for probablecases of bulimia by Cooper et at. (1987), and the mean scores of 129,9 ± 29,0 found for body image therapy patients and 123.1 ± 27.9 for obese dieters by Rosen et al. (1996). A lower cut-off score of < H2. based on 129-17 (standard deviation for probable bulimia nervosa cases), was used to denote non-cases of the abovementioned.

Eating attitudes and behaviors The Eating Attitudes Test (EAT-26) developed by Garner. Olmsted. Bohr and Garfinkel (1982) is a 26-item questionnaire which measures a broad range of eating attitudes and behaviours associated with eating disorders. Higher scores reflect more disordered attitudes and behaviours, but are not necessarily diagnostic of an eating disorder. The EAT-26 has been validated in different cultural and age groups (Evans, Dolan & Toriola, 1997; Garner et at.• 1982; Johnson-Sabine. Wood. Patten, Mann & Wakeling. 1988; Mumford et at.• 1991) The internal reliability (Cronbach's a. = 0,62; 95% CI: 0,54 0,7) for our sample was not as good as was reported for white British women (0,87) (Evans et al., 1997), Polish adolescents (0,84) (Wlodarczyk-Bisaga & Dolan,I996) or Nigerian students in Nigeria (0.70) (Evans et al., 1997). but was comparable to that found for adult Polish women (0.65) (Wlodarczyk-Bisaga & Dolan, 1996) and higher than found for adult women in Nigeria (0,54) (Evans et al., 1997). As the confidence interval for our sample did not embrace zero, it was assumed that the EAT scores are reliable and can be used as long as the results are interpreted with caution. Factor analysis showed that the EAT-26 had a first factor accounting for 16.2% of the variance in our sample. which is comparable with the 20.6% found for the Afro-Caribbean group (Dolan et al., 1990). Loading on this first factor was similar in our sample to that reported by Garner et at. (1982) for 9 out of 13 items. A cut-off score of ~ 21 and not ~ 20, as was suggested by Garner et at. (1982), was used to identify students with disordered eating attitudes and behaviours. This was based on the findings by Mann, Wakeling. Wood, Monck, Dobbs & Szmukler (1983) that the higher threshold was associated with the highest sensitivity and specificity in unselectedgroups in the general population.

Anthropometric measurements The student's weight and height were taken by trained and standardized fJeldworkers. The Body Mass Index (BMI = Weight (kg)/height (m)2) was calculated for each student. The cut-off points recommended by Jequier (1987), which comply with the latest recommendations of the United States National Heart, Lung and Blood Institute (0' Arrigo. 1998), were used to categorize the students' weight status. These cut-off points are as follows: underweight BMI < 20; normal weight: BMI ~ 20, < 2S; overweight:BMI ~ 2S. < 30; and obese: BMI ~ 30.

46

Dietaryrestraint The 21-item restraint scale of the Three Factor Eating Eating Inventory) developed by Questionnaire (TFEQ Stunkard and Messick (1985) was used to classify the respondents as restrained or unrestrained eaters. The total questionnaire's internal consistency and test-retest reliability have been found to be acceptable (l.aessle, Tuschl, Kotthaus & Pike. 1989). The internal reliabilityof the restraint scale for our sample (Cronbach's a. s O,S6; 95% 0: 0,46 - 0.65) was

S. Aft. J. PsychoI. 2001,31 (I) Downloaded from sap.sagepub.com at Stellenbosch University on March 30, 2016

=

moderate. However, as the confidence intervals did not embrace zero, it was assumed that the restraint scores are reliable and can be used as long as the results are interpreted with caution. The total TFEQ (51 items) had a first factor accounting for II % of the variance in our sample. The cut-off scores for the identification of restrained eaters suggested by Tuschl, Platte, Laessle, Stichler and Pike (1990), based on the results of a large sample of healthy women, were used. A score ~ 10 denotes a restrained eater and S 3 an unrestrained eater. Self-concept The lOO-item Adolescent Self-Concept Scale (ASCS) developed by Vrey and Venter (1983) was used to determine the self-concept of the respondents. The scale consists of six subsections including physical self, personal self, the self in relation to family and relatives, the self in relation to the social community, the self in relation to values and, finally selfcriticism Internal reliability (Cronbach's a) of the ASCS for our sample was satisfactory at 0,82 (95% CI: 0,78 - 0,86). Factor analysis showed that the first six factors explained 23,2% of the variance in the group. Suggested cut-off scores for the ASCS are as follows: low self-concept S; 62, medium selfconcept 63-78, and high self-concept ~ 79 (Vrey & Venter, 1983). Additional Questions To investigate specific weight management-related aspects, a short questionnaire was compiled, which included the following: I) self-reported weight; 2) weight reduction attempts over the past two years; 3) type of weight reduction methods used (16 listed methods); 4) weight pattern over the past two years (steadily increasing, steadily decre~sing, stable, cycler~; 5) think of current weight (under-weIght, normal weight, overweight, obese); 6) current weight goals (satisfied, lose 1-3 kg, lose ~ 4 kg, weight gain); and 7) fear of weight gain (very

scared, slightly scared, not scared at all). A weight cycler was defined as an individual who had lost and gained either 1-3 kg or more than 3 kg more than twice over the past two years. A Diet Score was derived from Question 4. Less acceptable weight reduction methods were afforded a higher score, with . higher Diet Scores thus denoting less acceptable weight reduction practices.

Statistical methods Data were analysed using the Statistical Analysis System (SAS) software (SAS Institute Inc., 1989). Means and standard deviation (SO) were determined for all the continuous variables. A Pearson's product correlation matrix was constructed to identify significant relationships between these variables. Subsequently a stepwise multiple regression model was constructed to predict the value of the dependent variable, BSQ, from a group of independent variables. Contingency tables were constructed to investigate the relationship between categorical variables, using BMI categories, origin (ruraVurban), BSQ categories, and dietary restraint categories as classification variables. The Chi-Square statistic was applied to determine the homogeneity of profiles, with a p-value of 5% indicating significant difference. Where necessary, categories were collapsed to improve statistical power. However, this was not possible in all cases as important information would have been lost. Therefore, in some tables one or more cells had an expected count of less than five, which could render the Chisquare an invalid test. These p-values are reported, but should be interpreted'cautiously.

Results Rural and urban origins Students with rural origins constituted 64,8% of the sample and those with urban origins 35,2%.

Table 1. Column percentages of BMI categories by specified variables BMI CATEGORIES

SPECIFIEDVARIABLE

Origin

20-24,9

25-29,9

:e30

,

n"=45

n=94

n=3O

n-9

! Rural

62,2

67,0

66,7

55,6

iI

37,8

33,0

33,3

44,4

I

n=46

n=94

n=30

n-9

!

17,4

2,1

3,3

0.0

i Nonnal

80,4

71,3

50,0

11,1

2,2

26,6

43,388,9

88,9

i Obese

0,0

0,0

3,4

0,0

n=43

n=83

n-28

n-8

i Increasing I ! Decraaslng

13,3

27,2

40,0

55,6

17,8

12,0

6,7

0,0

! StabI8

I

48,9

19.5

16,6

11,1

II

20,0

41,3

36,7

33,3

!

n=44

n=94

n=30

n=90

i sallsfled

68,2

36,2

26.7

0,0

2,3

30,9

6,7

0,0

4,5

27,6

66,6

100,0

25,0

5,3

0,0

0,0

n=44

n=93

n=30

n-9

45,4

59,1

56,7

44,4

18,2

24,7

36,7

44,4

36,4

16.1

6,6

11.1

I

Urban

i

Think of own Weight

!

UndelW9lght

! Overweight

Past weightpattem

i i

I

Cycler

I

Current weight goals

CHI-SQUARE