Healthy dietary practices among rural and semi-urban Blacks in the Northern Province of South Africa

The aim of the study was to investigate five healthy dietary behaviours in a sample of rural and semi-urban South Africans. The sample consisted of 200 adults, 100 from an semi-urban area (Mankweng) and 100 from a rural area (Tiberius) in the central region of the Northern Province of South Africa. The two geographically different communities were chosen by convenience and the participants in the two communities were choosen by cluster sampling. Results indicate that about a third (30% in semi-urban and 34% in rural) of the study sample are overweight and 18% are obese. A moderately high prevalence of six simple healthy dietary practices was found. However, there was a very low preva­ lence rate of eating fruits daily among both semi-urban (10%) and rural dwellers (9%). Semi-urban dwellers showed significantly higher healthy diet behaviour than rural dwellers in regard to avoiding fat, trying to eat fiber, limiting red meat, and limiting salt. Men reported more than women that they tried to eat fiber and they had more often breakfast everyday. Being semi-urban and female were significantly associated with the healthy dietary index, whereas age, BMI, educational level and marital status were not. The results give insight into dietary health behaviour practices and the factors that influence them, which have practical implications for dietary health promotion.


Introduction
The link between diet and chronic diseases such as cancer and cardiovascular disorders has been well recognised world wide (WHO, 1990: 10ff.).Overweight and high-serum cho lesterol levels, hypertension, and osteoporosis (i.e., decreased bone mass) increase the risk of cardiovascular disorders, stroke, and bone fracture, respectively (Edelman & Mandle, 1998: 155f.).As assessed by a national demographic and health survey, approximately 28% of men and 55% of women were over weight or obese in South Africa.In men, overweight and obes ity occur more frequently in the urban setting than in the non-urban setting.White men and the most educated men are the most overweight or obese of all men.For women, the overweight patterns do not differ much between urban and non-urban, although urban women tended to be more obese.
Women with the lowest level of education seemed to be the most obese, although this may be a function of age.African urban women have the highest rate of obesity, while Asian women have the lowest rate.Obesity has been found in a number of studies in all ethnic groups to predict the develop ment of hypertension and diabetes.The findings call for the control of obesity as a focus of community-based intervention programmes in South Africa (Medical Research Council, 1998: 15ff.).Similarly to the US, many urban South Africans consume too many calories and too much fat (especially saturated fat), cho lesterol, and sodium.Large meals consisting of high-kilojoule foods, and between-meal snacks, are the two dietary habits which are most often responsible for obesity in individuals and families.Ignorance of kilojoule value of various foods is to some extent a contributory factor, so is poverty, for fami lies with limited incomes have to buy cheap foods and as these usually consist of carbohydrates they are more likely to lead to overweight than proteins.They also consume insufficient complex carbohydrates and fiber.Such diets are one cause of high rates of obesity and diseases such as heart disease, high blood pressure, stroke, diabetes, and some forms of cancer.Recommendations to restrict salt and fat intake and increase complex carbohydrate and fiber consumption are central ten ets in public health nutrition guidelines (De Haan 1993: 15f., Edelman & Mandle, 1998: 156f.).In a recently nationally representative survey the Medical Research Council (1998: 2 If.) found that among persons above 15 years 11% of men and 13% of women were found to either have a blood pressure above 160/95mmHg or were taking appropriate medication to lower their blood pressure.A cal culation based on these prevalence rates and the census fig ures published for the South African population 15 years and older leads to an estimate of about 3.3 million hypertensive people in the country.
While the full etiology of any of these chronic diseases have yet to be understood, behavioural factors such as tobacco use, exercise, diet, alcohol consumption and preventive health checks are strongly implicated as risk factors (Steptoe & Wardle 1992: 486).Overconsumption of dietary fats, sugar and salt, and lack of fiber in the diet may lead to a number of chronic diseases including coronary heart disease and some cancers (Furie & Steyn 1995: 5f.).Steyn, Langenhoven, Joubert et al. (1990: 63 )  insights on what dietary behaviour may be related to over weight or obesity among a sample of rural and semi-urban Black South Africans.Consequently, simple dietary messages could be developed and included in health promotion pro grammes.It would also be of interest to know whether some of the dietary behaviours differ between rural and semi-urban dwellers.It may be antizipated that semi-urban dwellers may have adapted due to transition and urban life style different dietary behaviours.Differences between rural and semi-ur ban dwellers regarding dietary patterns could then also be addressed in health promotion programmes.

Research methodology Design
The study was a cross-sectional interview-administered house hold survey with an internationally validated questionnaire viewed by a trained postgraduate research assistant after per mission was sought from the local authorities and informed consent was obtained from the participants.A convenience sample of two geographically different communities in the central region of the Northern Province was chosen: one ru ral (Tiberius, 60 km west of Pietersburg) and one semi-urban community (Mankweng) representing different geographical and socioeconomic characteristics.Mankweng consists of six units and a population of 11212, Tiberius consists of 2 sec tions with a total population of 2000 using census data from 1996 (Statistics South Africa, 1998).From the six units in Mankweng one unit and from the two sections in Tiberius one section was chosen at random for the study.In the se lected clusters of each area one adult was interviewed in all households.One adult was chosen at random by using the birth data method (the adult who had birthday nearest to the interview date, was selected for interview).

Sample and procedure
The sample consisted of 200 adults, 100 from an semi-urban area (Mankweng) and 100 from a rural area (Tiberius) in the Northern Province of South Africa.The adults were inter-

Research instrument
The by only a limited number of items.It was designed to meas ure European health and behaviour, and was later extended to include developing countries as well (e.g., Peltzer, 2000: 46).The following dietary behaviours were assessed: (a) fre quency of consumption of red meat (daily, 2-3 times a week, once a week, less than once a week, never); b) frequency of consumption of fruit (daily, 2-3 times a week, once a week, less than once a week, never); c) addition of salt to food (usu ally, sometimes, occasionally, never); d) trying to avoid fat and cholesterol (yes, no); and e) trying to eat fiber (yes, no).The fiber and fat items were each followed by an open-ended question asking what foods the individual either avoided or ate.Ratings were also made of perceived body size (rated from very fat to very thin).Weight loss practices were recorded by using two items: "Are you trying to lose weight?" and "Are you on a diet?"The questionnaire was pilot tested on 30 adults and after 3 weeks re-administered.Test-retest reliability of .81 was found.Cronbach alpha and split-half reliability coefficient for the dietary questionnaire were .75 and .71,respectively, for this sample.
In relation to fruit consumption, the healthy practice was de fined as eating fruit daily (WHO, 1990: 25f.).For adding salt to the food, the unhealthy practice was defined as responding usually to this question, whereas any of the other responses (sometimes, occasionally, or never) were categorized as healthy.For red meat consumption the healthy practice was defined as eating meat less often than daily (including never).To produce a more quantifiable index and to simplify the pres entation, a healthy diet index score was calculated including (1) avoid fat, (2) eat fiber, (3) eat fruit daily, (4) limit red meat, and (5) limit salt.Data on age, marital status, height, and weight were also col lected.Self-reported height and weight was used to calculate body mass index [weight (kg)/height (m)2].

Data analysis
Using the SPSS version 10.0 the Pearson's product-moment correlation coefficients were calculated to evaluate relation ships between subscales.The Chi-square test was used for analysis of proportions.

Results
Table 1 indicates the sociodemographic characteristics of the participants.There were 43 men in the semi-urban area and 52 men in the village, and there were 57 women in the semi-urban area and 48 in the village.The ages of the participants seemed to have been evenly distributed between semi-urban and rural as well as between the two age groups of 18-34 and 35-64 years.The majority of the villagers had (some) primary education (one to seven years of schooling) whereas the majority of the semiurban dwellers had (some) secondary education (8 to 12 years of schooling).
Table 2 indicates the Body Mass Index and body weight per ception of the participants.About a third (30% in semi-urban and 34% in rural) of the study sample have overweight and 18% had obesity.The majority of the semi-urban dwellers (45%) felt that they had gained weight and 44% of the villagers felt that their weight had remained the same.There were no significant differences between semi-urban (X2=1.76;ns) and rural dwellers (X2=7.20;ns) concerning body mass status.Half of the semiurban sample was trying to lose weight and 37% was dieting Semi-urban dwellers showed significantly higher healthy diet behaviour than rural dwellers in regard to avoiding fat, try ing to eat fiber, limiting red meat, and limiting salt.Men reported more than women that they tried to eat fiber and they had more often breakfast everyday.There was a very low prevalence rate of eating fruits daily among both semiurban (10%) and rural dwellers (9%).
The interrelationship between dietary practices is illustrated with the Pearson correlation matrix in Table 4. Significant associations were only found between eating fiber and avoiding fat, and limiting red meat and having breakfast almost everyday.to lose weight; semi-urban dwellers were significantly more than villagers trying and dieting to lose weight, and also felt more overweight.
The prevalence of the individual healthy dietary practices for semi-urban and rural dwellers as well as men and women is shown in Table 3.
(Body Mass Index).Generally, dieting and trying to loose weight seemed to pro portionally decrease with a higher BMI.There were surpris ingly no significant associations between body weight status and individual dietary healthy practices.

Curationis May 2002
Dietary practices in relation to dieting status are shown in Table 6.Pearson Chi-Square showed that dieters were more likely to be avoiding fat and limiting red meat.
ban and rural Blacks in the Northern Province of South Af rica.However, there was a very low prevalence rate of eating fruits daily among both semi-urban (10%) and rural dwell ers (9%), which is a cause for concern.Generally semi-urban dwellers and men showed more healthy diet practices than rural dwellers and women did.The latter

Dietary behaviour, body mass index and sociodemographic correlates
To produce a more quantifyable index, and to simplify the presentation, a healthy diet index score was calculated in cluding (1) avoid fat, (2) eat fiber, (3) eat fruit daily, (4) limit red meat, (5) limit salt, and (6) having breakfast almost eve ryday.Findings are presented in Table 7.
Being semi-urban and male were significantly associated with the healthy dietary index, whereas age, BMI.educational level and marital status were not.

Discussion
The study found a moderately high prevalence of six simple healthy dietary practices among this population of semi-ur-finding is contrary to some other studies where women were found to be practicing healthier dietary practices than men (Wardle et al.1997: 445f.).
Although more semi-urban dwellers (37%) were dieting than rural dwellers (17%) were, it only had a significant influence on two healthy dietary behaviours: avoiding fat and limiting red meat.This was also found among European young adults (Wardle et al. 1997: 448).Furthermore, this study identified being semi-urban, female and having a lower education as associated with healthy di etary behaviour.Being semi-urban could be associated with higher socioeconomic status but not with lower educational level.Therefore, this result is mixed.Most other studies seem to indicate that higher socioeconomic status is associated with healthy dietary behaviour (Smith & Owen 1992: 735).For instance in Australia higher dietary fat intakes and lower di etary fiber intakes have been found among lower socioeco-45 Curationis May 2002 nomic groups compared to higher status groups (Smith & Baghurst 1992: 409).Steptoe and Wardle (1999: 391) found among a representative sample in Britain that fibre consump tion was inversely associated with socioeconomic status and that the higher education group ate more fiber, fruit, vegeta bles and cereals than the low education group.The rate of overweight and obesity status in this sample (32% and 18% respectively) is slightly lower to what was found by Walker (1995: 95f.) (44% obesity among female Blacks and Health and Behaviour Survey (developed by Steptoe & Wardle 1996:49-73) was designed as a broad survey of healthrelated behaviours and beliefs, each individual area was tapped 42 Curationis May 2002 The relationship between dietary factors and serum cholesterol values in the coloured population of the Cape Peninsula.South African Medical Journal.78: 63-67.WALKER, ARP 1995: Epidemiology and health implications of obesity in Southern Africa (In: Furie, J & Steyn, K Eds.1995: Chronic diseases of life style in South Africa.Cape Town: MRC Technical Re ports, pp 73-86).W ARDLE, J; STEPTO E, A; BELLISLE , F; DAVOU, B; RESCHKE, K & LAPPALAINEN, R 1997: Healthy dietary practices among European 2002 students.Health Psychology.16: 443-450.WORLD HEALTH ORGANISATION 1990: Diet, nutri tion, and the prevention of chronic diseases.Geneva: WHO.

Table 5
indicates the dietary practice in relation to the BMI

May EDELMAN, CL & MANDLE, CL 1998:
Health promotion throughout the lifespan.St. Louis: Mosby.Associations of social status and health-related beliefs with dietary fat and fiber densities.Preventive Medicine.21: 735 745.Cognitive pre dictors of health behaviour in contrasting regions of Europe.British Journal of Clinical Psychology.31: 485-502.The European health and behaviour survey: the development of an international study in health psychology.Psychol ogy and Health.11: 49-73.