Filetype PDF | Posted on 12 Jan 2023 | 2 years ago
Authentication
digital resources
107x Filetype PDF File size 0.19 MB Source: core.ac.uk
View metadata, citation and similar papers at core.ac.uk brought to you by CORE
provided by eCommons@AKU
eCommons@AKU
Faculty of Health Sciences, East Africa Faculty of Health Sciences, East Africa
October 2013
Development of a quantitative food frequency
questionnaire for use among rural South Africans in
KwaZulu Natal
T. Sheehy
University College Cork
F. Kolahdooz
University of Alberta
T. L. Mtshali
Izulu Orphan Projects
Tashmin Khamis
Aga Khan University, tashmin.khamis@aku.edu
S. Sharma
University of Alberta
Follow this and additional works at: https://ecommons.aku.edu/eastafrica_fhs_fhs
Part of the Public Health Commons
Recommended Citation
Sheehy, T., Kolahdooz, F., Mtshali, T., Khamis, T., Sharma, S. (2013). Development of a quantitative food frequency questionnaire for
use among rural South Africans in KwaZulu Natal.Journal of Human Nutrition and Dietetics, 27(5), 443-449.
Available at: https://ecommons.aku.edu/eastafrica_fhs_fhs/1
Journal of Human Nutrition and Dietetics
PUBLIC HEALTH NUTRITION AND EPIDEMIOLOGY
Development of a quantitative food frequency
questionnaire for use among rural South Africans in
KwaZulu-Natal
T. Sheehy,* F. Kolahdooz,† T. L. Mtshali,‡ T. Khamis§ & S. Sharma†
*School of Food and Nutritional Sciences, University College Cork, Cork, Republic of Ireland
†Department of Medicine, Aboriginal & Global Health Research Group, University of Alberta, Edmonton, AB, Canada
‡Izulu Orphan Projects, Empangeni, South Africa
§Faculty of Health Sciences, Aga Khan University, Nairobi, Kenya
Keywords Abstract
dietary assessment, rural, South Africa.
Background: South Africa is experiencing a dietary and lifestyle transition
Correspondence as well as increased rates of noncommunicable chronic diseases. Limited
S. Sharma, Aboriginal & Global Health Research information is available on the diets of rural populations. The present study
Group, Department of Medicine, University of aimed to characterise the diets of men and women from rural KwaZulu-
Alberta, 8303 – 112 Street, 5–10 University Natal (KZN) and develop a quantitative food-frequency questionnaire
Terrace, Edmonton, AB, T6G 2T4, Canada. (QFFQ) specific for this population.
Tel.: +1 780 492 3214
Fax: +1 780 492 3018 Methods: A cross-sectional study was carried out by collecting single 24-h
E-mail: gita.sharma@ualberta.ca dietary recalls from 81 adults and developing a QFFQ in Empangeni, KZN,
South Africa.
Howtocite this article Results: The diet of this population was limited in variety, high in plant-
Sheehy T., Kolahdooz F., Mtshali T.L., Khamis T. based foods (especially cereals and beans), and low in animal products, veg-
& Sharma S. (2014) Development of a etables and fruits. Amaize meal staple (Phutu) was consumed by over 80%
quantitative food frequency questionnaire for use of subjects and accounted for almost 45% of energy intake, as well as mak-
among rural South Africans in KwaZulu-Natal. ing an important contribution to fat and protein intake. Most of the protein
J Hum Nutr Diet. 27, 443–449 consumed by the study population was plant-based protein, with almost
doi:10.1111/jhn.12166
40% being obtained from the consumption of phutu and beans. A culturally
appropriate QFFQ was developed that includes 71 food and drink items, of
which 16 are composite dishes unique to this population.
Conclusions: Once validated, this QFFQ can be used to monitor diet-dis-
ease associations, evaluate nutritional interventions and investigate dietary
changes in this population.
Introduction increased industrialisation, urbanisation, acculturation,
etc. (Kruger et al., 2005). This has resulted in a shift to a
South Africa is a multicultural country of over 50 million less prudent, ‘Western’ dietary pattern, which is associ-
people, consisting of Black African, Asian, White and ated with an increased risk of noncommunicable chronic
Coloured individuals. Black African peoples make up diseases (Popkin 2006). Between 1940 and 1990, fat
almost 80% of the total population (Statistics South intake in black people living in urban settings increased
Africa, 2011) and include a diversity of tribal groups, from 16% to 26% of total energy, whereas energy from
such as the Zulu, Tswana, Sotho, Pedi, Venda and Ndeb- carbohydrate fell from 69% to 62% (Bourne et al., 2002).
ele. Over recent decades, populations in South Africa Results from the Transition, Health and Urbanisation in
have been undergoing a dietary and lifestyle transition South Africans (THUSA) study (MacIntyre et al., 2002;
caused by both political changes within the country itself, Kruger et al., 2005) across five strata of urbanisation
as well as the broader demographic and socio-economic (from rural to upper-class urban residential areas)
trends that are happening at a global level related to showed that energy from carbohydrate decreased from
ª2013The British Dietetic Association Ltd. 443
Development of a quantitative food frequency questionnaire T. Sheehy et al.
67% to 56% and energy from fat increased from 23% to quantitative food frequency questionnaire (QFFQ) for
31% between the lowest and highest strata. In addition, use specifically with this population.
more than half of all subjects, regardless of residence, had
intakes of vitamin A, folate, ascorbic acid, calcium, iron Materials and methods
and zinc that were less than 67% of the recommended
dietary allowance for their sex and age groups, with rural Study setting
subjects having significantly lower intakes of most mi- This cross-sectional study took place in six rural villages
cronutrients than urban subjects (Kruger et al., 2005). in Empangeni, KZN, South Africa. Villages were ran-
The introduction of national mandatory fortification of domly selected from a map of Empangeni, which is
commonly consumed staple foods (maize meal and wheat located 160 km north of Durban, the capital of KZN, and
flour) appears to have had a significant positive impact is within the local municipality of Umhlatuze.
on overall micronutrient adequacy (Steyn et al., 2008)
but, for some groups, especially the poor and among Subjects
rural populations, consuming a healthy diet remains
problematic because of limited access as well as the high Men and women (n = 81) from the six villages were
cost of healthier food choices (Temple et al., 2011). selected to participate in the present study. Within each
There is convincing evidence that high intakes of village, a fieldworker was asked to select people represent-
energy-dense, micronutrient-poor foods increase the risk ing the geographical area and to include people of both
of obesity and noncommunicable chronic diseases (World sexes and from all age groups (19–79 years). Fieldworkers
Health Organization, 2003). In South Africa, the burden selected people who lived close to the food stores, as well
of noncommunicable chronic diseases is increasing in as those who lived at long distances from stores. The
both urban and rural areas, especially in poor people liv- fieldworker was informed of the need for the sample to
ing in urban settings (Mayosi et al., 2009). This burden is be representative of the population residing in the villages
expected to increase substantially over coming decades and was asked not to purposely select relatives and
unless concerted prevention and control measures are friends. The majority of the villages had a fairly small
taken (Abegunde et al., 2007). Because diet is a major population with very similar socio-economic status and
potentially modifiable risk factor for noncommunicable access to food. Respondents were asked to attend a local
chronic disease, accurate assessment of dietary intake in centre for the interviews to be undertaken. The study was
at-risk South African populations is essential. approved by the University of Alberta Internal Review
Dietary intake can be determined by a variety of Board and University of KwaZulu-Natal. Written
methods, including 24-h recalls, food frequency ques- informed consent was obtained from all participants.
tionnaires (FFQs), food diaries and measurement of die-
tary diversity (Cade et al., 2002; Torheim et al. 2004). Data collection
FFQs have become the primary method of measuring
dietary intake in epidemiological studies because of their Data collection was conducted by locally-trained field staff
low cost, ease of administration, low subject burden and in June to December 2011. Training took place at Izulu
the ability to measure mean intakes over an extended Orphan Projects (IOP) and included practice interviews
period of time in large numbers of subjects (Willet, on IOP staff. Staff were trained by the principal investiga-
1998). Although it is common practice to adapt previ- tor (SS) and supervised during data collection. The inter-
ously developed FFQs, such as the Block (National Can- views were conducted in local language (Zulu). A
cer Institute) (Block et al., 1986) and Willett (Harvard) pre-established method (Sharma 2011) was followed
(Willett et al. 1987) questionnaires, for different study based on the steps outlined below.
populations (Tucker et al., 1998), it is well recognised
that food availability, accessibility and preferences differ, Compilation of a complete and accurate food list
often substantially, between settings and across different Subjects were asked to complete one interviewer-adminis-
ethnic population groups. Therefore, FFQs need to be tered 24-h recall. Interviews were conducted at the com-
developed specifically for each population to produce munity centres in the villages and systematically sought
valid and reliable dietary data (Sharma 2011). The aim and recorded information about all foods and drinks con-
of the present study was to characterise the diets of sumed during the preceding 24-h period. Data were
rural South African men and women in KwaZulu-Natal recorded on dietary assessment forms developed specifi-
(KZN) by identifying the most commonly consumed cally for the present study. An additional list of questions
foods, as well as other foods that make an important was included to prompt for easily forgotten foods, such
contribution to the diet, with the aim of developing a as sweets, alcohol and snacks. Questions were also
444 ª2013The British Dietetic Association Ltd.
T. Sheehy et al. Development of a quantitative food frequency questionnaire
included on smoking, employment status, use of dietary (United States Department of Agriculture, 2011) were
supplements, existing medical conditions, medication updated to include the weighed recipes for commonly
usage, and any special dietary practices that the respon- consumed local composite dishes. All analyses were
dents followed, such as weight-loss or low-fat diets. All undertaken using SAS, version 9.3 (SAS Institute, Inc.,
subject demographics and general characteristics, includ- Cary, NC, USA).
ing age, employment, smoking status and disease history,
were self-reported. Results
Determination of culturally appropriate portion sizes Subject characteristics
The amount of foods consumed was assessed using famil- Eighty-one participants completed a single 24-h recall.
iar household units (e.g. cup, spoon, glass, bowl), stan- Two subjects whose estimated energy intakes were extre-
1 1
dard units (e.g. slices of bread) or three-dimensional mely high [>20 920 kJ day (>5000 kcal day )] were
models (Nasco International, Fort Atkinson, WI, USA) excluded, leaving a final sample of 79 (34 men and 45
that had been carefully chosen to best estimate the women). Subject characteristics are shown in Table 1.
amount consumed. Each interviewer was given a set of
food models, as well as commonly used bowls, cups, Table 1 Demographic and other information among study
glasses and spoons, to use during the interview. To derive participants in rural KwaZulu-Natal
weights for the portion sizes that respondents reported Males (n = 34), n (%) Females (n = 45)
consuming in the 24-h recalls, the interviewer weighed
portions for all reported foods. Characteristics Mean (SD) Mean (SD)
Development of the draft quantitative food frequency Age (years) 44.0 (17.4) 49.6 (15.2)
questionnaire n (%) n (%)
Any food item reported on the 24-h recalls by more than Employment
one subject (apart from foods contributing little or no Yes 6 (17.6%) 11 (24.4%)
energy value, such as condiments and spices) was No 9 (26.5%) 13 (28.9%)
included on the draft QFFQ. Foods that did not appear No response 19 (55.9%) 21 (46.7%)
in the 24-h recalls but were considered relevant to rural Smoke
South Africans, such as seasonal foods, were also added Yes 14 (41.2%) 1 (2.2%)
(MacIntyre et al. 2002). Additional blank lines were pro- No 7 (20.6%) 38 (84.4%)
vided under each food group for respondents to list any No answer 13 (38.2%) 6 (13.3%)
Chewing tobacco
other foods or drinks that they consumed. Yes 1 (2.9%) 3 (6.7%)
No 26 (76.5%) 33 (73.3%)
Nutritional composition of mixed dishes No answer 7 (20.6%) 9 (20.0%)
To calculate the nutritional composition of commonly Usual amount eaten
consumed local composite dishes, weighed recipes were More than usual 1 (2.9%) 1 (2.2%)
collected as described previously (Sharma, 2011). Briefly, Less than usual 5 (14.7%) 5 (11.1%)
five community volunteers were asked to cook each com- Yes 27 (79.4%) 37 (82.2%)
No response 1 (2.9%) 2 (4.4%)
posite dish under the supervision of a trained investiga- Supplement use
tor. Individual ingredient weights (having allowed for Yes 1 (2.9%) 6 (13.3%)
inedible or discarded material) and final cooked weights No 22 (64.7%) 28 (62.2%)
of each of the five dishes were recorded by the investiga- No response 11 (32.4%) 11 (24.4%)
tor, and an average recipe was calculated for each dish. Conditions reported
All food weights were obtained using an electronic Hypertension (only) 2 (5.9%) 3 (6.7%)
kitchen scale (Aquatronic Baker’s Dream Scale; Salter Diabetes (only) – 1 (2.2%)
HIV 1 (2.9%) 8 (17.8%)
Houseware, Ltd, Fairfield, NJ, USA). Multiple conditions (>1)* – 5 (11.1%)
Other 1 (2.9%) –
Statistical analysis No reported conditions 30 (88.2%) 26 (57.0%)
Medications
Dietary data from the 24-h recalls were coded and analy- Blood pressure 2 (5.9%) 6 (13.3%)
sed using NUTRIBASE, version 9 (Cybersoft Inc., Phoenix, Diabetes – 5 (11.1%)
AZ, USA), which calculated energy and nutrient intakes Anti-retroviral therapy 1 (2.9%) 6 (13.3%)
per person. The food composition tables in Nutribase *Diabetes and hypertension.
ª2013The British Dietetic Association Ltd. 445
no reviews yet
Please Login to review.
