Food price perception, food and beverage marketing and the nutritional status of children 6–24 months in Obunga slums, Kisumu Kenya: a cross-sectional study

Study area

The Obunga slums are located in Kisumu County, on the Eastwards side; the coordinates are -004′44″ N and 34,045′53E. It is in a total land area of 1.39 sq. km. It encompasses five smaller regions, Kasarani, Central 1, Central 2, Kamaokowa and Sega Sega. Obunga slums are next to the Kisumu Industrial area; it has emerged due to a shortage of affordable housing in Kisumu City.

Study population

The study focused on children between 6–24 months residing in Obunga slums, plus their caregivers. The household was the unit of analysis, and respondents were the caregivers with children between 6–24 months.

Inclusion criteria included households with children aged 6–24 months residing in the Obunga slums.

Exclusion Criteria: Households with children aged 6–24 months residing in the Obunga slums had deformities and abnormalities (congenital disorders). As they are nutritionally vulnerable [18], The congenital disorders were established through observations, caregivers’ reports and child records in the Mother and Child Booklet.

Study design

A cross-sectional design was adopted, where data was collected once in March 2019 and analyzed. This design makes it easier to rapidly and effectively identify the relationship between the study’s dependent and independent variables and to collect quantitative data (Ferderer, 2005). The design will enable the constitution of a hypothesis that can be subjected to analytical study. The merits behind cross-sectional design include; exposure and outcome being measured simultaneously, data being collected once and then analyzed, and it describes both absolute and relative risks.

Sample size determination and the sampling procedure

Sample size determination

The sample size was determined according to Fisher et al., (1991) using the formula.

Where:

n = represented the minimum sample size (for a population > 10,000) required

Z = the standard normal deviate at the required confidence level (set at 1.96 corresponding to 95%, Confidence level adopted for this study)

p= population proportion estimated to be stunted in Obunga. This now stands at 40.2% (Okeyo, 2015)

q = 1-pd= level of statistical significance set (5%)

Therefore, on substitution

However, since the targeted population was 274 eligible households, the final sample size (nf) was adjusted as follows:

Where:

nf = desired sample size (when the target population is less than 10,000) 1.742

n= desired sample size (when the target population is greater than 10,000)

N= the desired sample size (target population)

A non-response rate of 20% was added to cover the anticipated non-responses and fouled (spoilt) questionnaires [7].

Sampling procedure

Listing was done to ascertain the actual numbers because Obunga slums have been due to the rapid migration in and out of the slums. Simple random sampling was then used to select 189 households from the 274 households listed in February 2019 as having children between 6-24 months. This was done in the following way.

1. 1.

   The first step was to assign all the households with children between 6–24 months, numbers 1–274, having determined the population size of 274 and a sample size of 189.

2. 2.

   Step 2 established a starting point by randomly opening a page and dropping a finger on the page with closed eyes.

3. 3.

   In Step 3, four pieces of paper were written to choose the direction (up to down, down to up, left to right, and right to left). The pieces of paper were folded, shaken, and the direction from left to right was chosen.

4. 4.

   In Step 4: The first unique 189 numbers were selected by reading from a table whose last three digits were between zero and 274. This was done because 274 is a 3-digit number.

5. 5.

   Numbers were not repeated once chosen.

6. 6.

   In Step 5, A new starting point was chosen; as we arrived at the end of the table, before meeting the target of 189 unique numbers, the direction was changed to up and down, and the target was achieved.

Data collection instruments

A questionnaire was developed for this study and is supplementary file 1. The questionnaire had the following sections.

Questionnaire

An interviewer-administered questionnaire collected data on food price perceptions and Food & Beverage Marketing.

Anthropometric assessment form

An anthropometric data collection form was used to gather information on the children’s height, weight and age.

Data collection procedures

1. A.

   Questionnaire

   The researcher collected data through a face-to-face interview in the selected households using a questionnaire built into the mobile app-kobo-collect. Data was collected on food price perceptions and food & beverage marketing.

   1. a.

      Food price perceptions: Data was collected on the consumption of the twelve food groups they included: cereals and grains; roots and tubers; legumes, seeds and nuts; milk and milk products; flesh meat and meat products; fish and other seafood; organ meat; eggs; Vitamin A rich fruits; other fruits; Vitamin A rich vegetables; Dark green leafy vegetables; Other Vegetables.

      Data collection was done according to the caregivers’ perception of the food price; Caregivers were requested to rate the food price from the twelve food groups [19] into either low, middle, or high.

   2. b.

      Food and beverage marketing: Data was collected on exposure to media. This was checked by the frequency of reading newspapers, listening to the radio, watching television, and accessing social media. Data was also collected on promotional practices observed by mothers on commercially produced complementary foods since the child’s birth, and if so, where they had seen or read the promotion. Finally, data were collected on the utilization of these foods, measured by caregivers reporting feeding their child any commercial food products before the interview day.

2. B.

   Anthropometric assessment

   This was measured through the recumbent length of each child. Each child’s length was measured twice to the nearest 0.1cm, and measurements were repeated when there was a deviation of >±0.5 cm. This was done through an infant/child length and height wooden measuring board by UNICEF (S0114530 Portable baby/child L-hgt mea.syst/SET-2). A child would be placed between the two slanting sides on their back. The head would gently be put against the top end, and the legs gently pushed downwards by the caregiver. The foot piece was slowly moved to the child until it pressed softly against the child’s soles, and the child’s feet were at right angles to the legs.

   Weight was then measured to the nearest 0.1g using SECA Model 881 digital scale (SECA GmbH, Hamburg, Germany). The children would have minimal clothing to avoid errors.

   The weighing scale was calibrated by placing a standard 20-kilogram weight on the scale every morning to ensure the scale could accurately measure 20 kg. If any error was seen, the scale was adjusted. The standard weight would be placed on the scale three consecutive times to ensure it has similar results three times, ascertaining its reliability. The anthropometric measures were done by taking two measurements of weight and two measurements of height; if the weight measure varied by plus or minus 0.1kg, it would be repeated. It would also be repeated if the height measure varied by plus or minus 0.1 cm.

Pre-testing

The pre-testing was done on 19 respondents, who accounted for 10% [20] of the calculated sample. After which appropriate adjustments were made to the tool. The pre-testing was done in the Nyalenda slums, an informal settlement in Kisumu County similar to the Obunga slums. The results obtained helped to rework the questionnaire and standardize it.

Validity and reliability

Content validity, which had to do with the instrument’s format, including clarity of printing, size of type, adequacy of workspace, appropriateness of language and clarity of directions [21], was achieved by giving the instruments to the nutrition professionals to go through. Both face validity and content validity were ascertained.

Test re-test reliability was used to assess the consistency of a measure from one time to another. The time between one test and the other was one week. The correlation increases with decreasing time gaps, whereas decreasing time gaps result in a lesser correlation. This is due to the two observations’ temporal correlation; the closer the time, the more similar the error-causing elements will be [21]. The validity of the weighing scale was ensured by placing a standard 20 kg measure on the weighing scale and calibrating it to ensure it read 20 kg. The reliability was confirmed by repeating this three times. Validity of the Height Board was done by placing a standard 1-m ruler on the height board and measures taken. This was repeated three times to ensure reliability.

Study variables

1. A.

   Independent variables

   The Independent variables were

   1. a.

      Food Price Perceptions: comprises data on the rating of food prices from the following food categories cereals and grains; roots and tubers; legumes, seeds and nuts; milk and milk products; flesh meat and meat products; fish and other seafood; organ meat; eggs; Vitamin A rich fruits; other fruits; Vitamin A rich vegetables; Dark green leafy vegetables; Other Vegetables [19]

   2. b.

      Food and Beverage Marketing: This comprised data on access to media (television, radio, newspapers, and social media) and the promotional practices observed by the caregivers on commercially produced complementary foods and utilizing these foods.

The information solicited was used to facilitate the assessment of broad food systems that can influence the nutritional status of children living in Obunga slums between 6-24 months of age.

1. B.

   Dependent variable

   The nutritional status of the child was the dependent variable. This was treated as both a categorical variable and a continuous variable. The categorical variables were wasting, stunting, and underweight, which reflects WAZ, LAZ, and WLZ below -2 standard deviations, below the population median, and overweight, the WLZ above two standard deviations above the population median. To measure the length for age Z Scores (LAZ), the child’s length and age were plotted against the WHO Length for age growth charts. To measure weight for length Z-scores (WLZ), the child’s weight and age were plotted against the WHO weight for length growth charts. To measure weight for age Z-Scores (WAZ), the child’s weight and age are plotted against the WHO weight for age growth charts.

Data analysis

Data was imported from Kobo Collect to Microsoft Excel. Anthropometric data and information were entered into the ENA for SMART Software. Scores for height and nutritional status were generated based on WHO Child Growth Charts and Reference 2007 charts for children aged up to two years. Then all the data was imported into the Statistical Package for Social Sciences (SPSS) Version 25 (Illinois, Chicago). Descriptive statistics and Inferential Statistics were used to analyze data. Frequencies and proportions presented data analyzed through descriptive statistics through tables and text. Binary logistic regression was used to determine the relationship between food price perceptions, food and beverage marketing and the nutritional status of children between 6–24 months in Obunga slums. Crude Odds Ratio (COR) and Adjusted odds ratio (AOR) with Confidence Interval (C.I.) were then computed in binary logistic regression based on a 95% level of significance. To test the strength of the association between nutritional status, food price perception, and food and beverage marketing in the Obunga slums.