Study setting
This study was conducted in rural Kyankwanzi district, central Uganda [43]. The total population of Kyankwanzi district is 214 693, of which 34% are illiterate, 48% are females, 16% are child mothers 12-19 years old and 19% are children 0-4 years old [43]. Children in this study area are fed CFs deficient in vitamin A, iron and zinc [9, 10]. The preparation of the CFs and child acceptability study were conducted at Ntwetwe Health Centre IV, Kyankwanzi district.
Description of the intervention
Ingredients for preparation of the complementary foods used in the intervention
This study formulated two homemade CFs, common bean pumpkin blend (BPB) and pumpkin blend (PB). The former was formulated from cooked common bean (Obwelu) and pumpkin (Sweet cream), whilst the latter (control CF) from cooked pumpkin, Sweet cream. The PB was selected as a control because pumpkin is commonly used as a single CF in Uganda [44].
These ingredients were chosen because common bean is rich in iron and zinc [22–25], and pumpkin is rich in PVACs [26–28]. Moreover, these ingredients are cultivated in rural Uganda and available in the local markets [20, 21]. Figure 1 shows BPB, PB and ingredients, common bean (Obwelu) and pumpkin (Sweet cream) used to prepare BPB and PB.
Figure 1: Study complementary foods and their ingredients (Here)
Preparation of BPB and PB
At household level, caregivers usually prepare homemade CFs based on consistency (thinness and thickness) of the food and the child’s age and development as recommended by the 2017 Food and Agriculture Organization of the United Nations (FAO) guide to conducting participatory cooking demonstrations to improve complementary feeding practices [45, 46]. The 2017 FAO guide to conducting participatory cooking demonstrations to improve complementary feeding practices, encourages participatory cooking demonstrations involving community nutrition and health workers, mother-leaders and peer counsellors [45]. To this end, BPB and PB were prepared by child caregivers (expert peer mothers). Community health workers identified 10 expert peer mothers from the local community and invited them to Ntwetwe Health Centre IV to participate in the preparation of CFs used in the acceptability study. Expert peer mothers were encouraged to prepare CFs using the locally acceptable home-based methods used in the community to prepare common bean and pumpkin for child consumption.
Common bean (Obwelu) and pumpkin (Sweet cream) were purchased from the local market with assistance from expert peer mothers. Expert peer mothers prepared Sweet cream by peeling and discarding seeds followed by boiling the pulp. For Obwelu, expert peer mothers used overnight soaking (for about 8 hours), followed by boiling (for about 1.5 hours). After cooking, expert mothers indicated that they prepare homemade CFs in their community based on consistency (thinness or thickness of food) suitable for the child’s stage of development. To this end, after cooking by expert peers mothers, research assistants mixed the ingredients to form CFs based on the consistency as suggested by caregivers and recommended guidelines for conducting participatory cooking demonstrations to improve complementary feeding practices [45]. Research assistants prepared three different varieties of BPB by mixing and mashing Sweet cream and Obwelu together. Table 1 shows the ratio of mixing Sweet cream and Obwelu that was used to prepare BPB varieties.
Table 1. Ratio of mixing Sweet cream and Obwelu to formulate BPB (Here)
After preparing, the three varieties of BPB were put on a table in three different serving dishes and presented to expert peer mothers. Based on consistency, expert mothers, one by one entered the room and where asked to choose one variety of BPB they would choose to feed their children, 6 to 24 months old [45]. All the 10 expert mothers unanimously selected BPB-3, prepared by mixing 2 parts of Sweet cream and 1 part of Obwelu. Mashed cooked pumpkin in Uganda is usually given as a single CF [44]. Therefore, pumpkin blend (PB) as a control was prepared from Sweet cream. Triplicate samples of prepared BPB (test food) and PB (control) were transported to METLAB East Africa limited laboratory, Kampala, Uganda for PVACs, iron and zinc analysis.
Vitamin A, iron and zinc analysis of BPB and PB
The PVACs content was analysed by high performance liquid chromatography (HPLC) as described in the HarvestPlus hand book for carotenoid analysis [47]. To analyse the vitamin A content, the Institute of Medicine (2001) bioconversion rates of PVACs to vitamin A, retinol (retinol activity equivalents) were used, i.e. 12 µg of β-carotene is equivalent to 1 µg of retinol, whilst 24 µg of α-carotene is equivalent to 1 µg retinol [48]. Iron and zinc concentrations of CFs were determined by flame atomic absorption spectroscopy (FAAS) as described elsewhere [10, 49]. Triplicate analysis for BPB and PB were done separately to get the mean content of PVACs, iron and zinc in each of the two CFs.
Micronutrient composition of BPB and PB
The mean concentrations of PVACs, iron and zinc were calculated per 100 g of edible portion of CF. Table 2 shows the PVACs, iron, zinc and vitamin A content per 100 g of edible portion of BPB and PB.
Table 2. Micronutrient composition of edible portion of BPB and PB [Here]
Study participants, enrolment, inclusion, and exclusion criteria
All children (aged 6 to 24 months old) coming for growth monitoring and immunisation at Ntwetwe Health Centre IV, Kyankwanzi district Uganda were screened for nutritional status and presence of any illness. Upon fulfilling the enrolment criteria (age 6 to 24 months, on complementary feeding) and obtaining consent for participation from the caregivers, the children were randomly allocated to two different study groups (BPB and PB) and children were enrolled. Children did not meet the enrolment criteria if their weight for age or weight for height z-score was < −3 standard deviations, if they had any childhood acute illness or features suggestive of any chronic disease such as tuberculosis, any congenital anomalies such as cleft lip or palate.
Sample size determination
A maximum 50 g of CF per serving is considered adequate for children in the age range of complementary feeding [42, 50]. Therefore, the sample size was determined to test the hypothesis that the mean consumption of CF during the acceptability test would be at least 50 g (50%) of the amount offered (100g). Assuming a mean difference of 5 g between test CF and control, and a standard deviation (SD) of 10 g in a normally distributed population of children 6 to 24 months old, a sample size of 63 for each CF would therefore allow us to reject the null hypothesis with 80% power. However, this was a crossover study meaning that participants consumed both test CF and control. In order to cater for loss to follow-up, an additional seven participants were added to the 63 to make 70 participants. To this end, the same 70 participants were enrolled in each group of BPB and PB.
Study design
This was a randomised crossover acceptability study. A total of 110 children from the growth monitoring and immunisation clinic at Ntwetwe Health Centre IV, Kyankwanzi district Uganda were identified for randomisation (Figure 2). Out of the 110 children, 70 were eligible and assigned to BPB (intervention) and PB (control) using simple random sampling according to computer-generated random numbers. Computer-generated numbers were given to participants by a research assistant who was located off site. On the first day, 35 children were assigned to each group of BPB and PB. A wash out period of one day was granted, and on the third day participants crossed over to the opposite CF group. Figure 2 shows the study design.
Figure 2. Study design (Here)
Measurement of study outcomes
Child acceptability was assessed by feeding the novel CF to the child, followed by measuring the amount of CF consumed, and duration taken to complete the CF [37–40]. The primary outcome of the study was to measure the amount of CF consumed by children. The two secondary outcomes were to measure the time taken by the child to consume the served CF; and to analyse the PVACs, iron and zinc intake of each child based on the amount of CF consumed. Caregivers were also requested to report any discomfort or adverse effects experienced by the children after being fed the study CFs.
Amount of complementary food consumed by children
This study ensured that children were offered the assigned CF (BPB or PB) at least 1 hour after they were last fed. A portion of 100 g of CF was offered to the child in a serving dish by the caregiver. The CF was considered acceptable if the child ingested at least 50 g of offered food [42]. The amount of food ingested was calculated by subtracting the left-over from the offered amount. Pre-weighed napkins were provided; any food that was regurgitated, vomited or spilled was swabbed, the napkin weighed and subtracted from the weight of the amount offered.
Duration of feeding
Duration of feeding was measured as described elsewhere [37]. Caregivers were asked to spoon feed their children the assigned CF until the child refused to eat. After a two-minute pause, the same food was offered a second time until s/he refused again. After a second two-minute pause, the food was offered a third time until refused again. After this third refusal, the feeding episode was considered terminated. The duration of feeding (excluding the intervening ‘pause periods’) was recorded by stopwatch, and the total duration of the feeding was noted. The feeding episode took place under the direct supervision of a trained research assistant to ensure that feeding was not forced. Children were considered as refusing intake if they moved their head away from the food, cried, clamped the mouth shut or clenched the teeth, or became agitated, spat out the food or refused to swallow as done elsewhere [37].
Micronutrient intake measurement
The micronutrient intake (MNI) for each child was calculated using the formula, MNI= A(g)*B/100, where A was the amount of CF (BPB or PB) consumed by the child and B was the nutrient composition in 100 g of CF food served to the child (see table 2). For example, if the child consumed 50g of 100 g of BPB served, then MNI for iron, zinc and vitamin A would be 50*1.99/100 (0.995 mg), 50*1.08/100 (0.54 mg) and 50*187/100 (93.5 µgRAE), respectively.
Measurement of background characteristics
Data on background characteristics such as age, gender and nutritional status of study participants were collected. Age was calculated in months based on the difference between the date of visit and date of birth. If exact date of birth of the child was unknown, the month and year of birth were estimated using a local events calendar. In such cases, age was calculated after imputing the day of birth as the 15th of the month, as recommended by 2019 WHO guidelines [51]. Date of birth was extracted from the child’s immunisation and growth monitoring chart. Nutritional status was determined using anthropometry, and diagnosed by Z scores based on the 2019 WHO recommendations for data collection, analysis and reporting on anthropometric indicators in children under 5 years old [51]. A child was stunted, wasted and underweight if his or her length for age Z score (LAZ), weight for length Z score (WLZ) and weight for age Z score (WAZ) was below -2 standard deviations of the WHO reference respectively [51].
Data analysis
Statistical and data analysis was done by STATA version 13.1. Background characteristics of the participants were evaluated by using descriptive statistics. The mean ± SD of the amount of the CF consumed, duration of consumption, and MNI was calculated. The paired t-test was used to detect the mean differences of outcome variables within participants between BPB and PB. The level of significant difference was set at a probability value of 5% (p 0.05).
Ethical approval
Permission to conduct the study was granted by the District Health Office, Kyankwanzi district, Uganda. In South Africa, ethical approval was obtained from the Biomedical Research Ethical Committee, University of KwaZulu-Natal, South Africa (Reference number: BE 438/19). In Uganda, ethical approval was granted by The AIDS Support Organisation Research Ethical Committee (Reference number: TASO-REC/066/19-UG-REC-009). This randomised control trial was registered by Pan African Clinical Trials Registry (www.pactr.org) as PACTR202002576768667. Informed and signed consent were obtained individually from caregivers of child participants in the study, and all data were coded to remove identifying information and ensure confidentiality.