Enhancing nutrition knowledge and dietary diversity among rural pregnant women in Malawi: a randomized controlled trial

We conducted the RCT from January to December 2016 in the Namkumba area in Mangochi, one of the 12 districts in the southern region of Malawi. Mangochi district was prioritised for this intervention as it faces major challenges in maternal and child undernutrition, according to the Demographic Health Survey of 2010 [19]. Nankumba area is situated on the southern end of Lake Malawi (Fig. 1). We included pregnant women attending ANC from Monkebay, Namkhumba, Malembo and Nkope health centres. The population of the area is about 150,000 and the inhabitants' livelihood is mainly fishing or farming (depending on geographical location) [20].

All consenting pregnant women who were between their 9^th and 16^th gestational week, both primi- and multiparous, were eligible for inclusion. Moreover, they had to be available during the study period and had to intend to give birth at the health facilities within the study area. Women carrying multiple foetuses and those with severe illnesses were excluded from the study.

Twenty clusters, defined as a village which did not share common borders with other villages in the study, were eligible to be included and were mapped out in the study area. The villages were assigned STATA-generated random numbers to allocate them to either the intervention or the control group, generating ten clusters in each study group. The women from these clusters were identified by the study team in their communities and referred to the nearest designated local health centre where their pregnancy and the gestational age were confirmed and where they were also invited to consent to participate in the study.

The intervention addressed some of the underlying and contextual causes of undernutrition as per the UNICEF’s framework [22], such as poor dietary quality and factors usually resulting from inadequate: (i) knowledge of the benefits of good nutrition and nutrition dense foods, (ii) skills in food preparation that ensure the inclusion of diverse food groups in the diet, (iii) food processing that enhances absorption of nutrients from the gut, and (iv) motivation to put in the efforts needed to adhere to the recommended dietary guidelines. In the intervention, we therefore applied nutrition education, cooking demonstrations and home-based counselling. We used lay counsellors who were women from the same community (and who were already involved in community activities) to facilitate the education and counselling sessions. Moreover, we trained the lay counsellors in the intervention processes, and to use standard guidelines to ensure the implementation of the intervention.

The education component was offered before cooking demonstration to groups of women (6–20 women per group). Nutrition education was oriented towards enhancing individual understanding of the importance of eating diverse food, communicating the effects of diversified diet for mother- and child health, and conveying information about the food groups and their function in the human body. The cooking demonstrations were focused on teaching the women how to prepare foods that integrated the SFG dietary guidelines. These sessions were done once a month, during the week and utilised a whole afternoon. Prior to the session, the women were assigned to prepare the powder ingredients, process maize flour and mobilise all other resources needed. Most women attended about 4 sessions before they delivered.

We demonstrated on how to use the recipes that included ingredients made from food supplements turned into powder seasoning to make the meals more diversified. The powders were from anchovy (Bonya, dried small bony fish), groundnuts (Nsinjiro) and dried moringa tree leaves (Chammwamba). Anchovies are a good source of calcium, iron, and zinc, groundnuts are a good source of proteins and fats, while moringa leaves are good source of protein, vitamin B₂, vitamin B₆, vitamin C, iron, vitamin A and magnesium [23]. The recipes combined foods with high content of nutrients that were lacking in their usual diet. The most limiting foods in these settings were legumes while animal foods and vegetables were usually inaccessible at upland and lake shore household. Nsima was usually taken with one relish, which made the diet comprise of only two food groups of the required six. The most accessible main meal combination from upland was Nsima (a thick porridge), usually taken with only vegetables, whereas in the lakeshore areas Nsima was commonly taken with fish.

The diets in the communities participating in the study were predominantly plant-based and thus contained a high content of phytate and polyphenols that inhibit the absorption of iron, zinc and calcium [24]. The women thus got demonstrations on food processing that enables increased absorption of nutrients, such as soaking beans overnight before cooking and soaking whole grain maize for three days before turning it into maize meals used for making porridge and the staple food Nsima.

The practices promoted in this intervention were harmonised with local dietary patterns and food beliefs. Recipe development was guided by linear programming analyses of dietary intakes of pregnant women in the area. The project conducted Trials of Improved Practices to identify challenges with use of the recipes over a longer time which could hinder adoption of the recipes. In line with this, the recipes were tested among a comparable group of pregnant women in a non-study area.

We also offered home-based nutrition counselling to the women once a week to enhance intake according to the proposed dietary guidelines. Counselling sessions were offered weekly based on the individual needs and involved women with their families, aiming to promote their capabilities to source the required ingredients [25]. Each counsellor was expected to support up to twenty women within walking distance from the counsellor’s home. More information regarding delivery of the intervention has been previously described [20]. No intervention-related harm was detected.

A questionnaire that was partly adapted from a validated general nutrition knowledge questionnaire [26] was developed to collect sociodemographic data, nutrition knowledge and food consumption behaviours. Socioeconomic status was assessed using a household asset index based on eleven household items according to their monetary value and given scores very poor: ≤ 1.25; poor: 1.25–3.75; well-off: > 3.75) [27].

Women were included in the study between nine and sixteen weeks of gestation and the average exposure to the intervention was about 6 months. We assessed the study groups at inclusion (baseline) and at two weeks before expected delivery (study end-point) for differences in nutrition knowledge, perceptions and dietary habits. To assess nutrition knowledge, we asked the participants if they knew the SFG dietary guideline and if they could mention the food groups. Each food group mentioned weighted equally with the score of one, making a total score of six. We rated nutrition knowledge as low for scores 1–3, moderate for scores 4–5 and high for score 6 [28]. To assess perceptions, we asked the participants whether pregnant women have higher risk for malnutrition (yes or no). The women were also asked to evaluate their own dietary habits to determine if they were at risk for malnutrition (yes or no). We further assessed dietary habits through their recalled 24 h quantitative food consumption, from which we evaluated their scores using both the SFG and the DDS. Dietary diversification with use of SFG guidelines was applied using the same rating used for nutrition knowledge. The DDS for women (MDD-W) was based on 10 food groups as proposed by the Food and Agriculture Organisation [29]. The MDD-W is a dichotomous indicator based on 10 food groups and is considered the standard for measuring population-level dietary diversity in women of reproductive age [30, 31]. Each food group was weighted equally with the score of 1; hence the maximum possible score was 10. The MDD-W was rated as either poor (scores 0–4), moderate (scores 5–7) or high (scores ≥ 8). We also assessed the women’s self-reported nutrition behaviour using four questions, asking if they consumed diversified foods, healthy drinks and snacks, whole grains and skipped the main meals more frequently or not.

For the RCT, we estimated the sample size based on the assumption that we would be able to detect a clinical meaningful mean difference in birth weight of 150 g between the control and intervention group [32]. Using a two-sided test, an alpha of 0.05 and a statistical power of 80%, we obtained a minimum sample size of 218 participants. In the current study we performed descriptive analyses using IBM SPSS Statistics 25 to assess the differences in baseline and study end-point characteristics between the two study groups, employing the Mann–Whitney U test for non-normal continuous variables and the Pearson chi-square test for associations between categorical variables. Further, the independent T-test was used to compare the mean nutrition knowledge score and dietary diversity score between the women in the two study groups.