Behavioural Nudges and Maternal Diet: Results from a Cluster-Randomised Pilot Trial among Pregnant Women in India

The study design and the interventions were approved by the Institutional Review Board of CHRIST (Deemed to be University), Bangalore (Reference ID - CU: RCEC/59/08/21). Informational nudges were designed to create awareness about a healthy diet focusing on meal regularity, meal frequency and diet diversity. The Treatment group was further split into Treatment arms 1, 2 and 3. The nudges designed for intervention are Informational Pamphlets (Based on the concept of MyPlate and modified as per local food habits) provided to all the treatment groups. The pamphlets distributed in the local language has been attached as supplementary file. Treatment arm 1 received the informational pamphlets only. Treatment arm 2 received in-house visits from ASHA workers every 5th day from the day of recruitment while 3 received phone calls from the research team to remind them about the importance of a healthy diet during pregnancy. This was done primarily to examine whether the source of information made any difference in the way women changed their food intake and the composition of food. Both ASHA workers and research team members were given identical scripts to be focused upon during their interaction with the pregnant women. Figure 1 shows the intervention design.
A gap of 28 days between baseline and end-line was maintained to measure the effect of the nudge. The difference in food intake and dietary diversity between baseline and endline were attributed to the behavioural interventions, after controlling for all other relevant factors. The information was collected based on a 24-hour recall of food and beverages consumed prior to the time of the survey. The data was collected twice within a week at both baseline and end-line, using a detailed questionnaire. The intention of conducting the recall twice was to capture variations, if any, in the dietary pattern of pregnant women. A single recall in a week is likely to miss these variations. The questionnaire covered demographic details, and information about intra-household family dynamics in food distribution, meal frequency and diet diversity.
The 24-hour recall approach in this study involved guiding participants to meticulously recount all foods and beverages consumed in the preceding 24 hours. This method ensured the collection of detailed data on portion sizes, preparation methods, and other relevant dietary information. To enhance the accuracy of portion size estimation, standard cups and utensils were displayed to all respondents, helping them gauge their intake relative to these standard measures.
Studies have shown that the 24-hour recall method can maximise response rates and reduce the burden on both respondents and researchers, while still providing high-quality data. The method is extensively used in national dietary surveys to provide detailed dietary intake data. This data is crucial for informing public health policies and nutritional guidelines.

2022-03-22

2022-04-18

Small Meals Frequency
Meal Diversity

Small Meals Frequency
Meal frequency during pregnancy is a critical aspect of maternal nutrition that may influence the risk of preterm delivery. Analysis of data from the pregnancy, infection, and nutrition study involving 2,065 women revealed that those who consumed fewer meals and snacks per day had a higher likelihood of delivering preterm, particularly following premature rupture of membranes (Siega-Riz et al., 2001). Other studies recommend that pregnant women consume smaller, more frequent meals rather than larger ones to prevent various food related complications (Abriha et al., 2014; Gete et al., 2020). The interventions in this study specifically targeted improving the number of small meals by emphasizing its importance and the possible low cost or readily available meals the respondents can consider. The number of smaller meals consumed by the pregnant women before breakfast, between breakfast and lunch, between lunch and dinner and after dinner are assessed using an ordinal logistic regression.
The dependent variable is the total count of four smaller meals consumed at various intervals at the endline. Each of these meals is recorded as a yes/no outcome and then summed up to form an ordinal measure of meal frequency. The independent variable is the combined effect of all three interventions, and baseline values are included as controls.





Meal Diversity

To assess changes in dietary diversity, two complementary metrics were employed: the continuous average dietary diversity score and the binary indicator for achieving Minimum Dietary Diversity for Women (MDD-W).
The Minimum Dietary Diversity for Women (MDD-W) is the primary focus, following guidelines from the Food and Agriculture Organization (FAO). MDD-W is a globally recognized indicator designed to evaluate diet quality and serve as a proxy for micronutrient adequacy among women of reproductive age (15-49 years) (Martin-Prével et al., 2015). It is a dichotomous measure indicating whether women have consumed at least five out of ten pre-defined food groups over the previous 24-hour period. These ten food groups are: 1) Grains, white roots and tubers, and plantains; 2) Pulses (beans, peas, and lentils); 3) Nuts and seeds; 4) Dairy; 5) Meat, poultry, and fish; 6) Eggs; 7) Dark green leafy vegetables; 8) Other vitamin A-rich fruits and vegetables; 9) Other vegetables; and 10) Other fruits. Achieving this five-group threshold is considered indicative of a higher likelihood of meeting micronutrient needs. The proportion of women achieving MDD-W was analysed as a key binary outcome.
In addition to the binary MDD-W indicator, a continuous dietary diversity score was also analysed. This score represents the sum of unique food groups (out of the ten listed above) consumed by each woman over the recall period, ranging from 0 to 10. The use of this continuous score provides a more granular measure of dietary diversity, allowing for the detection of changes in the number of food groups consumed, even if these changes do not result in crossing the MDD-W threshold of five food groups. Analyzing the continuous score alongside the binary MDD-W indicator offers a more comprehensive understanding of the interventions' impact on overall dietary patterns and quality.

This study employs a cluster-randomised field experiment design, wherein 22 Primary Health Centres (PHCs) were randomly allocated to one of four arms: a control (Arm 4) or one of three information-nudging treatments (Arms 1–3). Each woman is observed twice—at baseline (t = 0) and end-line (t = 1), allowing treatment effects to be identified via a two-period, multi-arm difference-in-differences (DiD) strategy. Under the parallel trends assumption (Angrist & Pischke 2009; Goodman-Bacon 2021), the DiD estimator is unbiased. The validity of this assumption is supported by randomisation at the PHC level and testing for the absence of significant pre-treatment imbalances will support the plausibility of this assumption.

Randomisation done using RAND command in excel

Public Health Centres (PHCs)

Yes

22 PHCs

415 pregnant women

Twenty-two Primary Health Centres (PHCs) were drawn at random from the 69 facilities in Chikkaballapur District that lie within 60 km of the District Health Centre. These 22 clusters were then randomised into four arms: Treatment 1 (pamphlet, n = 4 PHCs), Treatment 2 (ASHA home-visit, n = 4), Treatment 3 (reminder phone call, n = 4) and Control (passive, n = 10). Recruitment was carried out by the resident Accredited Social Health Activists (ASHAs), who maintain the antenatal registers for their catchment areas. Baseline enrolment yielded 440 pregnant women aged three to eight months’ gestation—199 in the control arm and 241 across the three intervention arms (81, 80 and 80, respectively).