How Does a Social Norms-Based Intervention Affect Behaviour Change? Interim Findings from a Cluster Randomised Controlled Trial in Odisha, India

"...findings show that a social norms-based intervention can be successful in increasing IFA [iron folic acid] use..."

The theory of normative social behaviour (TNSB) highlights the role social norms can play in influencing health behaviours and the circumstances under which they may do so. Guided by the TNSB, the Reduction in Anaemia through Normative Innovations (RANI) project designed a multicomponent behavioural intervention intended to increase iron folic acid (IFA) use and decrease anaemia among women of reproductive age in Odisha, India. This paper examines the effect of the intervention at midline to understand which components of the RANI intervention affected IFA uptake. The hope that delineating the effects of each intervention component will provide guidance for future social norms-based intervention designs.

Findings from formative research that examined social norms around IFA (e.g., via vignettes) use led to the design of the intervention, which included three components (described in Box 1 in the paper and in more detail at Related Summaries, below): (i) hands-on participatory learning modules, featuring games, conducted in small groups, (ii) dissemination of short videos focusing on iron consumption norms, and (iii) monthly haemoglobin testing for anaemia, followed by public display of (anonymised) community results via blood-shaped cards (different coluors indicating anaemia severity) and infographics appropriate for a low-literacy population. The focus of the three intervention components was more on injunctive norms messaging (that all women of reproductive age should take IFA) rather than descriptive norms messaging (that women are taking IFA).

Using a cluster randomised controlled design, the researchers collected baseline data in September 2019 and midline data in February 2020 from women of reproductive age in the control and treatment arms (n=3,800) in Angul, Odisha, India. Using nested models, they analysed data from the three different intervention components to determine the extent to which exposure to each of these components accounted for the overall intervention effect on haemoglobin and self-reported IFA use (documented in another midline paper, found at Related Summaries, below).

Participants from the treatment arm reported more IFA use at midline (32%) compared with the control arm (3%) (p≤0.001). However, haemoglobin levels were not statistically different in the two arms (11.7% vs 11.5%) (p=0.28). Nested logistic regression models predicting self-reported IFA use as a function of study arm reveal, for example, that, after controlling for demographic variables and IFA use at baseline, simply being in the treatment arm increased a woman's odds of taking IFA by nearly 17 times (odds ratio (OR) = 16.94; p≤0.001).

Also, exposure to each of the intervention components had a dose–response relationship with self-reported IFA use. For example:

• Each educational participatory learning session that women attended significantly increased IFA use. These sessions were designed primarily to raise awareness and improve knowledge about anaemia, iron-rich foods, and diet diversity. However, they also included discussions about gender roles, how eating order in the home disfavours women's health, and the need to remain strong to be able to take care of others.
• Exposure to the communication videos was significantly associated with increased IFA uptake. The four videos were designed for different audiences, including adolescents, husbands, mothers-in-law, and women of reproductive age, and yet were all intentionally shown to each of the groups. The overall strategy was to communicate the message that the entire community was engaged in the task of reducing anaemia.
• Haemoglobin testing provided individual health information, village-level health information, a comparison to other villages, and an indication of changes in health information over time. The public events, which the researchers describe as "highly demanded and motivational", seem to have had the effect of normalising women prioritising their own health, potentially shifting gender norms that solely focus on pregnant women's health or the health of the family. "This multilevel component, coupled with the other two components, may help women reach the tipping point to take and adhere to IFA."

There was very little difference in exposure to the intervention components by age, education, or caste, which implies that the RANI Project's outreach efforts did not disproportionately favour those with higher privilege in society.

The researchers conclude that, "While IFA use shows promise, haemoglobin levels may need more time to show significant changes, especially among non-pregnant women who were not taking IFA and who only take it weekly. End line results will elucidate more information about the full RANI intervention effects."