Impact of Community Masking on COVID-19: A Cluster-Randomized Trial in Bangladesh

"Even in places where it is obligatory, people tend to optimistically overstate their compliance for mask wearing. How then can we persuade more of the population at large to act for the greater good?"

In an attempt to stem the tide of the COVID-19 pandemic, the Bangladeshi government formally mandated mask use in late May 2020 and threatened to fine those who did not comply, although enforcement was weak to nonexistent, especially in rural areas. Thus, Innovations for Poverty Action (IPA) partnered with Yale Research Initiative on Innovation and Scale (Y-RISE), Stanford University Medical School, and a local non-governmental organisation (NGO) called Green Voice to test various strategies to increase mask-wearing. This paper reports on a large-scale randomised controlled trial (RCT) to evaluate a portfolio of encouragement strategies, assessing whether these strategies could increase mask usage and reduce symptomatic SARS-CoV-2 infections.

The researchers selected intervention elements that had a reasonable chance of persuading rural Bangladeshis to wear masks by consulting literature in public health, development and behavioural economics. Inspired by large literature in marketing and economics on the role of opinion leaders in new product diffusion, they emphasised a partnership with community leaders in mask distribution. Also, masks are a visible good where social norms are expected to be important, so the researchers consulted the literature that documented peer effects in product adoption. They decided to assess whether soft commitment devices encourage people to follow through with actual behaviour change, whether public displays can promote social norms, whether an altruistic framing inspires people more or less than self-interest, whether social image concerns and signaling can lead to higher compliance, and whether regular reminders are a useful tool to ensure adoption.

With this conceptual basis, and following two pilot studies, the researchers conducted the RCT in rural Bangladesh from November 2020 to April 2021 (N = 600 villages, N = 342,183 adults). There were 178,322 individuals in the intervention group and 163,861 individuals in the control group. The basic set of interventions implemented in all intervention villages consisted of 5 main elements:

1. One-time mask distribution and information provision (about masks) at households, including the showing of a brief video of notable public figures (the Honorable Prime Minister, the head of the Imam Training Academy, and a national cricket star) discussing why, how, and when to wear a mask;
2. Mask distribution in markets for 3-6 days per week during all 8 weeks of the intervention;
3. Mask distribution at mosques on 3 Fridays during the first 4 weeks of the intervention;
4. Mask promotion in public spaces and markets, where non–mask wearers were encouraged to wear masks (weekly or biweekly); and
5. Role modeling and advocacy by local leaders, including imams discussing the importance of mask-wearing at Friday prayers using a scripted speech provided by the research team.

The 4-part model to change social norms can be abbreviated as NORM = No-cost free masks distribution, Offering information on mask-wearing, Reinforcement in-person and in public, and Modeling and endorsement by trusted leaders.

In addition, within the intervention arm, the researchers cross-randomised villages to 4 village-level and 3 household-level treatments to test the impact of a range of social and behaviour change communication (SBCC) strategies on mask-wearing. All intervention villages were assigned to either the treatment or the control group of each of these randomisations. The village-level randomisations were as follows:

• Randomisation of treated villages to either cloth or surgical masks;
• Randomisation of treated villages to public commitment (providing households signage and asking them to place signage on doors that declares they are a mask-wearing household) or not, which was meant to encourage formation of social norms through public signaling;
• Randomisation of treated villages to noincentive, nonmonetary incentive, or monetary incentive of $190 given to the village leader fora project benefitting the public; and/or
• Randomisation of treated villages to 0 or 100% of households receiving twice-weekly text message reminders about the importance of mask-wearing.

The household-level randomisations, described in further detail in appendix D of the paper, were as follows:

• Randomisation to receive messages emphasising either altruism or self-protection;
• Randomisation to making a verbal commitment to be a mask-wearing household (all adults in the household promise to wear a mask when they are outside and around other people) or not (conducted in a third set of villages where there was no public signage commitment); and/or
• Randomisation to receive twice-weekly text reminders or not.

Mask-wearing and distancing - both physical and social - were assessed through direct observation at least weekly at mosques, markets, the main entrance roads to villages, and tea stalls. Individuals were coded as physically distanced if they were at least one arm's length from the nearest adult; social distancing was measured using the total number of adults observed in public areas. At 5- and 9-week follow-ups, the researchers surveyed all reachable participants about COVID-19-related symptoms. Blood samples collected at 10- to 12-week follow-ups for symptomatic individuals were analysed for SARS-CoV-2 immunoglobulin G (IgG) antibodies.

Key findings:

• The core NORM intervention increased proper mask-wearing from 13.3% in control villages (N = 806,547 observations) to 42.3% in treatment villages (N = 797,715 observations) (adjusted percentage point difference = 0.29; 95% confidence interval (CI) = [0.26, 0.31]). This tripling of mask usage was sustained during the intervention period and for 2 weeks after. After 5 months, the impact of the intervention on mask-wearing waned, but mask-wearing remained 10 percentage points higher in the intervention group.
• Adjusting for baseline covariates, the intervention reduced symptomatic seroprevalence by 9.5% (adjusted prevalence ratio = 0.91 [0.82, 1.00]; control prevalence = 0.76%; treatment prevalence = 0.68%). In villages randomised to surgical masks (N = 200), the relative reduction was 11.1% overall (adjusted prevalence ratio = 0.89 [0.78, 1.00]). The effect of the intervention was most concentrated among the elderly population; in surgical mask villages, there was a 35.3% reduction in symptomatic seroprevalence of COVID-19 among individuals 60 years old or older (adjusted prevalence ratio = 0.65 [0.45, 0.85]).
• Physical distancing increased from 24.1% in control villages to 29.2% in treatment villages (adjusted percentage point difference = 0.05 [0.04, 0.06]).  On average, physical distancing increased by 5 percentage points across all locations in intervention villages, but the change was larger in some locations than others. In markets, people were substantially more likely to physically distance (7 percentage points increase). In mosques, researchers observed no change (they observed little physical distancing in any mosques). The intervention had no impact on social distancing.
• In-person public reinforcement is a critical part of the intervention. As noted above, the researchers conducted two pilot tests before the RCT. In the first pilot, masks were only distributed. The second pilot included mask distribution, promotion, and in-person periodic reinforcement. Researchers found proper mask-wearing increased 28 percentage points when masks were not only distributed but promoted and their use was reinforced (pilot 2).
• People were just as likely to wear a surgical mask, which is less expensive to produce, as a cloth mask. However, local preferences matter: Mask colour had a significant effect on uptake, both for surgical and cloth masks. In villages where surgical masks were distributed, blue surgical masks were 3 percentage points more likely than green surgical masks to be observed. In villages where cloth masks were distributed, violet cloth masks were 6 percentage points more likely than red cloth to be observed.
• Some groups had higher take-up in mask use than others. The largest increase in mask use was in mosques, which saw a 38-percentage-point increase in mask use, while in all other locations it was roughly 25-29 percentage points. Researchers also found a larger increase in mask use in villages that had below-average initial mask-wearing. In those villages, mask-wearing went from 9% to 43% after, whereas in villages with above-average initial mask-wearing, it went from 18% to 43%.

None of the community- or household-level behavioural interventions or incentives led to additional increases in mask-wearing or physical distancing. That is, neither text message reminders, public signage, monetary and non-monetary incentives, altruistic messaging, or verbal commitments had any further effect on mask-wearing.

The NORM intervention was designed to be easily adopted by other NGOs or government agencies and required minimal monitoring. The materials have been made public in multiple languages to ease widespread adoption and replication by other implementers. Click here to access them.

Full list of authors, with institutional affiliations: Jason Abaluck, Yale University; Laura H. Kwong, Stanford University and University of California, Berkeley; Ashley Styczynski, Stanford University; Ashraful Haque, Innovations for Poverty Action (IPA) Bangladesh; Md. Alamgir Kabir, IPA Bangladesh; Ellen Bates-Jefferys, IPA; Emily Crawford, Yale University; Jade Benjamin-Chung, Stanford University; Shabib Raihan, IPA Bangladesh; Shadman Rahman, IPA Bangladesh; Salim Benhachmi, Yale University; Neeti Zaman Bintee, IPA Bangladesh; Peter J. Winch, Johns Hopkins Bloomberg School of Public Health; Maqsud Hossain, North South University; Hasan Mahmud Reza, North South University; Abdullah All Jaber, North South University; Shawkee Gulshan Momen, North South University; Aura Rahman, North South University; Faika Laz Banti, North South University; Tahrima Saiha Huq, North South University; Stephen P. Luby, Stanford University; Ahmed Mushfiq Mobarak, Yale University and Deakin University