A household-level constant visual deterrent advocacy campaign to reduce tobacco intake was conducted in rural Bangladesh.

To evaluate smoking tobacco expenditure by campaign components.

We conducted a single-blind clustered randomized controlled trial on 630 adult male household heads from 16 chars (riverine islands) in rural northern Bangladesh, between November 2018 and January 2019. Intervention allotment was randomized at the char level to minimize spillovers, with 8 chars in treatment and control groups each. The intervention provided households in treatment chars (n = 323) with two visual warning posters, detailing the health effects of tobacco on oneself and external actors, to be hung inside the household for 4 weeks. Households in control chars (n = 307) received nothing. Reported daily smoking (log) tobacco expenditure values were the primary outcome of interest.

Final analysis was conducted using 251 and 210 smokers in treatment and control chars respectively. The intervention reduced relative smoking tobacco expenditure by 12.8% (95% CI −31.45 to 5.81) but was not statistically significant (P-value = 0.163). Weak to moderate emotional reactions to the posters was identified as a reason for the statistical insignificance.

For a visual anti-tobacco intervention to have a substantial impact, it must induce strong emotional responses.

This study aims to assess the impact of a behavioural intervention, in the form of a self-monitoring record-keeping logbook, in reducing smoking tobacco expenditure amongst adult male household heads in rural Bangladesh.

The experiment was designed as a single-blind clustered randomised controlled trial utilising two-stage random sampling. A total of 650 adult male household heads were sampled from 16 chars (riverine islands) from Gaibandha, Northern Bangladesh, with eight chars in treatment and control groups each, between November 2018 and January 2019. The intervention consisted of a logbook to record daily smoking tobacco intake for 4 weeks provided only to participants in treatment chars (n = 332) while households in control chars received nothing (n = 318).

Final analysis was conducted using 222 and 210 households in the treatment and control chars respectively. The logbook intervention had a significant impact (P-value = 0.040) on reducing daily tobacco expenditure by 14% (α = 95%; CI: −0.273, −0.008) for the treatment group relative to the control group based on a difference-in-difference estimator. This is equivalent to a reduction of 20 cigarettes or 140 bidis smoked in a month.

Our minimal contact intervention successfully induced a reduction in smoking tobacco expenditure, which could effectively be incorporated with existing programs in the char regions of Bangladesh.