Differences in field effectiveness and adoption between a novel automated chlorination system and household manual chlorination of drinking water in Dhaka, Bangladesh: a randomized controlled trial

Abstract

The number of people served by networked systems that supply intermittent and contaminated drinking water is increasing. In these settings, centralized water treatment is ineffective, while household-level water treatment technologies have not been brought to scale. This study compares a novel low-cost technology designed to passively (automatically) dispense chlorine at shared handpumps with a household-level intervention providing water disinfection tablets (Aquatab), safe water storage containers, and behavior promotion. Twenty compounds were enrolled in Dhaka, Bangladesh, and randomly assigned to one of three groups: passive chlorinator, Aquatabs, or control. Over a 10-month intervention period, the mean percentage of households whose stored drinking water had detectable total chlorine was 75% in compounds with access to the passive chlorinator, 72% in compounds receiving Aquatabs, and 6% in control compounds. Both interventions also significantly improved microbial water quality. Aquatabs usage fell by 50% after behavioral promotion visits concluded, suggesting intensive promotion is necessary for sustained uptake. The study findings suggest high potential for an automated decentralized water treatment system to increase consistent access to clean water in low-income urban communities.

Fig 1. Schematic of passive chlorinator.

The device consists of (1) 10-liter chlorine reservoir (plastic jerrycan); (2) plastic flexible tubing (1/8 inch inner diameter) connecting the reservoir to the injection point; (3) a regulator to control the chlorine dose; (4) a brass non-return valve placed within the pipe, below the handpump and underneath the injection point; (5) two non-return valves at each end of the tubing (one shown located at the injection point and one hidden located just inside the chlorine reservoir outlet); and 6) a plastic funnel inserted into the pipe below the handpump to direct the flow past the injection point. Side view and top view show pipe cross-section with chlorine injection valve and funnel positioning.

Fig 2. Data collection schedule for households (hh) with children under five years of age in study compounds.

Solid black circles indicate water sampling for microbial and chlorine residual analysis, open circles indicate water sampling for chlorine residual analysis only. The duration of the follow-up data collection period (post-baseline and hardware delivery) was 10 months. Interviews were conducted at baseline and at five months of follow-up.

Fig 3. Proportion of samples from handpumps fitted with a passive chlorinator delivering water with detectable total chlorine residual (>0.1mg/L) over a 10-month follow-up study period.

Error bars show 95% confidence interval.

Fig 4. Proportion of households with detectable total chlorine residual in their stored drinking water (>0.1mg/L) by treatment group spanning a 10-month follow-up period.

Behavior promotions ended after month 4. Control group was not sampled past month 5. Error bars show 95% confidence interval of mean proportions.

Fig 5. Proportion of households with detectable E. coli contamination in their stored drinking water by treatment group spanning a 5-month follow-up period.

Error bars show 95% confidence interval of mean proportions.