Assessing the Health Impact of Water Quality Interventions in Low-Income Settings: Concerns Associated with Blinded Trials and the Need for Objective Outcomes

• Type of Material: Article
• Authors: Thomas Clasen, Emory University; Sophie Boisson, London School of Hygiene & Tropical Medicine
• Language: English
• Date: 2016-07-01
• Publisher: National Institute of Environmental Health Sciences (NIEHS)
• Publication Version: Final Published Version
• Copyright Statement: Publication of EHP lies in the public domain and is therefore without copyright.
• Final Published Version (URL): http://dx.doi.org/10.1289/ehp.1510532
• Title of Journal or Parent Work: Environmental Health Perspectives
• ISSN: 0091-6765
• Volume: 124
• Issue: 7
• Start Page: 886
• End Page: 889
• Research Categories: Health Sciences, Epidemiology; Health Sciences, Public Health; Environmental Sciences

Abstract

Background: A dramatic disparity between the results of blinded versus open trial designs has raised questions about the effectiveness of water quality interventions and other environmental interventions to prevent diarrhea, a leading killer of young children in low-income countries. Objectives: We summarize the results of blinded versus open trials of water quality interventions, describe evidence from a recent placebo-controlled trial in India suggesting that control households were put at risk from their participation, and suggest alternatives to blinded trials that could resolve continued uncertainty about the magnitude of the protective effect of water, sanitation, and hygiene (WASH) interventions without presenting ethical questions. Discussion: Concerns about reporting bias in open trial designs continue to cause uncertainty about the effectiveness of WASH interventions. However, evidence suggests that despite instructions to the contrary, placebos may encourage control group participants in blinded trials to cease practicing traditional water treatment practices in the mistaken belief that they are protected by an active intervention. Although objective outcomes such as pathogen incrimination, seroconversion, biomarkers, and anthropometry can be helpful, these are often costly, nonspecific, and unsuitable for evaluating programmatic interventions. Conclusions: Unless researchers can be assured that a placebo will not cause those in a control group to change their behavior in a manner that increases their risk, it is incumbent on researchers to use alternatives. Validated objective measures are needed for assessing the health impact of WASH interventions that are reliable, affordable, and suitable both for research and program evaluation.