Objectives: To determine whether installation of an ion-exchange water softener in the home could improve atopic eczema in children and, if so, to establish its likely cost and cost-effectiveness. Design: An observer-blind, parallel-group randomised controlled trial of 12 weeks duration followed by a 4-week observational period. Eczema was assessed by research nurses blinded to intervention at baseline, 4 weeks, 12 weeks and 16 weeks. The primary outcome was analysed as intent-to-treat, using the randomised allocation rather than actual treatment received. A secondary per-protocol analysis excluded participants who failed to receive their allocated treatment and who were deemed to be protocol violators. Setting: Secondary and primary care referral centres in England (UK) serving a variety of ethnic and social groups and including children living in both urban and periurban homes. Participants: Three hundred and thirty-six children (aged 6 months to 16 years) with moderate/severe atopic eczema, living in homes in England supplied by hard water (≥ 200 mg/l calcium carbonate). Interventions: Participants were randomised to either installation of an ion-exchange water softener plus usual eczema care (group A) for 12 weeks or usual eczema care alone (group B) for 12 weeks. This was followed by a 4-week observational period, during which water softeners were switched off/removed from group A homes and installed in group B homes. Standard procedure was to soften all water in the home, but to provide mains (hard) water at a faucet-style tap in the kitchen for drinking and cooking. Participants were therefore exposed to softened water for bathing and washing of clothes, but continued to drink mains (hard) water. Usual care was defined as any treatment that the child was currently using in order to control his or her eczema. New treatment regimens used during the trial period were documented. Main outcome measures: Primary outcome was the difference between group A and group B in mean change in disease severity at 12 weeks compared with baseline, as measured using the Six Area, Six Sign Atopic Dermatitis (SASSAD) score. This is an objective severity scale completed by blinded observers (research nurses) unaware of the allocated intervention. Secondary outcomes included use of topical medications, nighttime movement, patient-reported eczema severity and a number of quality of life measures. A planned subgroup analysis was conducted, based on participants with at least one mutation in the gene encoding filaggrin (a protein in the skin thought to be important for normal skin barrier function). Results: Target recruitment was achieved (n = 336). The analysed population included 323 children who had complete data. The mean change in primary outcome (SASSAD) at 12 weeks was –5.0 [standard deviation (SD) 8.8] for the water softener group (group A) and –5.7 (SD 9.8) for the usual care group (group B) [mean difference 0.66, 95% confidence interval (CI) –1.37 to 2.69, p = 0.53]. The per-protocol analysis supported the main analysis, and there was no evidence that the treatment effect varied between children with and without mutations in the filaggrin gene. No between-group differences were found in the three secondary outcomes that were assessed blindly (use of topical medications; nighttime movement; proportion showing reasonable, good or excellent improvement). Small, but statistically significant, differences in favour of the water softener were found in three of the secondary outcomes that were assessed by participants [Patient-Oriented Eczema Measure (POEM); well-controlled weeks (WCWs); Dermatitis Family Index (DFI)]. The results of the economic evaluation, and the uncertainty surrounding them, suggest that ionexchange water softeners are unlikely to be a cost-effective intervention for children with atopic eczema from an NHS perspective. Conclusions: Water softeners provided no additional benefit to usual care in this study population. Small, but statistically significant, differences were found in some secondary outcomes as reported by parents, but it is likely that such improvements were the result of response bias. Whether or not the wider benefits of installing a water softener in the home are sufficient to justify the purchase of a softener is something for individual householders to consider on a case-by-case basis. This trial demonstrated overwhelming demand for non-pharmacological interventions for the treatment of eczema, and this is something that should be considered when prioritising future research in the field. Trial registration: Current Controlled Trials ISRCTN71423189. Funding: This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 15, No. 8. See the HTA programme website for further project information. Results of this trial are also published at www.plosmedicine.org.