Point-of-use water treatment (household water treatment, HWT) has been advocated as a means to substantially decrease the global burden of diarrhea and to contribute to the Millennium Development Goals. To determine whether HWT should be scaled up now, we reviewed the evidence on acceptability, scalability, adverse effects, and nonhealth benefits as the main criteria to establish how much evidence is needed before scaling up. These aspects are contrasted with the evidence on the effect of HWT on diarrhea. We found that the acceptability and scalability of HWT is still unclear, and that there are substantial barriers making it difficult to identify populations that would benefit most from a potential effect. The nonhealth benefits of HWT are negligible. Health outcome trials suggest that HWT may reduce diarrhea by 30-40%. The problem of bias is discussed. There is evidence that the estimates may be strongly biased. Current evidence does not exclude that the observed diarrhea reductions are largely or entirely due to bias. We conclude that widespread promotion of HWT is premature given the available evidence. Further acceptability studies and large blinded trials or trials with an objective health outcome are needed before HWT can be recommended to policy makers and implementers.