Author Affiliations: Health Outcomes Research Group, Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York.
The study conducted by Emerson et al1 provides strong evidence of positive outcome bias, yet does not fully address how to approach the problem. Making reviewers and authors more aware of the problem of positive outcome bias, as Emerson et al1 suggest, could increase the number of articles published with null findings. But adding a counteracting bias to the already present positive outcome bias might lead to over- or undercompensation. Theoretically, one could only remove (as opposed to counteract) positive outcome bias by blinding reviewers to the outcomes of studies. A practical version of this would be to semiblind.
Mirkin JN, Bach PB. Outcome-Blinded Peer Review. Arch Intern Med. 2011;171(13):1213-1214. doi:10.1001/archinternmed.2011.56