Publication bias occurs when the outcome of research influences the decision whether to publish it. In the context of epidemiology, this bias can skew the scientific literature, leading to a distorted understanding of the relationship between exposures and health outcomes.