Research and scholarly methods: Mitigating information bias

Abstract

Whether designing a research study or considering changes to a patient's medication therapy based on published research, pharmacists must be knowledgeable about potential sources of bias in clinical research and how these threats to internal validity can be mitigated. They must be confident that the study results can be relied upon and represent the “truth” prior to generalizing the results to clinical practice. Bias can occur in the design, conduct, or analysis of a research study. The objective of this paper is to help pharmacists identify information bias, which is particularly important in observational research studies. Several common types of information bias are presented, including misclassification, interviewer/observer, self-reporting/recall, surveillance, lead-time, and immortal time. Examples from the medical literature and mitigation strategies are included to help pharmacists evaluate threats to internal validity.