HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Gerhard, T. Search for Related Content PubMed PubMed Citation Articles by Gerhard, T.

TOBIAS GERHARD, PH.D., is Associate Director, Center for Education and Research on Mental Health Therapeutics, Institute for Health, Health Care Policy, and Aging Research, and Assistant Professor, Ernest Mario School of Pharmacy, Rutgers University, 30 College Avenue, New Brunswick, NJ 08901 (tgerhard{at}rci.rutgers.edu).

Purpose. The concept of bias and its effects on study results are reviewed. Summary. Understanding bias is paramount for the conduct of sound research studies. In the context of research methodology, bias refers to the presence of systematic error in a study. This report serves as an introduction to the concept of bias and describes three major areas through which bias is introduced into research studies of health care interventions: (1) factors that relate to the exposure of patients to treatments in the population, (2) factors that influence inclusion of patients in the study, and (3) factors related to assessment and measurement. Methods to address bias in both design and analysis stages of a study are presented, including restriction, matching, multivariate analysis, instrumental variable estimation, and the prevention of misclassification of information. While some biases are specific to certain study designs and approaches, the general concepts of how bias is introduced into a study apply equally to randomized controlled trials, prospective observational or quasiexperimental studies, and retrospective studies. An important difference between the designs is how much opportunity the investigator has to avoid bias and influence accuracy of measurements, ranging from ideally full control in randomized controlled trials to almost no control in retrospective studies of automated databases. Conclusion. Investigators should aim to avoid bias in the design of a study, adjust for bias in the study analysis if bias cannot feasibly be avoided, and quantify and discuss the effects of residual bias on study results. Index terms: Clinical studies; Errors; Methodology; Quality assurance; Research