Biases in epidemiological studies: How far are we from the truth?

ABSTRACT In any research occurrence of ‘Bias’ is inevitable. Bias is an error that occurs in a systematic way during the design, implementation or interpretation of the study that deviates the results away from the actual facts. Biases are broadly of two types, i.e. selection bias and information bias. Selection bias occurs when the population selected for the study is not representative of the target population. Information bias occurs when there is a systematic difference in the way data is gathered from the subjects. Confounding is another factor which though not a bias may also cause deviation from truth. There are various methods of minimising bias like adhering to a strict protocol, good sampling methods, proper training of the researchers, double entry of the data and minimising follow up, whereas confounding can be minimised by randomisation, restriction, matching and statistical modelling. This article describes the various types of biases, their occurrence and how they can be minimised so that the research output is robust. Copyright © 2014, Indian Journal of Medical Specialities. Published by Reed Elsevier India Pvt. Ltd. All rights reserved.