Table 1.

Types of bias

Type of bias	Description	Example	Prevention
Selection bias	Some participants are more likely to be selected for a study. Included participants are not representative of the population	Unemployed people more likely to participate in a time-consuming study	Allocation concealment, sequence generation
Detection bias	A certain condition is more likely to be detected in a subgroup of participants due to systematic differences in how outcomes are determined	Detection of appendicitis by ultrasound in thin versus obese patients	Blinding of outcome assessment
Reporting bias	Positive results and correlations are more likely to be reported	Non-finding or negative finding is not published	Preemptive determination of outcomes of interest
Exclusion bias	A certain population is more likely to be excluded from a study	Pregnancy, vulnerable patients such as small children or elderly are not included	Preemptive definition of exclusion criteria and consideration of these during discussion of the results found
Attrition bias	Loss of follow-up of a certain subgroup of participants	Elderly people not reachable via email	Reporting of incomplete outcome data, intention to treat analysis
Performance bias	Systematic differences between the groups regarding the exposure or care other than the intervention	Group receiving a drug gets more frequent blood examinations	Double blinding

Adapted from [38]