BIAS

Why is bias important  ?

• You might have a knowledge that something appears to work, and that it might be useful in foods or in health or disease or any other way. But how confident are you that your knowledge is based on good science and not on biased observations?
• 'Assessment of the risk of bias'... and 'analysis of bias' are two different approaches that can provide assurance.

Assessment of the risk of bias

• Applicable to individual studies
• This involves, making judgments about:
• the method of sequence generation
• the method of allocation concealment
• the methods of blinding
• the completeness of outcome data
• the selectivity of reporting
• other possible sources of bias (see below)

Risk of bias analysis

• Applicable to multiple independent studies
• This involves quantifying the extent and probability of a source of bias
• Biases may be categorised in five common domains
• Selection bias
• Performance bias
• Attrition bias
• Detection bias
• Reporting bias
• Sources of bias assessed at INLogic Ltd have included:
• baseline data non-acquisition and inequality
• non performance of randomisation
• use of a biased randomisation method
• absence of double or triple blinding
• adequacy of placebo
• participant drop-outs
• adequate explanation of drop-outs
• adequacy of study design
• inclusion of adverse effects
• occurrence of reporting bias
• use of language
• source of funding
• setting, participants  independent of funding
• publication bias
• productivity bias

Methods and graphics employed during the assessment of bias

• Meta-analysis
• Meta-regression
• Meta-modeling
• Direct or indirectly:
• Fill and trim analyses
• Funnel plots
• Galbraith plots
• Standard plots
• Bubble plots