|
Bias is generally defined as any effect or inference tending to
produce results that depart systematically from true values. Bias
can occur at any stage of investigation, from study design to
publication of the results. It is possible (and desirable) to
design a bias-free study , but because studies are carried out
in the real world with practical and ethical constraints, it is
virtually impossible to conduct a study without bias.
More than
100 biases have been described (Sackett,
1979). Some of these can substantially affect symptom
study in advanced disease.
|