|
- Verification bias
This form of bias is one of the most problematic.
Verification bias comes about when the subsequent workup is based
on the result of the technology being tested, such that the gold or reference standard differs among cases included in the study.
This occurs because the standard of evidence may require an invasive test (e.g. biopsy or removal of organ) which in some cases is
impractical and/or unethical to perform.
Typically, positive cases receive a more rigorous workup than the negative ones.
Some studies may allow negative results to be accepted as being true without further followup or with a less convincing standard.
As a result, the true-negative and false-negative rates are
inaccurately reported.
Therefore, it is recommended that in cases in which a common reference standard cannot be
adopted, that a very assiduous clinical workup be conducted during the period of time over which one would expect the disease course to occur.
- Example
Consider a study that is examining the accuracy of lung cancer screening by CT. A biopsy or surgery cannot be performed in patients with negative
CT results to determine his or her true disease status. For these patients, it is necessary to pursue further clinical workup and/or imaging follow-up for a specified period
(2 years) to detect wrongly diagnosed cases. If an assumption is made that in those patients with negative CT results are disease free, the test specificity
will be greatly overestimated.
|
