Table 1.

Bias in diagnostic test accuracy studies

Element	Type of Bias	When does it occur?	Under- or overestimation of diagnostic accuracy?
Patients	Spectrum bias	When patient inclusion does not represent the intended segment of target condition severity spectrum.	Depends on difference between targeted and included part of spectrum.
	Selection bias	When eligible patients are not enrolled consecutively or randomly.	Usually leads to overestimation
Index test	Information bias	When the index test results are interpreted with knowledge of the results of the reference standard, or with more (or less) information than in practice.	Usually leads to overestimation, unless less clinical information is provided than in practice, which may result in underestimation.
Reference standard	Verification bias	When the reference standard does not correctly classify patients with the target condition.	Depends on whether both tests make the same mistakes.
	Partial verification bias	When a nonrandom set of patients does not undergo the reference standard.	Usually leads to overestimation of sensitivity, effect on specificity varies.
	Differential verification bias	When a set of patients is verified with a second or third reference standard; especially when this selection depends on the index test result.	Variable.
	Incorporation bias	When the index test is incorporated in a (composite) reference standard.	Usually leads to overestimation.
	Time lag bias	When the target condition changes between administering the index test and the reference standard.	Under- or overestimation, depending on change in patients’ condition.
	Information bias	When the reference standard is interpreted knowing the index test results.	Usually leads to overestimation.
Data analysis	Result elimination bias	When uninterpretable or intermediate test results and withdrawals are not included in the analysis.	Usually leads to overestimation.