Table 2

Criteria for study validation.

Population: Consecutive recruitment of an appropriate spectrum of eligible patients will be considered ideal. Convenience sampling, i.e. arbitrary recruitment or non-consecutive recruitment will be deemed inadequate. In the absence of any explicit information on the method of recruitment, the article will be categorised as unclear reported population enrolment.

Diagnostic test: The description of the diagnostic test will be considered ideal if the methodology is reported together with the measurement parameter and the cut-off level for an abnormal result. In the absence of any of the above information in the manuscript, then the diagnostic intervention will be considered as unclear reported.

Outcome measures: Blinding will be considered ideal if it is clearly reported that the results of the various tests were not divulged. Information on the number of patients recruited into the study and those whose outcome data were known will also be sought from the manuscripts. Withdrawal of patients from the study, missing data and lack of outcome data outwith the study hospital will be categorised as lost to follow-up. In particular we will look for evidence of verification bias where the rates follow-up and confirmation of outcome are different in patients with positive test results compared to those with negative test results.

Any available randomised trials will be assessed for validity separately to the diagnostic accuracy studies considering factors associated with bias in such trials, e.g. concealment of randomisation, sequence generation, blinding and follow-up.


Bachmann et al. BMC Surgery 2002 2:2   doi:10.1186/1471-2482-2-2

Open Data