Table 1:

Theoretical framework of how disease prevalence and test accuracy may be related10

FactorEffect on prevalenceEffect on accuracy
Clinical variability
Patient spectrum
  • Distribution of symptoms and severity may change with varying prevalence

  • Differences in symptoms and severity influences sensitivity and specificity

Referral filter
  • How and through what care pathway patients are referred may influence the spectrum of disease in the population

  • A change in setting and patient spectrum may also alter a test’s sensitivity and specificity

Reader expectations
  • Prevalence influences reader expectations: if one knows that the prevalence should be high, then one’s intrinsic threshold may be lowered

  • Changing one’s intrinsic threshold will influence accuracy

Artifactual variability
Distorted inclusion of participants
  • Excluding patients with difficult to diagnose conditions may influence the prevalence

  • Excluding patients with difficult to diagnose conditions will overestimate the accuracy of a test

Verification bias
  • If not all patients receive the (same) reference standard, this influences prevalence

  • Verification bias has an effect on test accuracy

Imperfect reference standard
  • Prevalence will be over- or underestimated

  • Test accuracy may be underestimated; the extent of which varies with prevalence