CH. 14 - sensitivity/specificity
Specificity Formula
TN (d) ---------- TN (d) + FP (b)
sensitivity formula
TP (a) ---------- TP (a) + FN (c)
Specificity
• Of the individuals who are truly *NEGATIVE*, what proportion are tested as *negative* using the screening tool (test) • probability that a test correctly classifies individuals w/p preclinical disease as NEGATIVE • High specificity = fewer people are given FALSE positive results • d/(b+d)
Sensitivity
• Of the individuals who are truly *POSITIVE*, what proportion are tested as *positive* using the screening tool (test) • *probability that a test correctly classifies as POSITIVE those who have preclinical disease* • High sensitivity = fewer people are given FALSE negative results • a/(a+c)
negative predictive value
• Of those who screen negative, how many are actually negative • Affected by the number of false negatives *d/(d+c)*
positive predictive value
• Of those who screen positive, how many are actually positive • Measure of positive yield • *As prevalence increases so does the PPV* *a/(a+b)*