EBP 2 week 2.1 diagnostic accuracy studies measurement properties

when would you use sensitivity?

• you do NOT want to miss pathology

• screening/ruling out matters

when would you use specificity?

• confirming a diagnosis (rule in)

• reducing false positives

when would you use likelihood ratios?

• making real clinical decisions

• estimating how much certainty changes

• choosing th most useful special test

when would you use predictive values?

• prevalence in YOUR setting matters

likelihood ratio

combination od sensitivity and specificity uncf

function of likelihood ratio

indicates by how much a given diagnostic test result will raise or lower the pretest probability of the specific disorder/ hypothesis

positive likelihood ratio

favoring or giving confidence in disorder given a positive test

positive likelihood ratio calculation

(sensitivity)/ (1- specificity)

negative likelihood ratio

favoring of decreasing confidence in disorder given a negative test

negative likelihood ratio calculation

(1-sensitivity)/(specificity )

large +LR (>5)

• odds favoring the diagnosis given positive test

  • helpful ruling IN condition

small -LR (<3)

• reduce odds favoring diagnosis given negative test

• helpful for ruling OUT condition

pre-test probability

chance (%) that a patient has condition before testing

pre-test odds

the ratio having the condition vs not having it before testing

pre-test odds calculation

pre-test= prevalence /(1- prevalence)

posttest odds

the odds that the patient has the target disorder, after the test results are known

posttest odds calculation

pretest odds x LR

posttest probability

the probability of the target disorder AFTER a diagnostic test result known

posttest probability calculation

(posttest odds/ [posttest odds + 1])

Receiver operating characteristic curve (ROC)

uses sensitivity and specificity across range of values