Input Values 

Use this to calculate positive
and negative likelihood ratios
for a test of given
sensitivity and
specificity and to calculate
the posttest probability of infection, given the
pretest probability and test performance. For these calculations, the pretest
probability of infection is the estimated of probability of infection based
on clinical signs and any other (such as previous test results) available information.
If no other information is available the estimated prevalence of infection in the
population of origin should be used. The posttest probability of infection
is the same as the positive
(or negative) predictive value for the test for the specified
pretest probability of infection.
Inputs required
 Test unit sensitivity;
 Test unit specificity; and
 Prior (pretest) probability of infection.
Outputs
 Table of pretest odds of infection, likelihood ratio, posttest odds of infection and
posttest probability of infection for positive and negative test results, respectively; and
 Plots of positive and negative likelihood ratios for varying sensitivity and specificity values
and varying prior probability of infection.
Formulae
LRpos = Se / (1  Sp)
LRneg = (1  Se) / Sp
PreTestOdds = p / (1  p)
PostTestOddsPos = PreTestOdds x LRpos
PostTestOddsNeg = PreTestOdds x LRneg
PInfPos = PostTestOddsPos / (1 + PostTestOddsPos)
PInfNeg = PostTestOddsNeg / (1 + PostTestOddsNeg)
where:
 p = Prior probability of infection;
 Se = Test unit sensitivity;
 Sp = Test unit specificity;
 LRPos = Likelihood ratio for a positive result; and
 LRNeg = Likelihood ratio for a negative result.
 PreTestOdds = Pretest odds of infection;
 PostTestOddsPos = Posttest odds of infection for a positive result;
 PostTestOddsNeg = Posttest odds of infection for a negative result;
 PInfPos = Probability of infection for a positive result; and
 PInfNeg = Probability of infection for a negative result.
References
Likelihood ratios
Likelihood ratios: pre and post test probability
Resources
Excel example
R function
Likelihood ratio calculator
