Example: HIV - Part 1

Let's look at an example:

hiv

HIV/AIDS prevalence world map in 2009 according to UNAIDS data
Image source

 

A screening test is applied to 10’000 people known at HIV high-risk exposure (prevalence = 12%) using an ELISA- Antibody test confirmed by Western Plot (Se = 99.99%, Sp = 99.99%). The results of this study are presented in the cross table below:

  Disease +
Disease - Total
Test + 1199
 		1
 1200
Test - 1
 8799
 8800
Total 1200
 8800 10000

 

 

 

 

 

 

Prevalence = 12%

Positive predictive value :
PVV= P (HIV + | Test +) = 1199 / 1200 = 99.9%

Conclusion - in case of high disease prevalence:

  • there is a high probability of infection
  • the test result is likely true positive
  • patient has a high probability to be truly infected

Thus, if Peter belongs to a population with such a high prevalence of HIV, the confidence in his positive test result would be relatively high.

« Previous | Next »