Test | Patients With Disease | Patients Without Disease |
Positive Result | a (true positive) | b (false positive) |
Negative Result | c (false negative) | d (true negative) |
How "reliable" is a test?
Example - High Incidence
Lets look at a sample of 1,000 patients
Test | Patients With Disease (300) | Patients Without Disease (700) |
Positive Result | 270 (true positive) | 70 (false positive) |
Negative Result | 30 (false negative) | 630 (true negative) |
Example - Low Incidence
Lets look at a sample of 1,000 patients
Test | Patients With Disease (10) | Patients Without Disease (990) |
Positive Result | 9 (true positive) | 99 (false positive) |
Negative Result | 1 (false negative) | 891 (true negative) |
Conclusion
Read this editorial, which gives insight into pretty much the only statistics SAQ that can be asked, after the 2013 curriculum revision - and which is also clinically relevant!
Predicting difficult intubation - worthwhile exercise or pointless ritual?
Yentis. Anaesthesia, 2002, 57, pages 105-109
e-journal available via the ANZCA Library