Likelihood ratios & area under the curve
Understanding likelihood ratios is part of the mastery of a diagnostic work up. Likelihood ratio help explain the significance of each test we perform and the difference they make to our patients.
The area under the curve is used during the development of a diagnostic test. The more accurate the test the better it separates those with the disease and those without it and the area under the curve can be used to quantify this.
This podcast covers
- Positive and negative likelihood ratios
- Fagan’s nomogram
- Categorical Likelihood ratios
- Area under the curve
Categories of likelihood ratios; Goodacre 2009 EMJ
| Likelihood ratio | Diagnostic value |
| 1 | None at all |
| 0.5 | Little clinical significance |
| 2 to 5 | Moderately increases likelihood of disease |
| 5 to 10 | Markedly increases likelihood of disease |
| 0.1 to 0.2 | Markedly decreases likelihood of disease |
| Over 10 | Diagnostic |
| Less than 0.1 | Rules out the disease |
An example of a Fagan’s nomogram By Mikael Häggström [Public domain], via Wikimedia Commons
![An example of a Fagan’s nomogram By Mikael Häggström [Public domain], via Wikimedia Commons](https://upload.wikimedia.org/wikipedia/commons/6/66/Fagan_nomogram.svg){kind=link}
Significance of the area under a curve
| Area under the curve | Significance |
| .90-1 | Excellent |
| .80-.90 | Good |
| .70-.80 | Fair |
| .60-.70 | Poor |
| .50-.60 | Fail |
Once you’ve listened help consolidate the topics covered with the questions below.