Overview
Youden's index, J, sums up a diagnostic test's performance in one number, combining sensitivity and specificity. It's often used to choose the best cutoff for a test.
Enter the sensitivity and specificity as percentages.
Worked examples
| Inputs | Result |
|---|---|
| Sensitivity (%)91Specificity (%)84 | Youden index J: 0.75Balanced accuracy: 87.5% |
Show 2 more examples
| Sensitivity (%)80Specificity (%)80 | Youden index J: 0.6Balanced accuracy: 80% |
| Sensitivity (%)99Specificity (%)60 | Youden index J: 0.59Balanced accuracy: 79.5% |
Youden index J by specificity
With the calculator's starting values, the youden index j is 0.75. The table shows how it changes with the specificity.
| Specificity (%) | Youden index J |
|---|---|
| 60% | 0.51 |
| 80% | 0.71 |
| 84%starting value | 0.75 |
| 100% | 0.91 |
Other inputs: Sensitivity: 91%. Change them in the calculator above for your own numbers.
How it works
J = sensitivity + specificity − 1. It runs from 0, no better than chance, to 1, a perfect test. Balanced accuracy, their average, is shown too. The best cutoff on a ROC curve is often the one with the highest J.
The formula
Practical tips
- 91% sensitivity and 84% specificity give J = 0.75.
- J treats false positives and false negatives as equally bad, which isn't always right.
- A test with J near 0 is no better than guessing.
Frequently asked questions
What's a good Youden index?
Above about 0.5 is useful, and above 0.8 strong, depending on the field.
Can J be negative?
Yes, if the test is worse than chance, which means the results are flipped.
How do I use it to pick a cutoff?
Calculate J at each possible cutoff and choose the highest.