Choosing a cutoff: sensitivity vs specificity
- Concept
Moving the value that separates positive from negative trades false negatives for false positives.
What it is
Many tests give a number (a blood marker, a score) and someone decides above which value the result counts as positive. Because values in healthy and diseased people overlap, lowering the cutoff finds more diseased people (higher sensitivity) at the cost of flagging more healthy people (lower specificity); raising it does the opposite. The choice depends on which error is more serious in each setting.
Worked example
- Data: a blood marker is measured in 100 diseased and 100 healthy people. High values suggest disease.
- Cutoff at 60 points: 80 diseased test positive and 92 healthy test negative. Sensitivity 80%, specificity 92%.
- Cutoff at 40 points: 95 diseased test positive and 70 healthy test negative. Sensitivity 95%, specificity 70%.
- Moving from 60 down to 40: false negatives fall from 20 to 5 and false positives rise from 8 to 30.
- Reading: the 40 cutoff suits screening (it misses 5 diseased people); the 60 cutoff suits confirmation better (8 false positives).
The two cutoffs in the example (100 diseased and 100 healthy)
| Cutoff | Sensitivity | Specificity | False negatives | False positives |
|---|---|---|---|---|
| 40 points | 95% | 70% | 5 | 30 |
| 60 points | 80% | 92% | 20 | 8 |
Markers
- cutoff
- Se
- Sp
In practice
- A laboratory's reference range is a cutoff: it defines who falls outside the range.
- Some guidelines use two cutoffs: below one the disease is ruled out, above the other it is ruled in, and the zone in between is investigated further.
Common pitfalls
- Believing a new cutoff improves the test overall: it only changes the type of error.
- Choosing the cutoff in the same data used to report performance, which overestimates it.
- Forgetting that if disease gives low values, the direction is reversed.
Test yourself
Think of the answer, then tap the question to see it.
If diseased people have high values, what happens when the cutoff is raised?
Specificity rises and sensitivity falls: fewer false positives and more false negatives.
Which cutoff suits screening for a serious, treatable disease?
One that favours sensitivity, even if false positives increase and are confirmed later.
Why is there no cutoff that removes both false positives and false negatives?
Because values in healthy and diseased people overlap; any cutoff leaves errors on one side.
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Sensitivity, Specificity, and Predictive Values: Foundations, Pliabilities, and Pitfalls in Research and Practice (Frontiers in Public Health, 2017); Receiver operating characteristic curve: overview and practical use for clinicians (Korean Journal of Anesthesiology, 2022).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.