Sensitivity in research measures how well a test identifies people who truly have a condition. Expressed as a percentage, it answers a simple question: of everyone who actually has the disease, how many did the test correctly flag as positive? A test with 95% sensitivity catches 95 out of every 100 true cases. The remaining 5 are missed, called false negatives.
How Sensitivity Is Calculated
The formula is straightforward: divide the number of true positives by the total number of people who actually have the condition (true positives plus false negatives), then multiply by 100 to get a percentage.
Sensitivity = True Positives / (True Positives + False Negatives) × 100%
A “true positive” is someone who has the disease and correctly tests positive. A “false negative” is someone who has the disease but the test incorrectly calls them healthy. Sensitivity focuses entirely on people with the condition. It says nothing about how the test performs on healthy people.
Why High Sensitivity Matters for Screening
When researchers design screening tests, they want to catch as many real cases as possible. A highly sensitive test rarely misses someone who is sick, which means few false negatives slip through. This is especially important for serious conditions where a delayed diagnosis changes outcomes. In breast cancer screening, for example, a missed case (false negative) can mean the difference between catching a tumor early and discovering it only after symptoms appear, when treatment options may be narrower and recovery harder.
A negative result from a highly sensitive test carries real reassurance. If the test catches 98% of true cases, a negative result makes it very unlikely you have the condition. This is the core principle behind using sensitive tests as a first-pass screen: they’re good at ruling things out. Clinicians sometimes summarize this with the mnemonic “SnNOut,” meaning a highly Sensitive test with a Negative result helps rule Out disease.
The Trade-Off With Specificity
Sensitivity and specificity pull in opposite directions. When you increase one, the other tends to decrease. Specificity measures how well a test identifies people who don’t have the condition. A highly specific test rarely labels a healthy person as sick (few false positives), while a highly sensitive test rarely misses a sick person (few false negatives).
Imagine a test that uses a blood marker to detect a disease. If researchers lower the threshold for a positive result, the test will catch more true cases, boosting sensitivity. But that same lower threshold also flags more healthy people as positive, dropping specificity. Researchers choose where to set this threshold based on the consequences. For a deadly but treatable disease, they lean toward high sensitivity because missing a case is worse than a false alarm. For conditions where a false positive leads to invasive follow-up procedures, they may prioritize specificity.
Why Sensitivity Stays Stable Across Populations
One useful property of sensitivity is that it remains stable regardless of how common or rare a disease is in a given population. This sets it apart from another measure called positive predictive value (PPV), which tells you the probability that a positive test result is correct.
Positive predictive value shifts dramatically with disease prevalence. In a population where a disease is common, a positive result is more likely to be a true positive. In a population where the disease is rare, even a good test produces more false alarms relative to true cases, and the positive predictive value drops. Sensitivity and specificity don’t have this problem. They are considered inherently stable features of a test, which is why researchers report them as the primary measures of test performance and why they’re comparable across studies conducted in different populations.
Sensitivity Analysis: A Different Use of the Term
Researchers also use the word “sensitivity” in a completely different context: sensitivity analysis. This has nothing to do with diagnostic tests. Instead, it’s a method for checking whether the conclusions of a study hold up when you change the assumptions behind it.
In a clinical trial, for instance, researchers make decisions about how to handle missing data, which patients to include, and what statistical model to use. A sensitivity analysis re-runs the analysis under different plausible scenarios to see if the results change. If the main finding holds across a range of assumptions, it’s considered robust. If it flips when one assumption changes, that’s a signal the conclusion may be fragile. This process typically involves three steps: selecting a set of plausible scenarios, computing the study’s performance under each one, and evaluating whether the results remain consistent.
When you see “sensitivity analysis” in a research paper, it’s the authors stress-testing their own work. It doesn’t produce a single percentage like diagnostic sensitivity does. Instead, it produces a range of outcomes that help readers judge how much confidence to place in the findings.
How Sensitivity Is Reported in Practice
Regulatory agencies like the U.S. Food and Drug Administration expect diagnostic test manufacturers to report sensitivity alongside specificity when submitting evidence for approval. The FDA recommends reporting both measures with 95% confidence intervals, which show the range within which the true sensitivity likely falls. A test reporting 83.5% sensitivity with a confidence interval of 74.3% to 93.2% means researchers are 95% confident the real sensitivity lies somewhere in that range.
No universal minimum sensitivity is required for FDA approval. The acceptable threshold depends on the condition, the intended use of the test, and what alternatives exist. A rapid screening test used in an emergency department might be held to a different standard than a confirmatory lab test. What matters is that the reported sensitivity is honest, well-documented, and appropriate for the clinical context.
When reading a study, look at sensitivity alongside specificity rather than in isolation. A test with 99% sensitivity but 30% specificity catches nearly every case but also flags a large number of healthy people. The best diagnostic tests balance both measures well enough for their intended purpose, and researchers choose cut-off points that reflect the real-world consequences of getting it wrong in either direction.