Sensitivity and Specificity

#Sensitivity concerns detecting the condition

Clinical sensitivity is the proportion of people with a defined condition who receive a positive test result. A highly sensitive test misses relatively few affected people in the setting where its performance was measured. A negative result in someone who has the condition is called a false negative.

Sensitivity is not the probability that a person with a positive result has the condition. It begins with the group known to have the condition and asks how many the test detects. Performance can differ with disease stage, sample timing and the characteristics of the people tested.

#Specificity concerns excluding the condition

Clinical specificity is the proportion of people without the defined condition who receive a negative result. A highly specific test produces relatively few false positive results in that group. A false positive means the test signals the condition even though the comparison assessment classifies it as absent.

Both measures rely on a way of determining who truly has the condition, often called a reference standard. That comparison can itself be imperfect. Study design, participant selection and differences in how tests are performed can influence the reported estimates, so a single performance figure is not a universal guarantee.

#Thresholds create trade-offs

For many numerical tests, a threshold separates positive from negative results. If higher values indicate disease, lowering the positive threshold generally increases sensitivity but reduces specificity. More affected people are detected, but more unaffected people also receive positive results. Raising the threshold generally reverses that trade-off.

The preferred balance depends on the purpose of testing and the consequences of each type of error. Follow-up testing may help clarify an initial signal. To understand what a positive or negative result means for someone being tested, the starting likelihood of the condition must also be considered.

#Common misunderstandings

Sensitivity and specificity do not directly tell someone how likely they are to have a condition after receiving a test result. Sensitivity starts with people who have the condition and asks how often the test is positive. Specificity starts with people who do not have it and asks how often the test is negative.

A highly sensitive test can still produce false positives, and a highly specific test can still miss cases. Neither measure, by itself, means a test is “accurate” for every purpose.

The likelihood that a positive result represents the condition also depends on how common the condition is in the population being tested. When a condition is uncommon, false positives can make up a substantial share of positive results.

Published estimates are not universal guarantees. Results can differ with the population studied, how samples are collected, and the reference standard used to establish who has the condition.

#Questions worth asking a clinician

• How sensitive and specific is this test for the condition you are checking for?
• If I have the condition, how likely is this test to miss it?
• If I do not have the condition, how likely is this test to give a positive result?
• How does my chance of having the condition before testing affect what a positive or negative result means?
• Would I need another test to confirm the result, and why?