What is Sensitivity vs Specificity?

It’s important to know the validity of a diagnostic test and understand whether the test actually measures what it was designed to measure. Learn the basics of sensitivity vs specificity, positive and negative predictive values, and likelihood ratios—vital indicators of accuracy of a test in diagnostics that provide information on risk stratification, allowing clinicians to best treat their patients.

What is sensitivity?
Sensitivity is the ability of a test to correctly identify people who have a given condition—the percentage of true positives. If the true status of the condition cannot be known, sensitivity and specificity can be defined relative to a "gold standard test".

How to calculate sensitivity?
It is calculated by identifying the number of people who test positive for a condition and dividing it by the total number of people with the condition, which includes those who tested positive and the false negatives—those who tested negative but actually had the disease.

For example, if a test has a sensitivity of 88%, it means that 88 of every 100 people with the disease actually test positive for the disease. But 12 of the 100 people test negative, false negatives, even though they actually have the condition.

Let’s say the gold standard test for a disease we’ll call Aetrionixia is TEST A.

In this case, 100 people have Aetrinoixia and test positive—the true positives. 135 people (100 + 35) actually have the disease.

What is specificity?
Conversely, specificity is the ability of that test to identify people who do not have the disease or condition—the percentage of true negatives.

How to calculate specificity?
It is calculated by identifying the number of people who test negative for a condition and dividing it by the total number of people without the condition, which includes those who tested negative and the false positives—those who tested positive but did not have the disease.

For example, if a test has a specificity of 95%, it means that 95 of every 100 people who test negative for the condition actually do not have the condition. But there are 5 people who test positive who are really negative for the condition.

Using the data for TEST A and Aetrionixia:

Positive and Negative Predictive Value Meaning
Like sensitivity, positive predictive value (PPV) deals with the number of people who test positive—but in this case, it’s how many true positives are there in the pool of all of those who tested positive.

For TEST A and Aetrionixia:

Similarly, negative predictive value (NPV) deals with the number of people who test negative—how many true negatives there are in the pool of all of those who tested negative.

One important point to consider is that sensitivity and specificity are inversely related—one increases as the other decreases—but they are generally considered stable for a given test. Positive and negative predictive values inherently vary with pre-test probability (e.g., changes in population disease prevalence).

Likelihood Ratios
Likelihood ratios (LRs) are also used to help interpret diagnostic tests. The likelihood ratios summarize the information from the sensitivity and specificity testing to provide information on how likely it is that a patient with a given test result has or does not have a condition.

There are two types of LRs:

• positive likelihood ratio (LR+)
• negative likelihood ratio (LR-)

What are positivity likelihood ratios?
Positive likelihood ratios describe the probability that a patient WITH the condition tests positive divided by the probability that a patient WITHOUT the condition tests positive. Essentially, the true positives divided by the false positives.

What are negative likelihood ratios?
Negative likelihood ratios describe the likelihood that a patient WITH the disease will test negative as compared with patients without the disease. Essentially, the false negatives divided by the true negatives.

Likelihood ratio values farther away from 1 indicate a stronger probability (LR- >0.1; LR+ <10).

Likelihood ratios are determined by dividing the probability that a person with a condition has a given test result by the probability that a person who does not have the condition has that same test result. They are rooted in sensitivity and specificity:

Using Aetrionixia and TEST A, LR+:

Thus, someone with Aetrionixia has a 4.11 times greater likelihood of being positive for TEST A than someone without Aetrionixia. Higher numbers indicate a greater likelihood of having the condition and a more accurate test.

And LR-:

Thus, someone without Aetrionixia is ~3 times more likely of being negative for TEST A than someone with Aetrionixia. In this case, lower numbers indicate a lesser likelihood of having the condition and a more accurate test.

Sensitivity, specificity, positive and negative predictive values, and likelihood ratios are all valuable when trying to understand test results. Patients, of course, want to know whether they have a given disease—to answer their questions, accurate testing is key.