Sensitivity, specificity, positive predictive value, and negative predictive value are core measures used to judge how well a medical test performs. They help students and clinicians decide whether a test is good for screening, confirming disease, or reassuring a patient after a negative result. These ideas connect statistics directly to patient care, because the same test can be useful in one setting and misleading in another.
Understanding them prevents common errors in interpreting lab tests, imaging, and bedside screening tools.
Sensitivity and specificity describe the test itself by comparing results to the true disease status, while PPV and NPV describe what a positive or negative result means for the patient in a specific population. The standard framework is the 2 x 2 table with true positives, false positives, true negatives, and false negatives. Disease prevalence strongly affects PPV and NPV, even when sensitivity and specificity stay the same.
In practice, highly sensitive tests are often used to rule out disease, and highly specific tests are often used to rule in disease.
Understanding Biostatistics Essentials
A test result begins with a measurement, then a decision rule. A blood marker, scan score, or swab signal may vary continuously from low to high. Someone chooses a cutoff point that labels values as positive or negative.
Moving that cutoff changes the mistakes the test makes. A lower cutoff catches more people who truly have the condition, but it labels more healthy people as positive.
A higher cutoff reduces unnecessary positive results, but it can miss people with mild or early disease. This tradeoff is important when a missed case could cause serious harm, such as an infection that needs urgent treatment.
Numbers become clearer when placed in a real group. Imagine testing one thousand people where only ten truly have a disease. Suppose the test detects nine of those ten people and correctly clears ninety five percent of people without disease.
About fifty healthy people will still receive a positive result. The clinic now has roughly fifty nine positive results, yet only nine are true cases. A positive result in this setting needs follow-up, not an immediate conclusion.
In contrast, a negative result is highly reassuring because almost everyone tested did not have the disease. This is why the same test can feel very different in a specialist clinic, where illness is common, than in a school screening programme, where it is rare.
Doctors often use tests in sequence to manage this problem. An initial test may be chosen because it is quick, cheap, or unlikely to miss a dangerous condition. Positive results may then be checked with a more precise test.
Pregnancy tests, rapid infection tests, newborn screening, and some cancer screening programmes follow this general pattern. A positive screening result does not automatically mean a diagnosis. It means the person belongs in the smaller group that needs further assessment.
Timing matters too. Testing very early in an illness may give a negative result because the marker has not risen enough to detect. Poor sample collection, damaged equipment, and human reading errors can create mistakes that published test statistics do not fully capture.
When studying these measures, pay close attention to the group named in the denominator. To judge how well positives predict disease, start with all positive results. To judge how well a test finds disease, start with all people who truly have disease.
Mixing up these starting groups is the most common error. Draw a table and fill in actual counts before calculating percentages. Check whether the question describes a screening population, hospital patients, or people with symptoms.
These groups have different disease rates and may have different severities of illness. Finally, remember that no test result stands alone. Symptoms, medical history, examination findings, and the consequences of a wrong decision shape what clinicians do next.
Key Facts
- Sensitivity = TP / (TP + FN)
- Specificity = TN / (TN + FP)
- PPV = TP / (TP + FP)
- NPV = TN / (TN + FN)
- Accuracy = (TP + TN) / (TP + TN + FP + FN)
- Higher prevalence increases PPV and decreases NPV, even if sensitivity and specificity do not change.
Vocabulary
- Sensitivity
- Sensitivity is the proportion of people with the disease who test positive.
- Specificity
- Specificity is the proportion of people without the disease who test negative.
- Positive predictive value
- Positive predictive value is the probability that a person with a positive test truly has the disease.
- Negative predictive value
- Negative predictive value is the probability that a person with a negative test truly does not have the disease.
- Prevalence
- Prevalence is the proportion of a population that has the disease at a given time.
Common Mistakes to Avoid
- Confusing sensitivity with PPV, because sensitivity is based on people who truly have disease, while PPV is based on people who tested positive. These answer different clinical questions and cannot be used interchangeably.
- Ignoring prevalence when interpreting PPV and NPV, because predictive values change with how common the disease is in the tested population. A test can have the same sensitivity and specificity but very different PPV in a low-prevalence setting.
- Using a highly specific test to rule out disease, because specificity mainly helps confirm disease when the result is positive. To rule out disease, a highly sensitive test is usually more helpful.
- Reading percentages from a 2 x 2 table incorrectly, because the denominator must match the definition of the statistic. For example, sensitivity uses TP + FN, not all positive test results.
Practice Questions
- 1 A screening test is given to 200 people. Of the 50 who truly have disease, 45 test positive. Of the 150 without disease, 30 test positive. Calculate sensitivity and specificity.
- 2 In a clinic, a test gives the following results: TP = 36, FP = 12, TN = 108, FN = 4. Calculate PPV and NPV.
- 3 A disease is rare in Population A and common in Population B. The same test with the same sensitivity and specificity is used in both groups. Explain how PPV and NPV are expected to differ between the two populations and why.