By Fatima Ahmad Qureshy
The ongoing pandemic has made us more aware of epidemiology and its principles. From healthcare departments, to media, to governments- we have seen a widespread use of biostatistics and epidemiologic studies, for better quality of public health.
According to Wikipedia:
Epidemiology is the study and analysis of the distribution, patterns and determinants of health and disease conditions in defined populations. It is a cornerstone of public health, and shapes policy decisions and evidence-based practice by identifying risk factors for disease and targets for preventive healthcare.
What better time to learn about this than now? In the recent series by Mededia, important concepts of biostatistics and epidemiology were covered. In this article, you will be learning the importance of tests and various factors involving them- including Sensitivity, Specificity, Positive Predictive Value and Negative Predictive Value.
Screening Tests Vs. Diagnostic Tests
To approach these concepts, first lets clear the differences between two main types of tests encountered in health: Screening Tests and Diagnostic Tests.
- In a screening test, we examine asymptomatic but at-risk individuals in order to classify them as likely or unlikely to have a particular disease. The goal is to identify individuals before the onset of symptoms. Screening tests almost always require a second, diagnostic test to confirm disease.
For example, in this COVID pandemic, you might have encountered temperature screening stations at airports. They are not confirmatory, but they allow to screen out the individuals at risk i.e. high grade fever.
- In a diagnostic test, the goal is to confirm whether the individual is free from disease or has disease. We examine the individuals who show symptoms or have tested +ve in screening tests.
E.g. PCR assay for COVID-19, will confirm the disease status for any individual screened out due to high grade fever.
Now coming to the properties of these tests:
Sensitivity
- The ability of a test to correctly identify all those who have the disease is known as sensitivity. In other words: “Probability of test results to show up as +ve for a diseased individual”.
- The screening tests usually have the higher sensitivity, because they are intended to screen out as many at risk individuals as possible. So, a test with higher sensitivity, will have higher false positives.
- The formula: Sensitivity = True Positive / Total Diseased = TP / (TP+FN)
Where, TP= True Positive , FN= False Negative
Specificity
- The ability of a test to correctly identify all those who do not have the disease is known as specificity. In other words: “Probability of test results to show up as -ve for a non-diseased individual”.
- The diagnostic tests usually have the high sensitivity AND high specificity, because they want to make sure that all non-diseased people are diagnosed as negative. A test with higher specificity, will have higher false negatives.
- The formula: Specificity = True Negative / Total Non-Diseased = TN / (TN+FP)
Where, FP= False Positive , TN= True Negative
Specificity and sensitivity both have the disease status in their denominator (Non-diseased or Diseased).
After the test results are received, the predictive values of those tests tell us about the chances that the diagnosed have the disease or not. This is where the predictive values come in handy.
Positive Predictive Value
- It is the probability that a person who tests (+ve), actually HAS the disease.
- The formula: PPV= True Positive/ All Positive = TP/TP+FP
Where, TP=True Positive , FP=False Positive
For example, PPV of Aptima SARS-CoV-2 assay is 61.8–89.8%.This means that if 100 people took Aptima SARS-CoV-2 assay and were diagnosed as positive, out of those 61-90 people would actually have COVID-19.
Negative Predictive Value
- Similarly, there is a Negative Predictive Value. It is the probability that a person who tests (-ve) does NOT have the disease.
- The formula: NPV= True Negative/ All Negative = TN/TN+FN
Where, TN=True Negative, FN=False Negative
PPV and NPV both have the test status in the denominator, unlike specificity and sensitivity.
