In the US, there are about 15 million cases of sore throat each year. Sore throat can be caused by either virus or strep bacteria. While most of virus infection will recover on its own, strep infection may cause serious illness in certain population and may require medication.
The rapid strep test (RST) in lieu of a lab culture.
Like all medical tests, RST is not 100% accurate, well not even close. So, what’s the value of the test? In this video, we will explore this question with Bayes Theorem.
First, let’s work on some of the common language involved in the medical field, such as False Positive. In our context, the test (or imaging) is used to inform us about the cause of illness. So, follow the jargon of Bayes Theorem, the test result is the information , and the cause of illness is the main event. Accordingly, during the induction stage (development and trial of the test), we learn the accuracy of the test, in the form of reliability as below.
Now, we can see that the test results (info) can be either Positive or Negative, and the cause of illness (event) can be either virus or bacterial. The reliability in this context becomes the accuracy of the test conditioned on the cause of illness. As such, True Positive (TP) means the test result of positive is correct (True) conditioned on bacterial being the cause. Correspondingly, False Negative (FN) means that the test result of Negative is incorrect (False) conditioned on bacterial being the cause. To summarize, reliability is about whether or not the test results (Positive or Negative) is correct (True or False).
Next, let’s provide some numbers to support our understanding of Bayes in the context of RST.
Click here to open in a new window.