Dynamic Chiropractic

Dynamic Chiropractic Facebook Twitter
Dynamic Chiropractic
Advanced Search
Wellness Blog
Dynamic Chiropractic
Dynamic Chiropractic PracticeINSIGHTS
Current Graphic

And Even More on Diagnostic Testing

Our last two blog posts, by myself and my colleague John Stites, have dealt with issues related to diagnostic testing. I discussed the importance of cut-off points and how changing them affects the sensitivity and specificity of a test. John discussed the concept of SPin and SNout and related cautions in interpreting them. Let me today add one more complexity to this situation.

As we know, sensitivity is a measure of the ability of a test to correctly identify someone with a a target disorder. And specificity is a measure of how well a test identifies someone without the disorder. For any given test, we can expect to see some false positive tests (where people test positive but do not have the disorder) and some false negatives (where people with the condition test negative). This occurs because there is no perfect test (a perfect test would always be able to identify those with the condition and those without). What we forget when we do tests is that we do them with a general idea of how likely someone might have a positive test to begin with. This matters, a lot.

I am indebted to the short book “The Laws of Medicine” for what follows (1). Consider: what if a test you wish to do has a false positive rate of 1 per thousand. That is, it will find one person in a thousand positive when that person does not have the test. But we wish to use our test in a population of people where the condition we are interested in involves 1 person in a thousand. That is, the infection rate is 1:1000. Perhaps you can see the problem. It leads to this fact- for every person who tests positive with the condition, there will be one who tests positive without the condition. Or put another way, there is only a 50% chance that a positive test is really positive. And this is not a useful test as a result. And if the false positive rate is 1% and the prevalence is half a percent, the test is wrong 19 out of 20 times! Conversely, if we have a population with a very high prevalence rate for the condition- which as chiropractors we do- we can reverse that false positive rate significantly. We can make it so that it is correct 95% of the time. This is because we know something about how frequent the condition is in our practice.

An interesting complexity to be sure!


  1. Mukherjee S. The laws of medicine: field notes from an uncertain science. New York, NY; Simon and Schuster TED Books, 2015:27-28.

Leave a Reply