Dynamic Chiropractic

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

The Importance of a Cut-Off Point

In the last column, my colleague Dr. John Stites provided additional insight into the concept of SpIN and SnOut, which are related to the larger concept of sensitivity and specificity. To remind you, sensitivity is the percentage of people who have the disorder in question and whom test positive on a given diagnostic test. Specificity is the percentage of those who do not have the disorder and whom test negative on the test.. Given this, let’s dive a little deeper once again.

Consider a standard medical diagnostic test, such as temperature. Now, I know that temperature is largely a less-than-useful test in the scheme of things, but it will serve to illustrate a point. Temperature, while actually a continuous measurement, has been dichotomized with regard to diagnosis. For our purposes here, if your temperature is above 98.6 degrees, we will call you “sick”; if below, you are “not sick.” The first question we can ask is, how was 98.6 degrees chosen? That specific measurement is known as a cut-off point; it defines everyone below the but-off point as “not sick,” while everyone above is “sick” All laboratory values have been dichotomized this way. Heart rate, blood pressure, BUN, PSA, you name it and we have a cut-off point for it.

Now, let us say for a given test, we have a measure of sensitivity and specificity. What happens if we change the cut-off point? Consider temperature. What would happen if we raised the cut-off point, so that now we would define you as “sick” if your temperature was above 102 degrees? I think you can see that this would likely raise the sensitivity of your diagnostic test; that is, nearly everyone who has a temperature above 102 degrees really is sick. But you can also see that our test is now calling people “not sick” when in fact they are; thus, by raising the cut-off point, we are actually lowering specificity. We are creating many false negatives here.

The reverse occurs when we lower the cut-off point. Now, we will increase our specificity, but we are increasing the number of false positives, so sensitivity now falls.

The actual cut-off point for all of the tests we use were defined only after a great deal of work, looking at large populations of people who both have and do not have conditions of interest. For sure this is important; we would not want, for example, to define someone as positive when they are not, if by doing so we would then refer them for additional invasive testing.


Leave a Reply