Understanding Diagnostic Tests-Sensitivity, Specificity & Predictive Values

We have looked at What is Normal in an earlier post.

Diagnostic tests are important. Much of Radiology and Imaging is made of diagnostic tests/imaging. We need to know what these tests mean, what the jargon surrounding the tests mean, the application of the results of these tests to clinical practice and care for an individual, and know how to evaluate new tests and the information around new tests.

In this post, we continue our exploration of few of the commonly used (or misused) parameters relating to diagnostic tests.

We will look briefly  at

  1. Sensitivity of a diagnostic test
  2. Specificity of a diagnostic test
  3.  Predictive Values of diagnostic tests

We will look at

  • What do these results mean?
  • What are their limitations?
  • How can we translate them into clinically useful bits of knowledge?
  • How can we communicate these results to our patients?

There are 3 presentations. We will recommend you view them in order of Sensitivity, Specificity and then Predictive Values.

We are also providing a link to a PDF file that you can download and use to practice the estimation of predictive values as well as to see how certain parameters can influence them.

 

Download PDF to practice the calculation of Predictive Values: Practice Predictive Values

Summarizing the Key Take Home Messages from this post

Bare Bones Key Take Home Messages:

Characteristics of a test

  1. How do we use Sensitivity in a clinical setting?
    1. If a sign, symptom or test has a sufficiently high sensitivity, a negative result rules out the presence of the target disorder. (SNOUT)
  2. How do we use Specificity in a clinical setting?
    1. If a sign, symptom or test has a sufficiently high Specificity, a positive result rules in the presence of the target disorder. (SPIN)

Note: Sensitivity and specificity are functions of the test.  They help to rule in or rule out but do not yet establish a diagnosis.

Highly Sensitive Test is Positive– Need for additional test to rule out false positiveCan follow with a highly specific test– if test is positive – may indicate it is not a false positive

Highly Specific Test is Negative- Need for additional test to rule out false negative-Can follow with a highly sensitive test-if test is negative-may indicate it is not a false negative

Translating Results into Clinical Practice

Predictive Values

  1. We can tell the patient- XX% of those with a positive test result may have the target condition of interest
  2. We can tell the patient- XX% of those with a negative test result may not have the target condition of interest
  3. We can be aware that the predictive ability of the test will decrease if it is applied in a low risk compared to a high risk population
  4. We can be aware that we might be better off choosing a more specific than sensitive test when we apply the test to a low risk population (example- screening programs)