The Power of Pre-Test Probability

What is pre-test probability?

Pre-test probability is the likelihood a patient has a disease before any diagnostic test is performed. Estimating pre-test probability is not an exact science, but rather something that is refined through clinical expertise. Let’s look at the importance of pre-test probability in a case.

You’re seeing a 64-year-old female with acute cough. Before you dive into the history you think of the possibilities: upper respiratory tract infection, bronchitis, pneumonia (and the list goes on). It’s flu season, so you know there is a higher prevalence of respiratory infections at this time of year. You learn she has a 4-day history of purulent cough and fever. At this point your suspicion for pneumonia is high. You rule out other emergent conditions and eventually diagnose her with pneumonia.

In this example you’ve used a combination of disease prevalence (pneumonia during the flu season in an older adult), the patient history, and physical exam to diagnose pneumonia. You question whether you should order further testing.

Before ordering any tests, you ask yourself the following:

1) What is the probability that this is pneumonia (low, medium, or high)?

2) Have I reached the threshold for treatment? (yes)

3) Will any test change my plan for treatment? (no)

4) Could ordering a test possibly do more harm than good? (Possibly yes, like incidental findings)

Further diagnostic testing isn’t warranted at this moment in the diagnostic pathway. You’ve reached the treatment threshold without having to order any diagnostic testing.

Meeting the treatment threshold

Source: adapted from Pellatt et al. (see reference list)

Let’s say the same patient comes back to you complaining of a chronic purulent cough, and she has lost 10 pounds unintentionally. You think, could this be TB or malignancy?  Your pre-test probability isn’t 100% , but is high enough you have reached the “test threshold.” At this point in time, a chest x-ray is reasonable to assess for a more sinister etiology of chronic cough. The chest x-ray comes back negative but the patient continues to have symptoms. Recognizing a chest x-ray is not very sensitive for lung cancer, you decide to order a CT scan to definitively rule it in or out.

Meeting the test threshold

Source: adapted from Pellatt et al. (see reference list)

This was a quick introduction to pre-test probability, but we will dive deeper in future blog posts.

Key Take Home Point: Think about the pre-test probability of a disease when ordering any diagnostic test – this is determined by the prevalence, history, and physical exam.

References/Readings

1. Pellatt R, Purdy E, Keijzers G. Review article: A primer for clinical researchers in the emergency department: Part XI. Inertia before investigation: Pre‐test probability in emergency medicine. Emergency medicine Australasia. 2020;32(3):377-82. (VIEW HERE)

2. Ranganathan P, Rakesh A. Understanding the properties of diagnostic tests – part 2: Likelihood ratios. Perspectives in clinical research. 2018;9(2):99-102. (VIEW HERE)

3. If you want to learn more about the methods that go into designing diagnostic test accuracy studies, check out this post on P-values to get stared