What's the Likelihood of Disease? Pre-test Probability as an Aid to Diagnosis

Tom G. Bartol, MN, NP


February 19, 2019

Estimating the Likelihood of Disease

When evaluating a patient with stable chest pain and no known coronary artery disease (CAD), how does the clinician decide whether to order imaging or other diagnostic tests? One approach involves a pre-test probability score.

Pre-test probability scores have long been used to aid in the diagnosis of CAD. The original score (developed in 1979[1]) integrated the variables of age, sex, and chest pain characteristics to estimate the likelihood that a stable patient with chest pain has obstructive CAD and to determine the need for further diagnostic testing.

In a recent editorial, Di Carli and Gupta[2] posit that pre-test probability models may no longer be as useful as they once were. Over the past two decades, the use of such scores has lowered the diagnostic yield from invasive and noninvasive coronary angiography, because tried-and-true probability prediction models have tended to overestimate the likelihood of obstructive CAD in contemporary populations. But rather than abandoning pre-test probability scores, new risk scores have been developed for modern populations, leading to a "battle of the scores" to find out which model best estimates the likelihood of obstructive CAD.[2]

At the same time, guidelines differ on whether clinicians should continue to use pre-test probability estimates for CAD. As we await further high-quality evidence on the validity of these tools, Di Carli and Gupta recommend an approach based on sound clinical judgment supported by guidelines and high-quality evidence, including the use of pre-test probability scores.[2]


It's easy for clinicians to click a button and order a diagnostic test to rule a disease in or out. Without considering pre-test probability, however, some of this testing may be unnecessary, and the results may not lead to an accurate diagnosis.

Pre-test probability is the likelihood of a patient having a certain condition, taking into account such factors as the prevalence of the disease in the population and the patient's signs and symptoms, gender, and ethnicity, along with known risk factors for the condition. For some conditions (such as CAD), prediction tools exist, but pre-test probability often is intuited using the clinician's judgement, experience, and "practice-based evidence."

With a pre-test probability score, a patient's chance of having the disease can be classified as low, intermediate, or high.[3] High and low pre-test probability scores are less useful. With a low pre-test probability score, any positive diagnostic test finding is more likely to be a false rather than a true positive, leading to even more testing, with the inherent risks of anxiety and cost. With a high pre-test probability score, a false-negative result is more common than a true negative.

Diagnostic testing is most effective when used in patients determined to have an intermediate pre-test probability for a condition. In these patients, there are fewer false-positive and false-negative results.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.