Prevalence and Predictive Value

Sensitivity and specificity of a laboratory test are independent and different from disease prevalence. We must understand the sensitivity and specificity of a test and the choice we made in defining normal in terms of trading false positives for false negatives. We also must understand the effect of prior likelihood. Given a particular patient, prior likelihood is how we should interpret the results. Many things can affect prior likelihood, including age of the patient.

The following scenarios illustrate the importance of considering the effect of age on disease prevalence:

  1. All 2 year olds are less than 4 ft tall but a 25 year old who is less than 4 ft tall needs to be evaluated.
  2. Hemoglobin of 9.0 g/dl would always be caused by a disease process of some sort in a 30 year old man, but is normal for a 1 month old infant.
  3. Alkaline phosphatase is elevated in all children because of bone growth but is a concerning lab finding in a mature adult.

Prevalence is one way to express prior likelihood. Prevalence is defined as the proportion of persons in a defined population in time with the condition in question. Another way of saying this is ‘prevalence is the pre-test probability’. Increasing prevalence of a disease does not affect the sensitivity or specificity of the test. Sensitivity and specificity are test properties that do not change with the incidence of disease. However, a test’s predictive value is affected by the prevalence of a disease in a population. For example, consider the following example:

Question 1:

A one year screening in a child comes back with hemoglobin of 9.9 g/dl. The history was remarkable for an intake of 32 ounces of whole milk a day. Based on the history, the screening test is:

  1. More likely to correctly identify iron deficiency
  2. Less likely to correctly identify iron deficiency
  3. The screening results are not affected by clinical presentation
Show answer >>

The correct answer is A.

The clinical presentation is very important in interpreting laboratory results. The patient in this case has a positive history with a very high milk intake. This history increases the prior likelihood that the child has iron deficiency and so increases the likelihood that the low hemoglobin is a result of iron deficiency.

Predictive value of a diagnostic test gives the frequency that a positive test result actually indicates a disease. The term is more useful when it is used to positively predict people with a disease. Predictive value is determined by the sensitivity and specificity of the test as well as the prevalence of disease in the population being tested.

Positive predictive value (PPV) is the probability that a patient who has a positive result really has the disease. In other words, PPV of a particular laboratory test is the probability that a person has a disease when restricted to the population that tests positive. Positive predictive value is a function of the sensitivity and specificity of the test AND the prevalence of the disease.