Example: Cryptosporidiosis prevalence
Let's look at an example:
You conduct a study to estimate the true prevalence of cryptosporidiosis in a representative sample of 248 lambs from a peninsular on St. Kitts island. You use a rapid screening test, which characteristics are given by the test producer:
- Sensitivity: 93%
- Specificity: 97%
Among the 248 lambs, 143 have a positive test result.
The apparent (test) prevalence corresponds to the “observed” number of
animals having a positive result among the population: 143/248 = 0.58
or 58%
How can we interpret this result?
Based on the formula above, the true prevalence is estimated to be:
TP = (0.58 + 0.97 - 1) / (0.93 + 0.97 - 1) = 0.61 or 61%
Conclusion:
The screening test in this situation slightly
underestimates the true prevalence (mainly due to the low sensitivity
of the test).