Specificity: Proportion of cases without the illness which do not have that specific sign (true negatives).
Example: Let’s take 100 children without pneumonia. If fast breathing is not present in 75 out of these 100 children without pneumonia, it has a specificity of 75%. These 75 children are “true negatives”. In this example, fast breathing is present in the remaining 25 children without pneumonia, who would then be incorrectly classified as having pneumonia if we relied only on fast breathing. These 25 children are “false positives”: they have fast breathing for reasons other than pneumonia. So:
Children without pneumonia (denominator): 100
Children without fast breathing (numerator): 75 (true negatives)
Specificity of fast breathing: 75/100 = 75%
[Children with fast breathing: 25 (false positives)]
Thus, if a “true negative” is a child without pneumonia correctly identified as having no pneumonia by the absence of fast breathing, and a “false positive” is a child without pneumonia incorrectly classified as having pneumonia because of the presence of fast breathing:
Specificity = [true negatives / (true negatives + false positives)
Note: specificity does not depend on the prevalence of the illness or condition in the population.
To sum up, specificity tells us what percentage of children without the illness we can correctly consider free of the illness (and, on the other hand, also what percentage we would incorrectly classify as having the illness when in fact they do not have it). Low specificity leads to over-treatment
Specificity: Proportion of cases without the illness which do not have that specific sign (true negatives).
Example: Let’s take 100 children without pneumonia. If fast breathing is not present in 75 out of these 100 children without pneumonia, it has a specificity of 75%. These 75 children are “true negatives”. In this example, fast breathing is present in the remaining 25 children without pneumonia, who would then be incorrectly classified as having pneumonia if we relied only on fast breathing. These 25 children are “false positives”: they have fast breathing for reasons other than pneumonia. So:
Children without pneumonia (denominator): 100
Children without fast breathing (numerator): 75 (true negatives)
Specificity of fast breathing: 75/100 = 75%
[Children with fast breathing: 25 (false positives)]
Thus, if a “true negative” is a child without pneumonia correctly identified as having no pneumonia by the absence of fast breathing, and a “false positive” is a child without pneumonia incorrectly classified as having pneumonia because of the presence of fast breathing:
Specificity = [true negatives / (true negatives + false positives)
Note: specificity does not depend on the prevalence of the illness or condition in the population.
To sum up, specificity tells us what percentage of children without the illness we can correctly consider free of the illness (and, on the other hand, also what percentage we would incorrectly classify as having the illness when in fact they do not have it). Low specificity leads to over-treatment
