Sensitivity, specificity, positive predictive value, negative predictive value

Sensitivity, specificity, positive predictive value and negative predictive value are all indicators of success for a measure: a value of 1 is good news, 0 is bad news.




Sensitivity = P(s₊d₊) / [ P(s₊d₊) + P(s₊d_-) ]

= P(s₊d₊) / P(s₊)

What % tests positive of all who have the disease?

Specificity = P(s_-d_-) / [ P(s_-d_-) + P(s_-d₊) ]

= P(s_-d_-) / P(s_-)

What % tests negative of all who do not have the disease?

Positive predictive value = P(s₊d₊) / [ P(s₊d₊) + P(s_-d₊) ]

= P(s₊d₊) / P(d₊)

What % has the disease, of all who test positive?

Negative predictive value = P(s_-d_-) / [ P(s_-d_-) + P(s₊d_-) ]

= P(s_-d_-) / P(d_-)

What % hasn’t got the disease, of all who test negative?