Search This Blog

Monday, May 7, 2018

Sensitivity, Specificity, PPV, NPV


Sensitivity is TRUE POSITIVE RATE.
 Diseased (TD) Healthy (TH)  Test Positive True positive (TP) False Positive (FP) PPV = ((TP))/((TP)+(FP)) Test Negative False Negative (FN) True negative  (TN) NPV = ((TN))/((FN)+(TN))  Sensitivity =((TP))/((TP)+(FN)) Specificity  =((TN))/((FP)+(TN))
2x2 table 
Sensitive test picks all diseased and also sometimes without disease. That means it won’t miss the disease.
Sensitivity is a measure of  test positive out of total diseased. 

Specificity is TRUE NEGATIVE RATE.
Specific test picks only the diseased one but may sometime miss the disease itself.
Sensitivity and specificity are not dependent upon the disease prevalence rate.
Specificity is measure of true negative out of total healthy patient. 

Predictive values depend upon the prevalence of disease and accuracy of the test.
Positive Predictive Value is the proportion of the people with positive test who actually have disease.
Negative Predictive Value is the proportion of the people with negative test who actually do not have disease.


PPV
NPV
Definition
% that a person with positive test is actually diseased.
% change that a person with negative test is actually disease free.
Use
Proceed with a patient with positive test
Proceed with a patient with negative test
Relation to prevalence
Low prevalence low PPV
High prevalence high PPV
High prevalence low NPV
Low prevalence High NPV

2 comments:

  1. Shameless. This is why dental hygiene is so poor in the uk and dentist retire at 35

    ReplyDelete