Sensitivity Specificity and Positive and Negative Predictive Values of Screening Tests

Sensitivity Specificity and Positive and Negative Predictive Values of Screening Tests

Evaluating Sensitivity & Specificity of Two Screening Tests:

First, download the Excel worksheet (“Exercise2.xlsx”), which displays made-up data from a study to determine the sensitivity and specificity of two serology (antibody) tests for the “Swim Flu”, a novel respiratory infection. There are twenty subjects, ten of whom had a confirmed diagnosis of “Swim Flu”, and ten of whom were confirmed disease-free.

Their serum antibody levels, as determined by Test A and Test B, are shown in the tables on the spreadsheet. High antibody level indicates a positive test and Low antibody level indicates a negative test. By entering various cutpoints for the two antibody assays in the yellow-highlighted cell, you will see side-by-side tabulations of the numbers of true positives (TP), false negatives (FN), false positives (FP), and true negatives (TN). By default, “10” has been entered as the antibody cutpoint. Calculate the sensitivity and specificity for Test A and Test B using the numbers from the table that is highlighted in pink on the Excel worksheet. Then enter “8”, “9”, “11”, and “12” and re-calculate the sensitivities and specificities for each of these cutpoints, entering the results in the table highlighted in green on the Excel worksheet. Finally, upload your responses to Blackboard under “EX2”

QUESTIONS

  1. Which of the two screening tests would you use to screen for “swim flu” and why?
  2. What antibody cutpoint would be a reasonable cutpoint for Test B assuming that you would prefer equal rates of false positives and false negatives?
  3. Calculate and interpret the positive predictive value (PPV) for Test A at antibody cutpoint of 10?
  1. Calculate and interpret the positive predictive value (PPV) for Test B at antibody cutpoint of 10?
  2. Discuss the implications of a 100% sensitive test. [Hint: Consider false positive or false negative rate].