Altering Diagnostic Test Cutpoints

Question 1

Measures of association

Answer 1

• for cohort, use RR
• for odds ratio, use OR
• for cross sectional, use RR or OR

Question 2

Diagnostic test endpoints

Answer 2

Many diagnostic tests are measured on a numeric scale, with the intention to classify them as normal or abnormal

Question 3

What is an example of a diagnostic test endpoint?

Answer 3

ELISA plates
- measured as intensity of the color response
- point at which the intensity of the color is regarded as indicative of disease, and a cutpoint is established
Rectal temperature
Neutrophil count

Question 4

Ideal diagnostic test populations have ____ separation

Answer 4

100%

• no overlap in true positives (sick) and true negatives (well)
• will also have achieved 100% sensitivity and 100% specificity

Question 5

Typical diagnostic test populations will have ______

Answer 5

Well and sick overlap

- areas under the curve of FN and FP

Question 6

If you want more sensitivity, you would move the cutoff point further to the _____

Answer 6

Left

- more specificity, move it toward the right

Question 7

How do you change the cutpoint?

Answer 7

Increase sensitivity/decrease specificity

Question 8

Is there a scenario where sensitivity and specificity increase at the same time?

Answer 8

No!!

Question 9

Increased sensitivity results in

Answer 9

Lower specificity, fewer false negatives, more false positives

Question 10

When is the cutpoint altered?

Answer 10

Informally with physical exam findings and bloodwork (done in our heads)

Question 11

Do commercial tests allow alteration of cutpoints?

Answer 11

NO!