Joined

·
234 Posts

True Positives - False positives

a. 20 - 9992

b. 20 - 89,928

c. 40 - 9992

d. 50 - 89,928

e. 60 - 9992

f. 60 - 89,928

g. 80 - 9992

h. 80 - 89,928

1 - 5 of 5 Posts

Joined

·
234 Posts

True Positives - False positives

a. 20 - 9992

b. 20 - 89,928

c. 40 - 9992

d. 50 - 89,928

e. 60 - 9992

f. 60 - 89,928

g. 80 - 9992

h. 80 - 89,928

you have to construct a 2X2 Table to determine the sensitivity and the specificity of the test.

They've done it on 2000 people 1000 with disease and 1000 without the disease and it was positive in 250 in diseased people and 100 in non diseased people.

So your 2X2 table look like this

click to enlarge

Sensitivity = TP/TP+FN = 250/250+750 = .25 or 25%

Specificity = TN/TN+FP = 900/900+100 = .9 or 90%

Now in the new population you also have to construct a new 2X2 table

They told you the prevalence is 80/100,000 so your total diseased people is 80 and the rest (100,000-80=99,920) are not diseased.

since we know the sensitivity is 25% so we expect quarter of the 80 to test positive that is 20 and the rest (80-20=60) to test negative.

since we know the specificity is 90% so we expect that 90% of the 99,920 to test negative and 10% (99,920X10%=9992) to test falsely positive.

so your 2X2 table should will be

click to enlarge

Obviously the answer is A. In fact you don't need to bother yourself with the specificity calculation because there's only one option with 20 as true positives so calculation of the sensitivity part would have been enough to answer the question

Joined

·
234 Posts

Thanks for that wonderful explanation. I appreciate

1 - 5 of 5 Posts

Join the discussion

USMLE Forums

A forum community dedicated to the United States Medical Licensing Examination. Come join the discussion about schools, exams, news, prep, reviews, accessories, classifieds, and more!

Full Forum Listing
Recommended Communities

Join now to ask and comment!