USMLE Forums banner

Positive Predictive Value Question

2K views 7 replies 4 participants last post by  mutant 
#1 ·
A screening test is developed for assessing vulnerability to developing a certain condition. A sample of 10,000 people between the ages of 18 to 24 is recruited from the general population and given the screening test. Of this sample, 200 individuals are identified as likely to develop the condition. The sample is tracked over the next twenty years. Forty of the original sample, although none of those selected by the test, are lost to follow-up. A total of 100 people from the sample eventually were diagnosed with this condition, of which 90 were correctly identified by the screening test. Based on this study, the positive predictive value of the screening test is best estimated as which of the following?

A. 45%
B. 60%
C. 75%
D. 90%
E. 100%
 
#3 ·
Answer

actually the answer is A (i got answer D which is totally wrong since i forgot about the 200 *careful!!!*)

so here is how it goes:
200 people are LIKELY to get the disease (keep in mind for later)
so after 20 years only a 100 turn to be positive, 90 are TRUE POSITIVE and the other 10 are FALSE POSITIVE, we add here the 100 others which didnt follow up which make them FALSE POSITIVE leading to 110 on that part of square so:
PPV: TP/TP+FP = 90/90+110 = 0.45 aka 45%
voila! ;)
 
#5 ·
actually the answer is A (i got answer D which is totally wrong since i forgot about the 200 *careful!!!*)

so here is how it goes:
200 people are LIKELY to get the disease (keep in mind for later)
so after 20 years only a 100 turn to be positive, 90 are TRUE POSITIVE and the other 10 are FALSE POSITIVE, we add here the 100 others which didnt follow up which make them FALSE POSITIVE leading to 110 on that part of square so:
PPV: TP/TP+FP = 90/90+110 = 0.45 aka 45%
voila! ;)
I understand the total is 200, and that 100 were Positive. And of those 100 that are positive, 90 were true positives, and 10 are false positives.

Wouldn't the remaining 100 be True negatives OR False negatives? (since they were not diagnosed with the condition) And in either case (TN or FN...neither is involved in calculation of PPV)

so TP/(TP+FN) = 90/(90+10) = 90%
 
#6 ·
ok, nvm. i understand that those 200 people...are all Positives. 100 got the condition, and of those, only 90 were TRUE Positives. Which means 10 plus the 100 are all False Positives...

so PPV = 90/(90+110) = 45%



good question, the 40 is a complete distractor. and if you read the question very fast, it's easy to miss that the Positives is 200 (not 100).
 
#7 ·
ReggieMiller i got the same answer as you, i found it a really good question as well, we have to filter which numbers are distractors and which are the ones needed to find the answer, so we have to be careful, its scary cause the 200 seemed so off the approach to the answer.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top