Epidemiology question

[yoskram]

Hello,
Would like input about a epidemiology question my non-medical friend came across in a study. I contend that it’s a trick question but I’d like to hear others’ thoughts. Thanks!

Here is the link to the article BTW

Medicine’s Uncomfortable Relationship With Math: Calculating Positive Predictive Value

www.ncbi.nlm.nih.gov

Sent from my iPhone using Tapatalk

---

Members don't see this ad.

 

[thetheatrecat]

Why do you think it's a trick question? If you test 1000 people, with a 5% false positive rate, 50 of those people will have a positive test (despite being negative for the disease). If the prevalence is 1/1000 (and the test is 100% sensitive, meaning it catches all of the positives), one person will correctly test positive. The positive predictive value is the number of true positives divided by the number of positive tests (in this case: 1/51=1.96%~2%). PPV increases with increasing prevalence.

 

[yoskram]

Why do you think it's a trick question? If you test 1000 people, with a 5% false positive rate, 50 of those people will have a positive test (despite being negative for the disease). If the prevalence is 1/1000 (and the test is 100% sensitive, meaning it catches all of the positives), one person will correctly test positive. The positive predictive value is the number of true positives divided by the number of positive tests (in this case: 1/51=1.96%~2%). PPV increases with increasing prevalence.

Because they don’t tell you that the sensitivity is 100% in the question... you’re supposed to assume this, which is why I think it’s kinda a trick question



Sent from my iPhone using Tapatalk