Hi Tom,
For 7c, when it's asking "at what range of prior probabilities of hereditary spherocytosis..." do we have to calculate treatment thresholds?
Regards,
Swaroop
I'll save an email an respond to both questions about problem 7 here.
RE: 7c: It might help to remind you that in this class we use the word "treatment" in problems like this in a very general way, where the treatment is the next thing you do, if the posterior probability after your test is sufficiently high. So in this case, the "treatment" is doing the osmotic fragility test, the threshold for which you estimated in part 7.b.
RE: 7d. The title for the table on the right should read: "Babies who did not have jaundice needing phototherapy." (It said "No jaundice requiring phototherapy.") The two tables in part d sum to the table at the beginning of the problem.
Tom,
On question 7f, to fill in the table for the likelihood ratios we have to only consider the babies with HS index >= 0.36? (not all the babies, correct?)
thanks,
Aida
No, for 7f those with HS >= 0.36 would be in the HS+ column.
Tom
So, for the table in 7f, across the top it says "HS+" Should this actually be HS index test positive (i.e. those with HS index scores greater than or equal to 0.36)? Or is this referring to positive for HS (hereditary spherocytosis)?
Thanks!
Whoops! My response to Aida was confused! Thanks for asking about this Eric!
The HS+ and HS-- columns in 7f should indeed refer to those who do and do not have HS.
I now understand the reason for Aida's question, which was whether the table should include everyone, or just those with HS index positive. The answer is that the table for 7f should include everyone. If we are trying to decide whether to do the HS index, we can't assume we know its result at the time we are making the decision to do it!
Sorry about my earlier confused and confusing answer.
Tom