terms in q3 qba hw

terms in q3 qba hw

by Laura Koth -

Can you clarify the following in more detail? I am getting confused because you use A, B, C, D and Es in the table in addition to your separate definitions of disease and drug treatment groups.

1. you have D+=AE in the stem. does "D" have anything to do with D in the table?

2. what is AE? are you multiplying A x E? or is this an abbreviation for "Adverse Effect"?

3. for the statement: "Association between diuretic and syncope, irrespective of Exposure: RRCD= 1.5"

is "CD" referring to the cells C and D in the table? or some other abbreviation?

4. Prevalence of diuretic among E+: PC+E+ = 0.68

is "C+E+" referring to the cell "C" or to the group Control or to use of drug "C" diuretic? Is E referring to a specific cell of E or just to the group of drug anti hypertensive?

Re: terms in q3 qba hw

by Francois Rerolle -

Sorry for the confusion.

The cells in the tables are indexed with a number, like D0 and D1 and have nothing to do with cases (D+), which are defined as adverse event (abbreviated AE) or controls (D-). C is the confounder and C+ (diuretic) and C- (non-diuretic) are its levels. E is the exposure and E+ (exposed) and E- (non-exposed) are its levels.


1) No, D+ is means cases

2) AE is adverse events, i.e cases, i.e D+

3) CD refers to Confounder Disease association not the cells in the table, which are C0, C1, D0 and D1

4) PC+E+ = 0.68 is the prevalence of the confounder (C) among exposed E+, i.e the prevalence of diuretic among E+


Hope that helps are sorry for confusion

Re: terms in q3 qba hw

by Laura Koth -

the term

Confounder Disease association

does not make sense. I have never heard this term before.

can you use a description rather than a short cut term?


Re: terms in q3 qba hw

by Sarah Raifman -

Hi Laura, I just saw your comment so thought I'd reply... I think Francois referred to that term because the question tells us that the RR(CD) is 1.5 -- meaning that there is a positive association between diuretic (confounder) and syncope (outcome). This is also what the stata option "drrcd" in episensi also refers to. 

Regarding your other question about getting CIs ... I'm not sure if this is what you're referring to, but I couldn't get the CIs for the bias-adjusted ORs using episensi in stata either. Is there a way to do this in episensi?  other than bootstrapping?  If you're referring to the last question (4a), then I think you can apply the same equation you use for the corrected RR column for the lower and upper CI bounds. 


Re: terms in q3 qba hw

by June Chan -

Sarah is correct and I added the slide that was supposed to pair with this question in the other Forum question trail. I'm sorry if Dr. Mayeda edited this out. I'd appreciate anyone else letting me know if they find it in her slide deck. thanks!