The smart way to learn. The smart way to teach.


Course PAID
Do you realy want to delete this discussion?
Forum >>

Dear Paul Please Advice

Dear Paul Please Advice Posted by Ajith S.

Please Advice and Guide I have no clue to get the ansvwer for this question - Suspected answer is ?C

Q. RCT, whether continuous intrapartum CTG is of benefit in low risk pop with 4/1000 adverse outcome. Reduction of 25% will be worthwhile. Alpha error 0.05, power 80%, sample size?
A. 200
B. 2000
C. 20000
D. 200000

with kind Regards

Posted by PAUL A.
Difficult – not sure how you are expected to work this out unless you are given tables and a calculator. Sample size calculations vary and one gets used to putting these parameters onto a stats package and getting an answer.

However, appropriate formula is:

n = [A + B] squared * [(p1*(1-p1)) + (p2 * (1-p2))]/[p1-p2] squared

– although there are many different variations.
for P = 0.05, A = 1.96 and B = 0.84 (80% power) – presume you are given these or an appropriate table. Do not think you should be expected to remember them.
Since proportion p1 (4/1000) and p2 (3/1000 – 25% reduction) are small, (1-p1) and (1-p2) approximate to 1for my calculation.
Using these, I get a figure of 54,000. I have entered these in my stats package and get a figure of 47,000. I will suggest that you are not actually expected to work it out in the time allocated. Rather, you are simply expected to know the type of numbers that will be involved in such a study in which case I will go with C.

Posted by Ajith S.
Thank you Ajith