number of samples needed to show difference between two samples I am an opthalmologist not trusting what the drug companies are telling me so need some help 
I am looking a papers on glaucoma where the pressure of the eye is measured.
The SD of the measurements is 0.8mm.  Company X  does a study on lets say 200 patients and says their drug lowered the pressure by 8mm  vs a standard drug which lowers the pressure by 6.9 mm
So I am wondering if the study shows a difference of 1 mm between the two groups , but each measurement ( patient)  has a measurement which has some intrinsic inaccuracy and a SD of 0.8  ,  how many patients would one actually need to be sure this is an accurate study 
Hope this is clear ,realize its pretty basic but appreciate any help.. 
Thanks
Rob
 A: Comparing new drug to fixed standard. Claim should be OK if new drug info is based on a clinical trial following FDA standards. Technically, even a sample size as small as 20 would suffice---provided
you believe such a small sample is truly randomly selected from an appropriate population.
A one-sample t test with $n = 200$ subjects and standard deviation 0.8mm,
would have essentially 100% power to detect a difference of 1mm, testing at 1% significance level. Output
from Minitab statistical software:
Power and Sample Size 

1-Sample t Test

Testing mean = null (versus ≠ null)
Calculating power for mean = null + difference
α = 0.01  Assumed standard deviation = 0.8

            Sample
Difference    Size  Power
         1     200      1

Put in another way, a 99% confidence interval with $n = 200$ has
a margin of error of about $\pm 0.15$mm.
Comparing two drug trials of about the same size. If you're comparing two trials, seeking 99% power:
Power Curve for 2-Sample t Test 

Power and Sample Size 

2-Sample t Test

Testing mean 1 = mean 2 (versus ≠)
Calculating power for mean 1 = mean 2 + difference
α = 0.01  Assumed standard deviation = 0.8


            Sample  Target
Difference    Size   Power  Actual Power
         1      33    0.99      0.991060

The sample size is for each group.

