Question about Sample Size Calculation

#1
Hi there,

I'm involved in a clinical research at an academic medical center and our statistician has been unavailable for a long time due to demand generated from other departments during covid, and it's holding back our IRB writeup for too long. We are just missing the paragraph about statistics and I'm wondering if I can get enough insight from the forum here to push an IRB through.

My question: sample size calculation.
Our study: We are going to measure 2 variables for one study population before treatment and after treatment at 6 months. For both variables and our population of interest, I know the mean, SD, and MCID.

Any expertise or input will help. Thank you!
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Does everyone get the treatment - AKA is there also a control group?

P.S., Please don't post the same questions twice.
 

hlsmith

Less is more. Stay pure. Stay poor.
#3
On yeah, how do you know the mean, SD, and MCID, if you don't have IRB approval? Also, please define MCID.

Lastly, these two variables are outcomes?

Thanks.
 
#4
Does everyone get the treatment - AKA is there also a control group?

P.S., Please don't post the same questions twice.
Thanks for responding. I posted in the other thread thinking that it was more appropriate. Won't do it again.

-There is no control group and everybody will get the treatment.
-There are some papers that we will cite which have studied the specific patient population and were able to ascertain the mean, SD, and MCID (minimal clinically important difference) of the two variables that we are looking at.
-The two variables will be our primary outcomes.
 

hlsmith

Less is more. Stay pure. Stay poor.
#5
Values for the pre-period? what about assumptions for the other period? How big of a change are you assuming and the variability of that change? What do their distributions look like for outcomes. Meaning some variables, length of stay, can be heavily skewed.

While everyone of post data collect right at 6 months or will it vary and matter?
 
#6
This is the part where I'm somewhat stumped. Everyone will have their data collected at 6 months. The two variables are numerical scores that we anticipate will be improved at least by the MCID, which is nearly equal to one standard deviation.
 

hlsmith

Less is more. Stay pure. Stay poor.
#7
There are two ways to go at this. One, just look at the differences and see if they exclude the null, '0'. However, the pre-data may inform how big of a change there is, meaning you may need to control for it. The example I like to use is, what if I had a weight loss program and most people are minimally overweight and a couple are obese. Obese people have the ability to lose more weight, thus it may be beneficial to try and control for it.

Is your sample a random sample or convenient?
 
#8
Nonrandom sample. We will be enrolling patients with the same subjective symptom (above a certain number on the severity score). This will be a very basic nonrandomized observational study.
 
#10
Correction: There will be an intervention for all patients at zero month and then 6 months later, we will reassess the two variable numerical scores. So not an observational study, technically speaking.