4-arm trial design

#1
Hi!
I am conducting a research project where I will be conducting a taste trial. I am wondering about the possibility of a 4-arm trial or whether its best to stick to 2-arms?

These would be the options:
2-arm: drug and formulation
4-arm: drug, formulation, food, food + formulation

Wanting to know the pros & cons of a 4-arm vs 2-arm trial in terms of running it and collecting data / sample size etc?
 

Karabiner

TS Contributor
#2
It depends on the theoretical background here, on the research questions, the study goals.
You did not mention any of these. Apparently, you do not have a 2X2 design but a 1x4 design,
but it is not quite clear.

Of course, 4-arm trials are usually more costly than 2-arm trials, in terms of larger sample sizes
needed and the setup/execution of the trial.

With kind regards

Karabiner
 

hlsmith

Less is more. Stay pure. Stay poor.
#3
You would also want to list out the purpose and comparisons a priori. Meaning, are you going to conduct all pairwise comparisons (e.g., 1 v 2, 1, v 3,..., 3 v 4. If so, that is 6 comparisons, and many times people will correct there level of significance to lower the risk of false discovery. This would mean you would need more subjects to find a results while lowering risk of type I error.
 
#4
You would also want to list out the purpose and comparisons a priori. Meaning, are you going to conduct all pairwise comparisons (e.g., 1 v 2, 1, v 3,..., 3 v 4. If so, that is 6 comparisons, and many times people will correct there level of significance to lower the risk of false discovery. This would mean you would need more subjects to find a results while lowering risk of type I error.
It depends on the theoretical background here, on the research questions, the study goals.
You did not mention any of these. Apparently, you do not have a 2X2 design but a 1x4 design,
but it is not quite clear.

Of course, 4-arm trials are usually more costly than 2-arm trials, in terms of larger sample sizes
needed and the setup/execution of the trial.

With kind regards

Karabiner
Thank you for the info.

Just a bit of clarification from my post - the 4 arms would be more so a 2x2 design testing the drug to the formulation and the food to formulation with food. In a way there are two aims - seeing if the formulation is better than the drug (which is kinda obvious as it would be) and then if there is a drastic difference between food that patients have and food+formulation .

I guess we are trying to determine if we should just go forward with drug + formulation or do all 4 (not too worried about costs, more figuring out what would be the best outcome to show in terms of data, like what would be significantly better to demonstrate and then go on to publish?