The FastDay Forum

[carorees]

Ben Goldacre has created a website where one can create one's own randomized trial. http://www.randomiseme.org/ I'm sure we could use it to answer some of our tricky questions.

Currently it seems to have a comedy feel as the first trial is to determine whether eating cheese late at night causes nightmares!

Ideas for a trial anyone?

http://www.randomiseme.org/trials/140 LOL

Effect of hunger on eating fast day food in 3 meals vs one meal ? Using some predefined scale for hunger / fullness etc.

[carorees]

I think that would be better as a cross-over trial so as to negate the effect of different people's assessment of hunger even if using a VAS.

Similarly, "Does skipping breakfast result in more hunger at lunchtime?" which was my first thought also needs a cross-over design.

"Does skipping breakfast result in more hunger at lunchtime?" has been done and published already I think.

I wasn't clear if the site actual did any randomising ? and can it not do cross-over by simply have the groups switch over and report in the other category.

[carorees]

I can't find one about breakfast vs no breakfast, only one about having a banana with porridge vs no banana.

Yes, the site randomizes participants. When you sign up to join a trial it assigns you to one or t'other group and gives you the instructions (i.e., eat a banana with your breakfast porridge every day, note how long until you feel hunger pangs)

I haven't checked about cross-over design yet

I meant published clinical studies, sorry. Bacon & eggs vs cereal breakfast would be another - hunger, snacking, lunch intake, weight change etc.

What seems to me the most pressing problem that needs research is what is the most effective method for maintenance.

I have just reached that stage and reading the various posts 6:1 does not appear to work and most people are trying various combinations to find something that works for them.
It would be great if on reaching your goal you were invited to take part in a trial to help determine the most effective approach.
A randomised trial would be needed I think.