What is a "randomized double blind crossover"?

Randomized: means the researcher does not assign how the subject will be treated. The treatment will be decided by either pre-arranged code in a opaque envelop or container.

Double blind: means researcher, the treatment handler, and even the patient does not know what kind of treatment is provided except the one who provide the treatment (such as pharmacist who is not one of the researcher that provided drug without any information that the doctor or nurse will know)

Crossover: means the treatment will be switched to the other treatment arm after a pre-determined period.

Correct me if I am wrong.

It is correct.

Randomized means the subjects are randomly assigned to different arms of the study which receive different treatment. As Ru-Jeng wrote the treatment will be decided by either pre-arranged code in an envelop. Double-blind means neither the physician nor the patients knows whether an active drug is given or an inactive placebo is given (pharmacist may know what is given, since pharmacist is not participating in the study). Cross-over design is the one in which patients receive 2 different treatments. When the randomized trial is a repeated measures design, the same measures are collected multiple times for each patient.

A crossover clinical trial is a repeated measures design in which each patient is randomly assigned to a sequence of treatments, including at least 2 treatments of which one is the standard treatment or placebo. A crossover design has two advantages over a crossover longitudinal study. Firstly, the influence of confounding variables is minimized since each patient serves his or her own control. Secondly, all crossover designs have 'balance' that means all subjects should receive the same number of treatments and that all subjects participate for the same number of periods. Interestingly, in most crossover clinical trials each patient receives all treatments.

Cardiff University provide an e-Learning stand alone Masters level module on Research, Statistics and Evidence Based Practice, which is appropriate for different healthcare professionals working in any clinical areas. This might be something you're interested in developing further.

The downsize with this design is the presence of a carry-over effect.

Regarding the reply of Ru-Jeng Teng and Omer Iqbal ; both are OK; but missed the concern raised by Tim Constantin.

So if I may add, that in any crossover study design we have a wash out period.

In case both active drugs tested, have different wash out period; then we use logically the longest wash out period.

In case we use placebo; the wash out period will be more simply determined.

Please see an example of study using this design.