Clinical trials are required to document that a drug works (and to get an idea of safety and side effects as well). What goes into the design of a clinical trial and what does it mean?
Scientists use clinical trials to more or less prove that a certain drug, procedure or device works. By proof is meant that (statistically speaking) if the trial was repeated again and again, there is a very high likelihood that the results would be the same almost every time.
Numbers, such as an improvement in psoriasis PASI scores, are often reported with a "p-value" quoted, for example: p<0.001. What this tells us is that the likelihood of that beneficial result being found and erroneously reported when in fact the drug had no effect is very small, less than 1 in 1,000. When "p-values" are small, the study has statistical significance. Larger studies with more patients and low p-values have more power and may carry more weight or believability than smaller trials with equally low and significant p-values.
Measuring results, which is often done by human observers, can result in bias that makes results less believable. In order to try and remove bias, trials can be blinded to keep participants and observers from knowing if they are getting the treatment or not. A trial where neither party knows whose getting active treatment and where a mock or placebo treatment is given is called a double-blind, placebo-controlled trial.
Another trial design often used is a cross-over trial, during which one group of patients is treated with a sequence of treatments, separated in time, and the response to the different treatments is compared.
For a more in-depth look please follow some of the links below:
