[johnt]

A number of drugs that have been successfully tested on animals (with chemically induced PD) have shown disappointing results on humans. And, some PD drug trials have been ended, because the results are not statistically significant, even though some participants have reported considerable benefit.

To add to this, we can all see PwP with different symptoms and different rates of progression.

Taking all this together must raise the possibility that there is more than one type of PD, and following from this the possibility that a therapy that helps one group of PwP might have no benefit for another or, even, make them worse.

This makes things difficult for traditionally run clinical trials: the results of one group being watered down by another, leading to statistical significance being lost and, therefore, a good drug for some being abandoned.

Fortunately, there are other testing techniques which overcome these problems. They focus on the individual, rather than the group.

As the simplest example, you could have a trial with one person. First, to get control figures they would be measured without the drug. They would then take the drug and be measured again. (For PD even this is non-trivial: given the fluctuation of symptoms, to get statistical significance each phase needs to be quite long.) Of course, taken alone it would be dangerous to assume that this result could be generalized to all PwP.

Another technique is to use cross-overs where there are a number of placebo/drug phases. These can be blinded.

As a final example, you can have variable length trials in which a participant stays in the trial until, based on a mixture of their data and that of the group, it becomes unlikely that he or she will benefit. In this form of testing the cohort size is gradually reduced, but people who are showing a good response stay in.

See
"Translational medicine in Parkinson’s Disease: Novel clinical trial designs have been proposed for evaluating novel anti-dyskinetic medications. Strengths and limits of the N-of-1 design."

http://www.pdonlineresearch.org/resp...signs-have-bee

Does anyone know the extent to which this form of clinical trial design is used in PD?

John