How should we interpret Nan's apparent good response and the other lady's apparent non-response to the treatment?

I don't see either response as conclusive for either party. I see the situation probabilistically. Based on prior experience the two ladies went into the trial with a probability of success. On the basis of the results we have seen, we need to adjust the probability for both. For instance, we may of thought that Nan went into the trial with a probability of success of 0.2, and on the basis of the response so far we need to increase this to 0.5, say. (This approach is called Bayesian.)

We can add to our knowledge by increasing the number of new patients or by increasing the number of data points for existing patients.

There are all sorts of confounding events that could have affected the responses: the type of PD; the placebo effect; and the weather being obvious ones.

I agree with jeffreyn that there's a long list of difficulties that need to be overcome before someone can be allowed to continue with the treatment. But, I don't think that these are insurmountable.

If we want more people to volunteer for trials then we need to make their experience better. A compassionate response should be the least that we can expect.

John