My work has been to test theory by using agents that target molecular apoptosis pathways and note the response. So if you think that pathway X has blocked apoptosis and you give agent A with the result being neuropathy and eventual healing it strongly supports the theory provided the agent doesn't do this in normals and in people with autoimmune disease (AD). I've already been screened for AD and it has been ruled out.
AD cannot be ruled out. Only known AD can be ruled out. There are secondary markers though, the absence of which can suggest no AD. How is your sIL2r for example? There are lots more, but they escape me at the moment.
Healing from an agent only suggests the theory is right. You have to demonstrate mechanisms, and the mechanisms are active. Then move to large scale dbRCTs. Even then it can be wrong.
You may also be using theory driven assumptions, eg:
So if you think that pathway X has blocked apoptosis and you give agent A ...
Let us suppose agent B is working, and agent D, and agent H. The other agents don't work. Why don't they work? Why do these specific agents work (presuming this is accurate for now)? Many agents have multiple chemical effects, and variable response in different organs. How do you know which effect is working?
For example, when I was investigating quaifenesin there were lots of responders. Yet the model relied on assumptions about alterations to intracellular and mitochondrial phosphate. These were implausible. So an effect had to be via another pathway, or accidental, or something. If they involved phosphate then they did so via another mechanism.
Such alternative mechanisms can also include things we have not discovered yet.
As I pointed out in my blog (link given earlier in this thread) even the success of an agent does not prove the theory behind the agent is correct.
Lets take Rituximab. Clearly it works very successfully in about 6% of patients. We do not know why ... yet. It also fails completely in about 30% of patients. Again we do not know why.
Quoting an earlier comment of yours:
The main problem is that many of the supplements (alpha lipoic acid, resversatrol, knotweed, milk thistle, quercitin, green tea and tobacco) being pushed by the gurus as helpful are actually harmful because they are anti-apoptosis and suppress the reactions; so people think they feel better because there are fewer free radicals but it really is just that they suppressed their immune system.
Let us take resveratrol/knotweed. Anyone thinking that resveratrol has as its primary effect an antioxidant effect is not following the science. The primary mechanism is via indirectly raising cAMP, therefore shifting the Ca++/cAMP axis. This profoundly alters intracellular signalling. The argument against resveratrol could be made on that basis though ... its a blunt instrument.
Or take alpha lipoic acid. In addition to being a critical antioxidant, its an essential factor in glucose metabolism.
I am less sure about milk thistle, green tea or quercetin. Tobacco I think is just a bad idea.
There is also chance. Things change, sometimes outside of our control. So if a change occurs, how do you know its not just some unanticipated or unmeasured factor? That is why science requires it be duplicated many times in many subjects. This is in hope that random factors can be ruled out. However even then there can be systematic unanticipated factors that skew an entire study. This could even be from something as simple as using the wrong assay.
One final point:
...it strongly supports the theory provided...
This suggests that there may be risk of confirmation bias. Hypotheses have to be robustly tested, in addition to having supporting data. N=1 is not sufficient. It is not clear how many people these factors have been tested on from your comment.
Please note this particular thread is relevant to experimental design, and reliability of information. A thread on a given hypothesis should probably be separate.