The best and easiest place to start is by starting to document everything you can about this plant. Every preliminary investigation paper I've read starts by citing a history of use, case studies, anecdotal evidence, etc. supporting the investigation of the plant as a medicine.
You'll almost certainly find out interesting things. Perhaps different practitioners prepare or use the substance differently, or use it for different ailments.
Sounds like you already have your first case study, so just document that well and then get the who, what, why, where, when, and how of this plant's overall history of use.
A good place to start is to look at other similar studies. Check out Mullein leaf (Verbascum thapsus, Great mullein, Common mullein). There's a long history of use for inflammatory, lung conditions, hemorrhoids, etc.. They investigated the anti-inflammatory properties by subcutaneously injecting mice with an irritant and mullein extract then measured the injection site inflammation against the controls. They found it to be quite effective.
You may find in your background research some property of the plant that is easier to test than immunological effects.
A clinical study is probably beyond your reach at the moment, and is probably overkill at this point. But if people are using this remedy on a regular basis around you there's probably some research you can start right away. Taking lancet sized blood samples from people before and after treatment and then doing a white blood cell count might show something interesting.
Working with blood is something to be very careful with though. I'd suggest having the subjects lance themselves, put a drop on a slide, then apply the cover slip, and place it in a storage box for you. This way you can wait for any extra blood on the slides to dry and only work with the samples in your lab under safe conditions. Touching blood in the field is too dangerous, as can be seen by the large number of accidental sticks suffered by nurses each year.
With some good background info and a proposal, I bet you can find a local hospital lab that would help you with the counts, maybe even run all of them for you. I'm not sure what format the automated count machines take though.