Here is where some scientific study could be very useful. Does breaking the spell really work, or does it only appear to work? Any kind of "fake" treatment or therapy you can think of will sometimes appear to work if we don't meticulously record each trial, because people remember all the successes and forget all the failures. Even a sham treatment will "succeed" sometimes, because most ailments in most people will eventually get better, at least temporarily, given enough time, and it is natural for people to seize on these moments and interpret them as strong evidence the fake treatment "worked". Even if the fake treatment doesn't work, people don't often interpret this as evidence the treatment is fake .... it just didn't work "this time".
Another question: if a spell does work, or if breaking spells does work, does it work by breaking a "real" spell, or does it work by making the person
believe a spell on them has been broken? This is an interesting distinction. For example, if you give patients something that does absolutely nothing, like a sugar pill -- what is called a "placebo" -- this will have some positive statistical effect on people's ailments. It can even be a pretty significant effect. It works better if people believe it will work. This is called the placebo effect, it has been well documented. See:
Placebo Effect: A Cure in the Mind: Scientific American and
placebo effect - The Skeptic's Dictionary - Skepdic.com
You can control for the placebo effect by measuring the treatment outcomes of people who are treated for spells vs. people who are not treated for spells, where neither group knows they are being treated for spells. Or, you could make both groups *believe* they are being treated for magic spells, but one treatment is fake and the other is "real".
The only question is, is the effect strong enough to be measured? In high quality studies with hundreds of rats and sometimes thousands of human subjects, it has been possible to measure effects as tiny as the rate of cancer changing from 0.025% to 0.050%, due to the influence of some drug, if I remember one study I saw. If magic spells are real we should be able to measure them, UNLESS their effects are so weak, and so rare compared to other factors like washing your hands or getting enough iron in your diet, that we can't even measure it at all. That's still an interesting result, definitely worth knowing. If magic spells have a negligible effect on people's health compared to other factors, then it *might* exist but for all practical purposes, it is not something to be worried about.