I share your concerns, yet can't think of a solution. I think the only thing that can be done is to factor in a large amount of doubt, and amortize that doubt as more definitive studies come in over time.