In cases like this, I think there are two issues that can help with this.
The first is a precautionary principle can be adopted. If it looks like there could be a problem, then a precautionary principle would suggest that you should assume there may well be one. That doesn't mean you necessarily shut everything down. Perhaps the recommendation to the minister would not go that far if there is so much uncertainty that we have to invoke the precautionary principle. It's well known that we should not be using that as a reason for inaction.
The other way to deal with this, though, is to look at the weight of evidence. We do that in science all the time. You could try to shoot down an individual study and say that it's just a correlation, for example, but eventually, the number of studies that are all pointing in the same direction by independent researchers could become overwhelming. The advice to the minister could be, “There is uncertainty in a complex problem, but the weight of evidence points to this, and, therefore, here are the potential options.”