The wording “X causes Y” is slippery. Saying “X increases the probability of Y”, is another thing. Thinking about smoke, with “smoke causes cancer” I expect every smoker to have cancer. It isn't so. If the effect of X on the probability that Y occurs is too much small, maybe it's just that we haven't enough statistical data (yet) to see it. Then, even if the effect could be statistically seen, telco companies could always say that it's safer than driving a car… Or a new business for body-cellphones shields could be made… Or, more likely, some people will pay more attention at to where they put their cellphones.