Incidence rate per 200,000

Over on this thread (http://www.nuc.berkeley.edu/node/2260), Professor Chivers agreed that using the Linear No Threshold Model, there's a 1 in 200,000 possible cancer incidence for the 0.7 Bq/L measured in the UCB milk sample.

My question is what does the ratio of cancer per 200,000 have to increase to before it's "a level of public concern?" Is there a standard risk/regulatory tolerance that's used as an acceptable limit in the US?

Thanks