Not the original commenter, but the math is (making some implicit, but arguably reasonable assumptions):

Probability that someone in the population has schizophrenia = (1870/500000) = 0.00374

Probability that someone does NOT have schizophrenia = (1 - 0.00374)

Then if we assume that blind people have the same rate of schizophrenia as the population, Probability that 66 blind people ALL don't have schizophrenia = (1 - 0.00374)^66 = 0.78

1870/500000*66 = 0.24684. However, it's "nearly half a million", so let's call it 30000 as a conservative estimate: that's still 0.4114 children in expectance, which isn't very many.