I think you've underestimated the affected population by not including those who aged out of the group close to the beginning or aged in close to the end. An 18 year old at the start of the period would have been 32 by the end, meaning (32-10) = 22 years of girls were affected. Using your population numbers, I get around 38,000 ( = 13,900/8 girls per year group * 22 year groups).

Which puts it at a still horrifically high 1 in 27 and the same conclusion.