>>129694176 that they found. You can't use 6.8m as the denominator, because you don't know how many people didn't report the side effect or didn't detect them. We can't estimate true prevalence without knowing the number of people who would've reported if they'd noticed something (a likely unknowable quantity). In other words, you're trying to solve:
6/x = y/6.8million
Without knowing either x or y.
