What I’m arguing is that we don’t need complete phase 3 trials to establish efficacy - we can establish that based on early data from those (and earlier) trials

What I’m arguing is that we don’t need complete phase 3 trials to establish efficacy - we can establish that based on early data from those (and earlier) trials, while accepting larger error bars on the risk of side effects.