But wait… Remember our original double-blinded trial was also “randomized”. That means patients with this mysterious Condition IE would be expected to land with equal frequency in both treatment arms. 20 should have also gone into the Eylea arm. It means we should have seen 90 good responders over there. But we only saw 70.
The checkbook doesn’t balance. And the most likely reason is that “Condition IE” doesn’t really exist and in our exuberance over the new kid Eylea, in our desperation to help patients failing on a medication, we’ve made biased observations in our follow on studies that are now un-randomized, un-blinded and anecdotal. Happens all the time. Maybe some of those 20 would have gotten better had they stayed on Lucentis. Maybe some snuck in from outside the original 100. Maybe some got other meds or PDT. Maybe they got better from a placebo effect - which can be quite powerful for both patient and doctor. Who knows? Otherwise we have to imagine a Condition anti-IE that curiously occurs with the same frequency as Condition IE and makes some cases of AMD susceptible to Lucentis but resistant to Eylea. We end up looking for snipes and fairies.
It’s why the FDA, as opposed to say the Dr. Oz show, requires hard data from double-blinded, randomized, controlled scientific trials. And it‘s why “superior to” proven by such a trial carries a lot more weight than “as effective as”, expert opinion, or anecdotal reports. In general, switching patients who’ve failed one medication to an “as effective as” medication from the same class is like rotating bald tires. Maybe it’s worth the time and effort… if that’s your only option.
It’s important to point out that this “cross-over” of the 30 patients was not actually done in View 1 or 2. It’s a hypothetical part of our thought experiment. Maybe 20, maybe 2, maybe none would have responded in real life. But Nexus is a real life cross-over of sorts. So let’s see what that might show us. (2 b continued)