I'm inclined to agree with Gator on experimenting a bit with slightly lower doses. I didn't really want to say that since I have been overly cautious at times with Bix on reducing the dose, so I didn't want to put my bias (play it safe, and rule out rebound, and then raise the dose more confidently if it's ruled out) on you in case I was being overly conservative. But I'm happy to second the idea.
As for holding the dose a few cycles, I see merit in it for dose increases, but not as much for dose decreases. If you decrease and can see in the 1st cycle it is too little, no point in holding there for longer (unless you are waiting for rebound to clear or something like that). If it appears to be too much, for safety I think you have to decrease, although it is possible that in a few cycles when overlap wears off if would be the right dose, but I couldn't in good conscience advise someone to hold a dose that is clearly too high on that basis.
With dose increases I found the overlap can build up nicely and give you a dramatic improvement around cycle 3 or 4. With Bix it was fairly predictable once I had enough data. With him, I learned to basically ignore the results I got in cycles 1 & 2 on a higher dose, and look to cycles 3 & 4 for good data. But ECID, and Bix got good duration on PZI, so overlap was always a key factor for him in how he would respond. If a cat isn't getting good overlap, it's probably academic as I doubt you'd see much improvement at cycle 3 or 4 without that.
As for onset, I was curious in terms of calculating how fast that drop by +3 happened. If PZIR onsets at +1 then it's not so steep, but if it doesn't kick in til +2.5 say.... well then, yowza! I never could figure it out for Bix, as he would get an earlier drop with good overlap vs. not, so it was hard for me to sift through the different factors and identify it. I hadn't read the study, that's interesting about the data points moving around in the first 45 days. Weird.
Thanks Gator for clarifying things - I agree there is a more aggressive approach one could take, or the more cautious one. I tend to err on the side of caution as long as you don't have anything like ketones in the picture. With Bix, I ended up going back to the highest dose that I felt 99% sure was not causing rebound, and then working back up from there. It probably wasn't the quickest way to get him regulated, but it worked, and stopped me from going around in circles on what to do with the dose.
[ETA: On 2/23 & 2/24 with the 1.5 my guess would be that was shooting through rebound. I'm thinking the reason it stayed a U-curve towards the end of the cycle is because the dose was high enough to cancel out the rebound going on. That could be why the next cycle was all reds on the 0.8 - just not enough insulin to even touch the rebound left over from the last cycle. Not 100% sure, but that's what looks like a possible explanation to me.]
