I want to move my injection day. How much drift does the label actually allow, and in which direction?

My weekly dose currently lands on a Thursday and that is now the worst day of my week for it. I want it on a Sunday. That is a three-day move and I can see two ways to make it:

• Shorten once. Go Thursday, then the following Sunday — a four-day interval — and weekly after that.
• Lengthen once. Go Thursday, then the Sunday ten days later, and weekly after that.
• Or drift. Move it by one day a week for three weeks, so intervals of 6, 6, 6 days.

The labels permit a day change subject to a minimum interval, so all three appear to be allowed on paper. But allowed and equivalent are different things, and I would like to know which perturbation is smaller and why. Ideally with the exposure arithmetic, because I suspect the two directions are not symmetric and I cannot see immediately which way round it goes.

A reason to prefer the single-step approaches over the drift that has nothing to do with exposure: a drifting schedule destroys your ability to attribute symptoms.

People on this class typically develop a reproducible weekly symptom pattern — some describe a first-48-hours effect, some a mid-week trough in appetite suppression, some nothing at all. Whatever the pattern, its value is that it is stable, because a stable pattern lets you notice a change. A new symptom appearing on day 3 is informative if day 3 has looked the same for eleven weeks.

During three weeks of drift, the day of the week and the day of the cycle are both moving, and you have lost the ability to distinguish "this is the day-3 effect" from "this is a Wednesday effect" from "this is new". If anything goes wrong during a drift, the drift is a confounder you introduced yourself.

Practical version of the argument: make the change once, on a week where nothing else is changing, and hold everything else constant across it. Not simultaneously with an escalation, not in the week of a new vial, not while travelling across time zones. One variable at a time is not a fussy laboratory principle here, it is the only way a weekly cycle with a sample size of one is interpretable at all.

On the direction, I mildly disagree with the accepted answer's framing on practical rather than arithmetic grounds. Shortening once means the change is complete immediately and the following week is already normal. Lengthening once means a three-day stretch where exposure is at its lowest point in the whole schedule, and for anyone whose appetite signal is noticeably time-varying across a normal week, that stretch is the part of the week they will notice most. The arithmetic favours lengthening by a few percentage points; the experience reported by people who have moved a day tends to favour getting it over with. Both are inside the label. Neither is a decision this board should be making for anyone.

Time zones, since the question is really about calendars and nobody has raised the case that actually catches people out.

Crossing time zones changes the clock but not the elapsed time. Flying east across eight hours and dosing "same day, same local time" is a 16-hour-shorter interval; flying west is 32 hours longer if you also keep the same nominal weekday. Neither of these is remotely enough to matter for a 7-day-half-life agent — a 16-hour compression on a 168-hour interval is a 10 % interval change, producing about a 3 % peak change by the arithmetic above.

What does matter, and only for the cold-chain question rather than the pharmacokinetics, is that a long trip converts a refrigerated storage problem into an ambient one. That is a different discussion, but it is the reason travel weeks are the weeks people actually get wrong: not because of the schedule, because of the logistics.

One genuinely useful convention if you travel often: track the schedule in elapsed days from the last administration rather than in weekdays. A weekday is a label on a calendar; the drug responds to elapsed hours. Once you are counting elapsed days, a time-zone shift becomes visibly irrelevant and a genuine two-day slip becomes visible instead of hiding behind the fact that it is still nominally Thursday.