Ask a clinic owner why patients drop off a GLP-1 or hormone program and you'll usually get one of three answers: cost, side effects, or "they got what they wanted and left."
Those are real. They're also the stated reasons, collected from the small minority of patients who bother to tell you. The larger group doesn't cancel. They just don't reorder. And when you look at when that happens, it clusters — somewhere after the initial excitement burns off and before the results are undeniable.
Call it week nine. The exact number varies by program. The mechanism doesn't.
The three phases of a treatment program
Weeks 1–4: novelty carries everything. The patient has just made a decision and spent money. They're motivated, they're paying attention, and the newness of the routine does the work. Adherence in this window tells you almost nothing about the patient's long-term behavior.
Weeks 5–10: the trough. Novelty is gone. Results, if they're coming, are real but undramatic — the kind you notice in a photo from two months ago, not in the mirror this morning. The routine has become a chore competing with every other chore. Meanwhile the first refill decision arrives, which means the patient has to actively choose to keep paying.
Week 11 onward: identity. Patients who cross the trough stop treating the program as something they're trying and start treating it as something they do. These patients are dramatically cheaper to keep than to replace.
The whole retention problem is that middle window. Everything a clinic can control comes down to getting more patients across it.
Why patients go quiet in the trough
Three things are happening at once, and they compound:
Progress becomes invisible. Weight loss slows and plateaus. Symptom relief becomes the new normal — a patient who no longer wakes at 3am doesn't celebrate that every night, they just forget it used to happen. Real progress stops producing the feeling of progress.
The routine loses its scaffolding. In week two, taking a dose is an event. In week eight, it's laundry. Missed doses start as accidents and become a pattern, and each missed dose makes the next one easier to miss.
The refill decision arrives at the worst possible moment. The patient is asked to spend again precisely when the program feels least remarkable. If reordering also requires effort — a phone call, a portal they've forgotten the login for, a message to your front desk during business hours — you've added friction to a decision that was already wobbling.
By the time a patient consciously decides to stop, the decision was made weeks earlier by drift.
What actually moves the needle
The interventions that work all attack the same thing: making an invisible process visible again, and removing effort from the moment of decision.
Make progress legible. A patient who logs a dose and sees a twelve-week streak has evidence that they're the kind of person who follows through. A weight chart that shows 15 pounds down since March reframes a plateau week as part of a trend rather than a failure. This isn't gamification for its own sake — it's replacing a feeling the biology stopped supplying.
Put the reorder inside the routine. Not an email campaign, not a call from your staff — the buy button in the same place the patient already goes, appearing when supply runs low. The best refill prompt is one the patient encounters while doing something they were doing anyway.
Catch drift early, with data you already have. A patient who logged every dose for six weeks and has now logged none for ten days is a different problem from a patient who was always sporadic. The first is recoverable with one well-timed message; the second needs a conversation. Most clinics can't tell them apart, because nobody is logging anything.
Ask before they decide. A short check-in before the refill — how's it going, any side effects, anything you want to change — does two jobs. It surfaces the fixable problems (a dose that's too aggressive, a side effect nobody mentioned) while they're still fixable. And it reminds the patient that someone is paying attention, which is most of what "care" feels like from the patient's side.
The uncomfortable part
All of this requires patient-level data your clinic probably isn't collecting. If your record of a patient's program is an order history, you know they bought something twice and then stopped. You can't see the six weeks of quiet drift that preceded it, because there was nothing recording it.
That's the real argument for a patient portal, and it isn't about having an app. It's that the tracking which keeps patients engaged is the same tracking that tells you which patients are slipping — the same daily action serves the patient and the clinic at once.
Everything else is a campaign fired into the dark.
Ideas here reflect how we've seen clinics run retention programs; every clinic's numbers differ, and the right intervention depends on your treatment mix and cadence. Nothing above is medical advice.