Prime Therapeutics found only about 1 in 12 patients who started a GLP-1 for obesity were still on therapy at three years. Every dollar you spend on acquisition runs through that number first.
Key Takeaways
- Prime Therapeutics reported that only 8 percent – about 1 in 12 – of people who started a GLP-1 for obesity without diabetes were still on therapy at three years, and 14 percent for the high-potency obesity-approved products.
- The same research showed one-year persistence improving sharply, from 33 percent of 2021 initiators to 63 percent of early-2024 initiators, which means persistence is movable rather than fixed.
- Published health-system data reports materially different weight outcomes by persistence, which makes retention a clinical quality issue and a financial one at the same time.
- Median months enrolled is the highest-leverage input in your model. Improving it costs less than acquiring the equivalent revenue and it compounds.
Prime Therapeutics has now published three years of persistence research on GLP-1 therapy for obesity, and the headline finding from June 2025 is the most important number in this specialty: only 8 percent – about 1 in 12 – of people who initiated a GLP-1 for obesity without diabetes were still on therapy at three years. For the high-potency, obesity-approved products the figure was 14 percent. The prior year’s analysis found 85 percent of new initiators were no longer taking the drug after two years.
Read alongside that, though, is a finding most people skip. In the same June 2025 release, Prime reported one-year persistence rising from 33 percent among 2021 initiators to 63 percent among those starting in the first quarter of 2024 – a near doubling, which Prime attributed largely to the resolution of the drug shortages. Persistence is not a fixed property of the patient population. It is substantially a function of supply, cost, coverage and program design, and three of those four are things a practice can influence.
The financial consequence is arithmetic. If your contribution is $180 per enrolled patient per month, a median tenure of four months yields $720 per patient and a median tenure of nine months yields $1,620. Fully loaded acquisition cost does not change between those two scenarios. Nothing in your marketing budget produces a swing that large for that little money. And unlike acquisition, retention compounds: the patient who stays through month nine is also the patient who refers.
Attrition Is Concentrated, and It Is Early
The drop-off is not evenly distributed across the year. It clusters in the first eight to twelve weeks, and the reasons are predictable enough that you can build a program around them. Side effects during titration, arriving before meaningful results and before the patient has built any loyalty to the program. Cost shock, when an introductory price expires or a coverage determination comes back denied. Access friction, when a refill lapses or a prior authorization stalls. And expectation mismatch, when the patient came in with a number in their head that nobody ever discussed with them at intake.
Each of those has an operational counter, and none of them is clinical care. Published health-system research from the Cleveland Clinic group examining adults initiating semaglutide or tirzepatide reported materially different weight outcomes by persistence status – patients who discontinued within three months, those who discontinued between three and twelve months, and those who persisted a full year were not in the same place at the end. That finding matters because it collapses the false distinction between a retention problem and a quality problem. In this specialty they are the same problem.
Cost and coverage sit underneath most of it. KFF’s 2025 employer survey found only 16 percent of firms with 200 to 999 workers covering GLP-1s primarily for weight loss, rising to 43 percent at firms with 5,000 or more. KFF has also reported that 56 percent of GLP-1 users say the drugs were difficult to afford, including one in four who say very difficult. Your patient’s twelfth month is a financing question as much as a clinical one, and a program that never raises financing until the patient is already gone is leaving revenue on the table.
You cannot manage attrition you have never measured. Most practices know how many patients they enrolled last month and cannot say how many of last spring’s cohort are still here.
Measure Survival by Cohort, Not Headcount
Active patient count is a vanity number because it nets new enrollments against departures and hides both. Replace it with four measurements, run monthly:
- Cohort survival curves. Group patients by enrollment month and report the percentage still enrolled at 30, 60, 90, 180 and 365 days. Six months of this and the shape of your problem is obvious, including whether it is getting better or worse.
- Median months enrolled, trended. The single number to put in front of the leadership team. It goes directly into the contribution model.
- Time-to-second-visit and time-to-first-refill. Leading indicators. A patient who has not returned by day 45 is largely gone, and you want that flagged on day 32, not discovered at the quarterly review.
- Reason at exit, captured consistently. Cost, side effects, coverage denial, results, life event, lost contact. Guessing at the mix is how practices end up solving the wrong problem loudly.
Program Design That Holds Patients
The design changes that move these numbers are unremarkable individually. Set expectations explicitly at intake, in writing, about pace and about what happens when therapy stops – patients who are surprised at month three leave at month three. Front-load contact into the first eight weeks, when the risk is concentrated, rather than distributing visits evenly across the year. Own the refill calendar as a practice function with a named owner and a worklist, rather than treating it as the patient’s responsibility. Build a structured side-effect response pathway so a patient who feels unwell on a Saturday reaches your protocol instead of the internet. Discuss cost trajectory before it becomes an emergency, including what happens when an introductory price expires. And build the non-pharmacologic components – nutrition, activity, behavioral support – as things the patient is actually enrolled in, not brochures, because those are what carry a patient through the periods when medication access is disrupted.
One structural note: KFF’s 2025 employer survey found that 34 percent of firms covering these drugs for weight loss require the enrollee to meet with a dietitian, case manager or therapist, or participate in a lifestyle program, in order to receive coverage. Employers are, in effect, mandating the wraparound service line. A practice that has built and documented that program has something to sell that the direct-to-consumer platforms structurally do not.
From the Field
A three-location medical weight loss group in the Midwest, about 1,400 active patients, was enrolling well and could not understand why revenue had gone flat. There was no retention reporting at all – the dashboard showed active patients, which had been stable for a year because new enrollments were almost exactly replacing departures. Cohort analysis showed median tenure at 4.6 months, with more than a third of patients gone by day 75. Our fractional COO engagement rebuilt the schedule template to front-load three contacts into the first six weeks, created a refill worklist owned by a named medical assistant at each site with a 45-day trigger, wrote and rehearsed an intake expectation script with the front desk at all three locations over two weeks, and installed a day-32 no-return flag in the EHR. We ran the weekly retention huddle ourselves for the first quarter, then handed it to the practice administrator. Median tenure reached 7.1 months over the following three quarters. Enrollment volume was essentially unchanged; revenue was not.
Advice, Execution, and Which One You Need
The analysis in this article is not proprietary. A capable consultant will build your cohort curves, identify the drop-off points, and hand you a program redesign and a workflow map. If you have a strong practice administrator with real authority and open hours, take the plan and run it – that is frequently the better economic choice and we say so on the call. Consulting is not a downgrade; it is a different allocation of who does the work.
But retention is won in behavior, not documents. The gap between a workflow map and a changed practice is a person standing at the front desk while the new intake script gets said out loud fifteen times, sitting with the medical assistant while the refill worklist is worked for the first time, and running the retention huddle in week three when everyone’s enthusiasm has worn off. That is what a fractional executive does after the analysis: the same diagnosis, then the hours inside your practice with your staff until the habit outlives the engagement.
Ask the direct question before you sign with anyone, us included: will this person be working in our systems with our people, or delivering a plan for us to implement? Both are honest offers, and the right answer depends entirely on whether you have the internal capacity to execute. The difference is hours per week and who does the work – not quality.
Sources
- Prime Therapeutics, Only 1 in 12 Remain on a GLP-1 Drug for Obesity at Three Years (June 25, 2025) — https://www.primetherapeutics.com/w/prime-therapeutics-leading-research-shows-only-1-in-12-remain-on-a-glp-1-drug-for-obesity-at-three-years
- Prime Therapeutics, 1 in 7 Stays on GLP-1 Drugs for Weight Loss After Two Years (July 10, 2024) — https://www.primetherapeutics.com/w/prime-continues-to-lead-industry-on-glp-1-research-1-in-7-stays-on-glp-1-drugs-for-weight-loss-after-two-years
- Gasoyan H, et al., Obesity (2025) – Discontinuation and weight outcomes among adults initiating semaglutide or tirzepatide — https://onlinelibrary.wiley.com/doi/10.1002/oby.24331
- KFF, 2025 Employer Health Benefits Survey (October 22, 2025) — https://www.kff.org/health-costs/2025-employer-health-benefits-survey/
- KFF, What to Know About the BALANCE Model for GLP-1s in Medicare and Medicaid and the Medicare GLP-1 Bridge (May 11, 2026) — https://www.kff.org/medicare/what-to-know-about-the-balance-model-for-glp-1s-in-medicare-and-medicaid/
- American Medical Association, 2025 Prior Authorization Physician Survey — https://www.ama-assn.org/system/files/prior-authorization-survey.pdf
More in the Medical Weight Loss Series
- The Cash Price of GLP-1s Fell by Two Thirds. Where Is Your Margin Now? — what a GLP-1 program actually earns once the drug stops being the margin.
- Three Compliance Fronts That Can End a Weight Loss Practice – and None of Them Are Clinical — the compounding, telehealth and advertising rules that moved under this specialty between 2024 and 2026.
- Insurance or Cash in Obesity Medicine: Run the Numbers First — what the coverage landscape actually supports, and how hybrid models are built.
- Scaling With APPs and Second Sites: What Actually Breaks — what breaks when you add a provider or a second location, and the order to fix it in.