Two routes to the same weekly injection are usually compared as though the only variable were where the conversation happens. On one side, a cash subscription that arrives by mail. On the other, an appointment, a prescription and a trip to a pharmacy counter. Set the subscription price against the copay and the cheaper one wins. That arithmetic fails at the first step, because the two routes mostly do not dispense the same drug — a difference in kind rather than degree, set out in the approval-status article.
The price gap is a product gap
Across the 477 sellers recorded here, 460 publish at least one price for a compounded preparation and 20 publish a price for the branded product. That is 958 compounded rows against 39 brand rows, at medians of $199 and $408 a month. The ratio is 2.05.
Almost none of that is the delivery mode. A compounded preparation is not FDA-approved, which means the agency does not verify its safety, effectiveness or quality before it is marketed, and it is not sold through the formulary system at all.[1] The gap between $199 and $408 is the price of that distinction, and it would still be there if both products were handed over in the same room. What the manufacturers themselves charge for cash purchases of the approved products is dated and quoted in the program-terms article.
What the subscription figure says it covers
Reading each row’s own note for what the seller states is inside the number gives a sparse picture. Shipping or delivery is named as included on 42 rows across 29 sellers. The clinical visit, consultation or provider review is named as included on 19 rows across 12. Needles, syringes or an injection kit appear on 11 rows across 8.
At the other end of the same market, 25 rows across 16 sellers publish a figure that is not a medication price at all — a care membership, a program start fee, a consultation charge or a monthly medical fee, with the drug billed separately and priced nowhere. So the visit is simultaneously the line item most buyers assume the subscription absorbs and the line item a minority of sellers charge for on its own, with the medicine left unquoted.
The item neither route prices is bloodwork
Of the 220 sellers written up individually, the laboratory question was settled for 72. Five state that bloodwork is included, eight state that it costs extra, and 59 say nothing about it.
On the in-person side the same item is usually billed by whoever draws the blood, under whatever coverage applies, which is a separate contract from the one that pays for the drug — the logic of which sits in the coverage article. Neither route, in other words, reliably puts a number on the test most prescribers want before and during treatment. A comparison that omits it from both sides is at least symmetrical; one that omits it from the cash side only is not.
One route settles the pharmacy question by default
A counter has a name, a license on the wall and a state board behind it. On the mail-order side that fact has to be published, and mostly is not. Of the 220 sellers written up here, the question of whether the preparing pharmacy is named was settled for 201: 32 name it and 169 do not — roughly one in six.
That is not a price difference and it changes what a price means. A buyer who cannot identify the compounder cannot check its licensure, its inspection record or its disciplinary history, and cannot hand the name to a prescriber holding the rest of the chart. The in-person route supplies that identification as a side effect of geography; the cash route supplies it when the seller chooses to.
Does the mode change the result?
The outcome literature says the two deliver comparably, with caveats. A systematic review and meta-analysis of four randomized trials covering 576 Black women with obesity compared remote against in-person weight-loss interventions and found no statistically significant difference in weight change (SMD −0.22; 95% CI, −0.68 to 0.24), in percentage weight loss (SMD −0.80; 95% CI, −3.86 to 2.26) or in body-mass index (SMD −0.26; 95% CI, −1.61 to 1.11), with evidence certainty ranging from moderate to very low.[2]
A Veterans Administration program evaluation compared pharmacotherapy-based obesity care delivered in person against the same care delivered by telephone. Over nine months, a newly enrolled virtual cohort (n=149) lost 14.4 ± 17.0 lb against 16.7 ± 21.0 lb in a historical in-person group (n=180), p = 0.44. Inside the cohort that transitioned mid-course, the in-person phase produced 6.5 ± 18.2 lb and the virtual phase 2.5 ± 13.3 lb, p = 0.22.[3]
Both comparisons point the same way and neither reaches significance. “No significant difference” is the correct summary of that study and it conceals a consistent direction in the point estimates, on small numbers with wide standard deviations. Quoting only the summary would be a true sentence describing a tidier picture than the data supports.
What direct-to-consumer telemedicine does deliver has been measured on its own terms. Among 966 patients prescribed liraglutide through a direct-to-consumer platform, 85.6% reported losing more than 2 kg by day fifty, averaging 4.9 kg; adherence to the prescribed regimen was 94.1%, adverse events were reported by 39.8%, and 86.4% wanted to continue.[4] That is a retrospective, self-reported, fifty-day snapshot of one molecule, which is a narrower claim than it first reads as.
Where the comparison breaks down
Three places, and the first is the largest. The compounded route is not the approved route, so the cheaper figure buys a preparation whose contents and quality nobody has verified in advance. That is not a reason to reject it and it is a reason to stop calling the difference a saving on convenience.
The second is that only one side of the comparison publishes a number. This site records 997 published cash figures. There is no equivalent published total for an office visit plus a lab panel plus a pharmacy fill, because each of those is priced per payer, per plan and per market; a household comparing the two is comparing a quoted price against an estimate of its own. Nothing here can supply the missing half, and inventing a national average would be worse than leaving the gap visible — which is the rule the methodology sets out.
The third is geography. 204 of the 477 sellers publish a state list at all; of those, 119 cover all fifty states and 85 cover fewer. A cash subscription is conditional on a license map that the buyer cannot see until the intake, while a local clinic is conditional on the buyer being able to reach it. Those are different constraints, and neither appears in a monthly figure. The ways a monthly figure can mean less than it looks like are collected in the presentation article.
Making the two routes comparable
Four line items settle it, and each one has a written answer on the cash side: the medication, the clinical visit, the supplies, and the shipping. Ask which of the four the subscription figure contains, and price the remainder separately. Then do the same on the clinic side, where the medication line is whatever a covered fill costs after a deductible and the visit line is whatever the office charges.
The default should run the unflattering way. Naming an inclusion is the exception in this data rather than the rule — 29 sellers for shipping, 12 for the visit, 8 for supplies, out of 477 — so an item a page does not mention is better treated as unpriced than as absorbed. That is not an accusation. It is what a blank field means when the field is usually blank, and the same principle governs how a missing figure is recorded here rather than filled in with a flattering guess.
Only after both lists are complete does a difference mean anything. The published cash figures, once every correction is applied, are collected on the price board, and a specific quote turns into an annual total through the cost calculator. The comparison that follows is between two totals, not between a subscription and a copay.
What this does not establish
Coverage is not census. Most of the 477 sellers here have not been individually adjudicated on what their figure includes, and a line item that appears in no note has not been checked and found absent — it has not been checked. The inclusion counts above describe the sellers who stated something, which is a self-selected and probably more forthcoming group.
Nor does any of this recommend a route. The randomized evidence on delivery mode is small and of mixed certainty, the observational evidence is either single-site or commercially operated, and none of it was designed to compare a compounded cash subscription against an approved product dispensed under coverage. That comparison has not been run, and the honest reading is that the money question and the outcome question have different answers on different evidence.