How Compounding Pharmacy Partnerships Actually Work for Telehealth Brands

What the pharmacy is actually doing in your program

In every arrangement we could verify, the compounding pharmacy describes itself the same way: the dispensing pharmacy of record, filling a valid, patient-specific prescription written by somebody else. That somebody else is your program's clinician. The pharmacy is not seeing your patient, not making the clinical decision, and not holding the patient relationship. It is receiving a prescription, compounding against it, and shipping.

That division matters more than it sounds, because it tells you where each obligation lands. The prescribing decision, the consultation record, and the patient relationship are yours (or your clinical entity's). The compounding itself, the sterility and potency of what ships, the state pharmacy licensure behind the shipment, and the pharmacy's own regulatory standing are the pharmacy's. What is genuinely shared is the part most agreements are vaguest about: what happens to your patients when the pharmacy has a problem.

So the useful mental model is not vendor and customer. It is two regulated operations, each of which can end the other's ability to sell. Your marketing can put a pharmacy's certification at risk, and a pharmacy's inspection can stop your fills with no notice you control.

• The pharmacy holds: compounding, sterility and potency, state pharmacy licensure, its own inspection and disciplinary record
• You hold: the prescriber relationship, the consult record, patient communications, marketing claims, and the brand the patient blames
• Shared, and usually underwritten: recall handling, patient notification, and what your program does during a supply stoppage
• Neither of you holds: FDA approval, because a compounded drug does not have one

The question that maps the whole relationship in one answer: if a batch fails potency testing after it ships, who tells my patients, in whose name, and on what timeline? Everything about how seriously a partner takes quality is in the specificity of that answer.

What is published, and what is not

On 2026-09-08, we read the public sites of all eleven compounding pharmacies in our directory, taking each pharmacy's own site rather than a summary of it. The surfaces are not uniform and the sample should not be described as though they were: most publish a partner or provider page, others had to be read from the public homepage because they publish no partner-facing equivalent, and the single pricing figure we found sits partly on a patient-facing page. The result is narrower and more useful than the version of this section we published first, which generalised from two pages and was wrong about one thing.

Pricing is rarely published, and the document you would actually negotiate against is rarer still: one of the eleven shows a figure of any kind, and a partner rate card is published by zero of the eleven. ProCompounding publishes a $10 flat-rate shipping charge on its provider page and cash-price floors on a patient page, and its partner rates are still behind a request form. The other ten publish nothing, several in so many words: Belmar says to contact them for formulary and pricing, Strive says an account representative walks through the tier structure, Empower's pricing link is a contact form. Minimums are the same story from the other side: one of the eleven addresses them at all, and Wells answers that there is no minimum order quantity or value.

Cold chain is the strongest version of the point. Four of the eleven name temperature-controlled shipping. Zero publish a temperature range, a packaging qualification, an excursion policy or a validation summary. For a category that ships biologically active injectables, the specification you would need in order to compare two partners was not published by any of them.

Licensure is where our first draft was wrong, and the correction matters more than the error. We wrote that nobody publishes the roster that would substantiate a coverage claim. One pharmacy does, comprehensively: ProCompounding publishes 63 permit rows covering all fifty states and the District of Columbia, each with the permit number, permit type and pharmacist-in-charge, and each with a link into that state's own board lookup so you can verify it rather than believe it. Four more publish a per-state roster without permit numbers, and two publish an individual state licence number inside a compliance document. So the correct thing to tell a buyer is not that this information is unavailable. It is that it is available from a minority, that asking for it is therefore reasonable rather than pushy, and that one competitor has already set the bar at a level the others have not met.

Turnaround is also published more often than not: seven of the eleven give a number, from a same-day claim to a three-to-five business day window. So of the six things a buyer most wants to compare, two are widely published, one is published by a useful minority, and three are effectively private.

The practical consequence: ask every pharmacy for all-in cost per fill at three monthly volumes in identical units, and ask for the licensure roster with permit numbers. The second request is not unreasonable, because at least one pharmacy in this category already publishes exactly that, unprompted, on its own website.

503A or 503B decides what they can legally make for you

Compounding runs on two different statutory chassis, and which one your partner sits on changes what it may compound, not merely how it operates. Section 503A covers compounding by a licensed pharmacist in a state-licensed pharmacy against patient-specific prescriptions. Section 503B covers registered outsourcing facilities, which may compound without a patient-specific prescription but, unlike 503A pharmacies, get no exemption from current good manufacturing practice requirements.

Both are restricted from making copies of commercially available drugs, and this is where most program assumptions break. Under 503A, a compounded drug is treated as essentially a copy when it has the same active ingredient, an easily substitutable strength (FDA reads that as within ten percent), and the same route of administration, unless the prescriber documents a change that produces a significant difference for that specific patient. FDA's own guidance gives examples of what qualifies, such as removing a dye for an allergy or a liquid form for a patient who cannot swallow a tablet.

Two details in that guidance do real damage to common pitches. First, FDA states directly that a lower price is not a sufficient basis for the significant-difference exception, which disposes of the most popular commercial argument for compounding a marketed drug. Second, the 503B test is stricter and shaped differently: a compounded drug is identical or nearly identical to an approved drug when the active ingredients, route, dosage form, strength and excipients match, and FDA states that this cannot be excused by a prescriber's determination of a clinical difference for an individual patient. Be precise about the reach of that, because the popular version of it is wrong: it governs the identical-or-nearly-identical branch only. Section 503B(d)(2)(B) covers a different case, a compounded drug that is not identical or nearly identical but contains a bulk drug substance that is a component of an approved drug, and there a prescriber's determination of a clinical difference for the patient does count. So on 503B you cannot reason your way out of a copy, while a genuinely changed formulation has its own pathway.

For GLP-1 programs specifically, the shortage-era arrangements are over and have been for more than a year. Semaglutide's shortage was resolved in February 2025 and tirzepatide's in December 2024, and FDA's enforcement-discretion windows closed behind them through spring 2025. With neither drug on the shortage list, the ordinary copy restrictions apply in full. There is also a live proceeding worth tracking rather than assuming: in April 2026 FDA proposed excluding semaglutide, tirzepatide and liraglutide from the 503B bulk drug substances list, finding no clinical need. As of September 8, 2026 that proposal was still pending, with the comment period closed and no final rule published. It is not in force, and a partner who tells you it is settled in either direction is telling you something they cannot know.

Licensure is per state, and the mechanisms are not the same

A pharmacy shipping into a state it is not licensed in is not a paperwork problem, it is an unlicensed dispensing problem, and it lands on the shipment you sold. The complication is that states do not use one mechanism, so a partner's compliance in one state tells you very little about another. Three examples, each checked against the state's own material, show three different shapes.

Florida requires a non-resident pharmacy registration for any pharmacy that ships, mails or delivers a dispensed drug into the state, and requires that the pharmacy and its prescription department manager be licensed where the pharmacy sits. Ohio licenses the same activity as a non-resident terminal distributor of dangerous drugs, requires an Ohio-licensed pharmacist as responsible person for a pharmacy that ships compounded drugs in, and imposes a self-reporting duty that is unusually useful to a buyer: the pharmacy must report any FDA warning letter, injunction or decree to the board within 72 hours of issuing or receiving it.

California is the one to understand in detail, because it has the sharpest tool. Its board can issue an immediate cease-and-desist order against a pharmacy compounding sterile products when it has a reasonable belief, based on an inspection or investigation, of an immediate threat to public health or safety. That order can stop shipments into the state within days and before a full hearing. California also requires compounding for use in the state to meet current United States Pharmacopeia standards, and its requirement that non-resident pharmacies have a California-licensed pharmacist-in-charge became operative on July 1, 2026.

The practical consequence for a program is that supply risk is not uniform across your map. A single board action can remove your largest state overnight while the other 49 keep selling, which means your continuity plan has to be per state, not national.

Ohio's 72-hour rule is a free diligence tool. A pharmacy that must self-report federal warning letters to a state board is a pharmacy whose regulatory history is checkable rather than negotiable, and the same letters sit in FDA's public warning letter database under the firm's name.

Certification makes your pharmacy your problem

If paid acquisition is anywhere in your plan, LegitScript certification is the gate, and its standards do not stop at your own operation. Its affiliates and partners standard states that a merchant's partners, including partner pharmacies responsible for fulfilling prescription medication to patients, are generally required to be LegitScript-certified or accredited by another recognized body, with limited exceptions at LegitScript's sole discretion.

That single sentence rewires the buying decision. A pharmacy that is cheaper and uncertified is not cheaper, it is a blocked advertising channel with a discount attached. And because certification runs on its own multi-month clock, discovering the constraint after signing costs a quarter you cannot buy back.

The application itself is a useful preview of what a serious partner should already have assembled, because it asks the pharmacy for exactly what you should be asking: every resident and non-resident license number with expiration dates and the pharmacist-in-charge for each, whether it subcontracts compounding to another facility and that facility's license, which independent lab tests its batches for sterility and potency and how often, what it does when a batch comes back sub-potent or non-sterile, what share of its volume ships out of state, any product recalls or adverse patient outcomes in the past year, and any FDA or DEA inspection findings from the past five years.

One nuance to hold onto, because it changes how much weight a badge deserves: this is an application-and-disclosure review. The pharmacy self-reports its licenses, its testing cadence and its history. Certification is meaningful evidence that a partner submitted to scrutiny and disclosed, and it is worth requiring. It is not an independent audit of every fact asserted, and it is not a substitute for reading the primary records yourself.

How these partnerships actually fail

The failure modes are not exotic, and they are documented in public records rather than inferred. Read one thing into them carefully, though: the first, third and fourth below come substantially from the record of a single pharmacy, Boothwyn Pharmacy, across a federal warning letter, a Pennsylvania penalty and a California board decision, all cited in full below. They are worth studying because they are unusually well documented, not because one firm proves a category. Treat each as a failure mode to ask about, not as a base rate.

The most consequential is releasing product before test results return. A 2025 state board decision documents sterile preparations shipped to patients before bacterial endotoxin testing came back, with results that then exceeded limits by two to three times or more, and no patient notification and no recall at the time. The same matter documents end-product testing data for several lots that falsely represented the lots had met testing requirements, with the pharmacy's own sterile lab director identified in the record as responsible for the falsification. The corrective SOP the pharmacy later adopted is the whole lesson in one line: no GLP-1 or testosterone-containing sterile preparation is released until passing sterility and endotoxin results are received and approved.

The second is potency drift, which is invisible to your patient and to you. Federal warning letters in this period name specific out-of-specification results on GLP-1 products, including semaglutide at roughly 80 percent of labeled potency and tirzepatide lots in the 86 to 89 percent range. A patient on an underpotent dose experiences your program as ineffective, churns, and tells people the brand did not work.

The third is unapproved space and undisclosed operations. One state action found roughly 30,000 doses of injectable weight-loss drugs compounded at over a thousand doses a day using rooms and a second building that had never been inspected or approved, and that staff did not tell inspectors compounding was happening there. It ended in a one million dollar civil penalty, probation, and a requirement to obtain national compounding accreditation before probation lifts.

The fourth is the cold-chain contradiction, which is the easiest of all to check yourself: compounded GLP-1 preparations labeled to be stored frozen while the approved products' own labeling says not to freeze and not to use if frozen. If your partner's storage instructions contradict the reference product's, ask what stability data supports theirs, and expect a document rather than an assurance.

Notice what all four have in common. None would have been visible in a demo, a sample vial, or a reference call with a happy customer. All four were visible in inspection records, board decisions and warning letters that were public and free to read.

The pattern behind the pattern: every one of these failures is a quality-system failure that shows up first as a paperwork answer. A partner who can produce a release SOP, a lab name, a testing cadence and five years of inspection history in one email is telling you something real. A partner who needs a week is telling you something too.

What belongs in the agreement

Most of what founders remember from a pharmacy sales call is unenforceable, because it was a description rather than a commitment. These are the terms worth insisting on in writing, and each one exists because its absence has cost somebody a program.

• The release rule: no product ships before passing sterility and potency results are received and approved, stated as an obligation rather than a practice
• The notification chain: who contacts patients on an out-of-specification result, in whose name, within how many hours, and who pays for the remake
• The licensure roster: the actual list of states with license numbers and expirations, warranted current, and updated to you when it changes
• The disclosure duty: the pharmacy tells you about warning letters, board actions, recalls and adverse outcomes on a defined clock, not when you ask
• The subcontracting answer: whether any of your volume is compounded by a facility other than the one you signed with, and its name and license
• The certification commitment: current LegitScript certification or recognized accreditation maintained for the term, since your advertising depends on it
• The continuity plan: what happens to in-flight patients during a stoppage, including who holds the records and how fast fills can move
• The cost structure: all-in per fill at defined volumes, in named units, with the conditions under which it changes

How EmbedCare handles this layer

EmbedCare runs an operated stack, which means the pharmacy relationship is ours to hold rather than yours to assemble: owned supply behind the program, flat product rates fixed in a signed partner agreement with medication included on GLP-1 programs, and the licensure, quality and certification obligations above sitting on our side of the line. Partners get the outcome the diligence in this guide is trying to reach without running the diligence themselves.

That is a reason to look at us, not a reason to skip the questions. If you are assembling your own stack, the companion checklist turns this guide into the specific questions to ask, and asking them is the single highest-leverage hour in the whole build. Ask us the same ones.