Operations

Why payment processors close telehealth accounts, and billing that survives

September 28, 2026 · 9 min read · Updated September 29, 2026

The first sign of trouble with a telehealth merchant account is usually an email. Payouts are paused pending review, a percentage of settlements is now being held in reserve, or the account is closed with funds retained for a period stated in the terms. Operators who go through this tend to blame HIPAA, or assume the processor objects to medicine in general. Neither is the reason. Processors underwrite money risk, and a direct-to-consumer telehealth brand combines three things that money-risk models dislike: the card is never present, the billing recurs, and the product includes prescription drugs shipped to a home. The fix is partly choosing a processor that underwrites the category on purpose, and mostly designing the billing so the risk the processor fears does not materialize.

This post covers why standard processors flag telehealth, what high-risk status costs in reserves and rates, what actually drives chargebacks in subscription programs, the descriptor and cancellation design that keeps ratios down, how HSA and FSA cards fit, and where certification and the card networks come in.

Why standard processors flag telehealth

The clearest public statement is Stripe's restricted-business list. Under the heading for pharmaceuticals, medical devices and telemedicine, it lists "Telemedicine and telehealth services," "Online pharmacies, including SaaS platforms," "Card-not-present prescription-only products and pharmaceuticals," and "Prescription delivery services" as restricted categories that require additional due diligence before an account is approved. Separately, its prohibited list includes "Pseudo-pharmaceuticals or nutraceuticals that are not safe or make harmful claims," and its general prohibitions reach businesses making "outrageous claims" or using "deceptive testimonials." Other mainstream processors publish similar lists with similar wording.

Read those categories together and the underwriting logic is plain. Card-not-present means the processor cannot rely on a chip or a signature, so fraud and "I did not authorize this" disputes are easier to file and harder to defend. Recurring billing means a single unhappy patient can generate several disputes across several months. Prescription products mean regulatory exposure: if a state board or the FDA acts against the clinic, the processor is left holding refunds for treatments it has already settled. And weight-loss and sexual-health marketing has a long history of claims that draw regulator and card-network attention. None of this involves HIPAA. A processor handling a card charge is not evaluating the clinic's privacy program; it is estimating how much of the clinic's revenue might come back as disputes and refunds after the clinic is gone.

Underwriting for the restricted category therefore asks for the things that reduce that estimate: provider licenses, the legal structure that separates the brand from the professional entity, the pharmacy relationships, a certification such as LegitScript, the refund and cancellation policy as written, sample descriptors and receipts, prior processing history if any, and financial statements. A brand that has these ready before applying is treated differently from one that discovers the list after its first payout hold. The pre-launch checklist on this site puts the merchant application in sequence with the rest.

What high-risk status costs

A brand that is approved as a high-risk merchant, whether by a mainstream processor after extra review or by a specialist acquirer, pays for the risk in ways that do not all show up on the rate card. The processing rate is higher than a retail merchant pays, and it is usually quoted as interchange plus a markup rather than a flat percentage, so it moves with the card mix. There is often a rolling reserve: a percentage of each settlement is held back and released months later, which means a growing brand is permanently lending its processor a slice of revenue. There can be a monthly volume cap that has to be renegotiated as the brand grows, longer settlement delays, monthly minimums, and an early termination fee.

The largest cost is the one nobody prices: a closure. Processor terms typically allow funds to be held for an extended period after termination to cover future disputes. A brand whose only account closes loses the ability to bill its existing subscribers, which in a medication program means refills stop unless a second account exists. Experienced operators keep a second merchant relationship warm from the start, split volume between them, and treat the merchant account as a single point of failure to be engineered around rather than a utility.

What drives chargebacks in subscription programs

Card-network monitoring programs measure a merchant's disputes as a ratio of transactions, and the thresholds are set low enough that a mid-sized brand can cross them without noticing, which brings fines and, eventually, termination. Chargebacks in telehealth subscriptions come from a short list of causes, and every one of them is a design flaw before it is a fraud problem.

  • An unrecognized charge: the statement descriptor is the MSO's legal name or a processor default, and the patient files a dispute against a charge from a company they have never heard of.
  • A forgotten subscription: the patient signed up for a program, stopped engaging, and treats the third month's rebill as a surprise.
  • A cancellation that did not stick: the patient asked to cancel by email or chat, the request was not processed before the next rebill, and the dispute is the patient's way of finishing the job.
  • A consult that did not result in a prescription: the patient paid, the provider declined for clinical reasons, and the refund either did not happen or took long enough that the patient disputed first.
  • A shipment that did not arrive on time: cold-chain delays, pharmacy backorders or address errors, with a patient who has already been charged and is not being told what is happening.
  • A price change the patient did not see coming: an introductory month converting to the full rate without a reminder, or a dose increase that changed the price without a fresh consent.

Very little of this is card fraud in the classic sense. It is communication and process failure that the dispute system catches because the dispute system is the only lever the patient believes will work. The retention post on this site, from visits to recurring care, makes the case that failed payments and drift are retention problems; disputes are the same problems viewed from the processor's side.

Descriptors, receipts and cancellation design

Each cause on that list has a specific fix, and together they form the billing design a processor wants to see in the application. The statement descriptor should carry the brand name the patient knows, plus a phone number or a short URL, and it should be set per product where the processor allows dynamic descriptors, so a consult charge and a medication charge read differently on the statement. Every charge should generate a receipt that repeats the amount, the next billing date and a working cancellation link. A reminder should go out before each rebill, and a longer, plainer one before any introductory price converts or any annual term renews.

Cancellation should be self-serve, in the patient portal, completed in the same session the patient started it, with a confirmation email. A retention offer is fine as long as the cancel button stays visible next to it. A pause option, where the patient can suspend billing and shipments for a stated period, prevents a large share of cancellations from becoming disputes, because a patient who can pause has no reason to fight a charge. Refunds for consults where the provider declined to prescribe should be automatic and same-day, triggered by the clinical outcome rather than by a support ticket, and the refund policy should say so in words on the checkout page.

Very little of this is card fraud in the classic sense. It is communication and process failure that the dispute system catches because the dispute system is the only lever the patient believes will work.

Behind the patient-facing design sits the recovery work: automatic card updates through the networks' updater services, retries on a sensible schedule, a grace window before refills stop, and a human touch for the cards that keep failing. Tessic Health's subscription billing includes failed-payment recovery for this reason, and its storefront and patient portal carry the brand's name rather than the platform's, so the name on the receipt is the name the patient signed up with. Whatever platform a brand uses, the questions to ask are whether it can set descriptors per charge, whether cancellation can be completed by the patient without a ticket, and whether refunds can be triggered by a clinical outcome.

Accepting HSA and FSA cards

Health savings account and flexible spending account cards are ordinary debit cards on the major networks, so a telehealth brand does not need a special agreement to accept them. Whether a charge goes through depends on how the merchant is coded. Merchants assigned a medical category code, such as a physician's office or a pharmacy, are generally treated as eligible at the network level, and the charge is approved on the assumption that the expense is medical. Merchants coded as general retail need an inventory approval system that flags eligible items at checkout, which is impractical for a clinic and is the reason a brand's merchant category code should be set correctly at onboarding rather than defaulted.

Eligibility of the expense itself is a separate matter that the patient's plan administrator decides. A consult is a medical service. A prescription medication is a medical expense. Whether a weight-loss medication or a hormone program is covered by a particular plan can depend on documentation, and some administrators ask for a letter of medical necessity from the treating provider. Supplements and skincare products without a prescription usually do not qualify. The brand's job is to issue itemized receipts that separate the consult, the medication and anything else, to make a medical-necessity letter easy for the provider to produce when a patient asks, and to avoid telling patients that a charge is HSA-eligible when the answer belongs to their administrator.

Certification and the card networks

The card networks run brand-protection programs that can penalize acquirers for merchants selling prescription products outside the networks' rules, which is one reason acquirers are cautious about the category. LegitScript certification is the credential Visa and Mastercard recognize for telehealth and online pharmacy merchants, and it is the same certification Google, Meta, Microsoft and TikTok require before prescription-drug ads run. LegitScript's published fees for healthcare certification are $975 per website to apply, $2,150 per year, and $2,500 for expedited review. For a brand, certification does double duty: it is a prerequisite for advertising, covered on the ad-approval guide, and it is the document that moves a merchant application from the restricted pile to the approved one.

Tessic Health prepares and files the LegitScript application as part of setup, alongside the MSO and friendly-PC structure that the certification review examines. A brand building independently should budget for the application early, because the review looks at the same things the processor does, and both want the professional entity, the pharmacy relationships and the marketing claims in order before approving anything.

Questions operators ask

  • Can a telehealth brand use a mainstream processor at all? Often yes, after the additional review that the restricted-business category triggers. The application goes better with licenses, structure, certification and a written refund and cancellation policy attached, and it goes worse when the brand's site is still running before-and-after claims.
  • What should the reserve look like? A rolling reserve expressed as a percentage of settlements, released on a stated schedule, is standard for the category. The figures are negotiable with processing history, and they should be revisited after the first few months of clean ratios.
  • Should consults and medication be billed as separate charges? Usually yes. Separate charges with separate descriptors make statements legible, make HSA and FSA receipts cleaner, and let a declined-consult refund happen without touching a medication charge.
  • Does a subscription need a new consent every time the price changes? A price change the patient did not agree to is a dispute waiting to happen and, in states with auto-renewal statutes, a legal problem. Treat any change in amount as needing fresh disclosure and consent.
  • What if the account is closed anyway? Funds are held for the period the terms state, and the brand bills through its second account while it appeals. The lesson is to have the second account before it is needed.

Payment processors are not hostile to medicine. They are hostile to surprise charges from unfamiliar names for products that might not arrive, and a subscription telehealth brand can look exactly like that from the outside if its billing is not designed. A recognizable descriptor, a receipt with a cancel link, a pause button, an automatic refund for a declined consult and a certification on file are what turn a restricted-category applicant into a merchant the processor is happy to keep.