A telemedicine platform can have clinicians licensed in a dozen states, a tidy EHR, a pharmacy partner and real patient demand — and still watch its payment account get frozen on day nine. The founder usually assumes something broke. Nothing broke. The account was never underwritten in the first place. It was onboarded by an aggregator that reviews risk after the money starts moving, and once the model became visible — remote prescribing, auto-refills, a card charged every 30 days without the patient in the room — the account fell outside what that aggregator is willing to hold.
This is one of the least-understood operational risks in digital health. It is not a compliance failure and it is not fraud. It is a mismatch between how high-risk commerce actually works and how instant-signup payment tools are designed. Understanding the difference saves weeks of revenue and a great deal of avoidable panic.
The structural reason aggregators decline these verticals
Mainstream platforms like Stripe, Square and PayPal operate largely as payment facilitators. Thousands of small merchants sit beneath a shared master account, and the facilitator absorbs the aggregate risk. That model is excellent for a coffee shop and poor for a prescribing platform, because the facilitator cannot price individual risk — it can only exclude whole categories. Telemedicine, pharmaceuticals, supplements with therapeutic claims, peptides, nutraceutical subscriptions and prediction markets tend to land on restricted-business lists for the same cluster of reasons:
- Future delivery and recurring billing, which create chargeback exposure long after the sale settles.
- Regulatory variability across states and countries, which the facilitator cannot monitor merchant by merchant.
- Products where a dissatisfied customer disputes the clinical outcome rather than the transaction.
- Card-not-present volume with higher-than-average refund and friendly-fraud rates.
A dedicated merchant account flips the arrangement. You get your own MID, your own risk profile, your own reserve terms, and an acquiring bank that reviewed your model before approving it. The trade-off is real: paperwork up front, slower onboarding, and underwriting that asks uncomfortable questions. In exchange, the account is far less likely to be shut off mid-quarter. For anyone weighing the difference between facilitator convenience and a true merchant account, the U.S. Chamber of Commerce’s guide to choosing a credit card processing provider lays out the basic structural vocabulary worth knowing before you fill in a single application.
The underwriting file: what to assemble before you apply
Most declines in high-risk underwriting are not substantive. They are incomplete-file declines, and an incomplete file often gets coded as a risk decision that follows the business around. Build the package once, properly:
- Corporate and banking basics — formation documents, EIN letter, ownership breakdown with ID for anyone holding 25% or more, and a voided check or bank letter matching the legal entity name exactly.
- Processing history — three to six months of statements if you have them, including the chargeback and refund ratios. If you were terminated elsewhere, disclose it. Underwriters find it anyway, and a volunteered explanation reads very differently from a discovered one.
- Clinical and licensure proof — state medical licenses for your prescribing clinicians, the list of states you actually serve, your telehealth modality (synchronous video, store-and-forward, asynchronous intake), and how you handle states where your model is not permitted.
- Pharmacy and fulfillment chain — who dispenses, who ships, and who is on record as the merchant of record for each charge.
- The full customer journey — screenshots or a staging login showing intake, consent, pricing disclosure, subscription terms, cancellation path and refund policy.
- Website compliance surface — terms of service, privacy notice, HIPAA-relevant disclosures, a visible customer service phone or email, and a descriptor plan so the line item on a patient’s statement is recognizable.
MCC selection is not a formality
Telehealth typically sits under MCC 8099 (health practitioners, not elsewhere classified) or 8011 (physicians’ offices). The code you are assigned influences interchange treatment, FSA/HSA eligibility, issuer risk scoring and whether certain card-brand monitoring programs apply to you. A platform miscoded as general retail or a supplement seller loses benefits it should have and inherits scrutiny it should not. Get the code right at setup; changing it later is slower than you expect.
Recurring refills, tokens and the Account Updater problem
Refill billing is where telehealth economics live and where disputes concentrate. Three practical rules reduce most of the damage. First, store credentials as tokens in a vault, never in your own database, and keep the token portable enough that a processor change does not force every patient to re-enter a card. Second, run Account Updater against your recurring file so expired and reissued cards do not generate a decline cascade that looks like involuntary churn. Third, make the renewal visible: a pre-billing email a few days ahead, a clear descriptor, and a cancellation flow that takes one click rather than a support ticket. Card networks increasingly expect that transparency, and it quietly removes a large share of “I didn’t authorize this” disputes.
FSA/HSA acceptance and why it is worth the setup work
Benefits cards use an auto-substantiation system that checks the merchant category and, in many cases, line-item detail against eligible-expense rules. Accepting them properly means correct MCC classification, itemized receipts a patient can submit, and clean separation between eligible clinical services and ineligible retail items sold alongside them. The payoff is conversion: a consultation that can be paid with a benefits card converts measurably better than one that cannot, and patients who use benefits funds tend to cancel less often.
Chargeback defense as an operating discipline
Card brands watch dispute ratios, and sustained excess puts a MID into monitoring programs with fees and remediation deadlines attached. The defensible position is built before the dispute arrives: timestamped consent records, intake answers, clinician notes referenced by ID rather than reproduced, shipment tracking, cancellation logs and the exact subscription language the patient agreed to. Pre-dispute alerts let you refund a transaction before it becomes a chargeback, which protects the ratio even when it costs the sale. Automated fraud scoring on the front end — velocity checks, mismatched geolocation, disposable email patterns — stops a meaningful slice of the volume that would otherwise become a dispute six weeks later. Automation helps here, but it needs human review on the edge cases, a point that overlaps with the broader hidden risks of leaning on AI at work without someone accountable for the output.
Where a specialist desk earns its keep
Generalist sales agents submit high-risk files to generalist banks and the file dies in a queue. A specialist desk knows which acquirers currently have appetite for remote prescribing, how to present an asynchronous intake model so it reads as clinically supervised rather than unsupervised, and when a hybrid business should be split across multiple MIDs instead of forced into one. 2Accept works in exactly that lane, arranging merchant accounts for high-risk industries through a network of acquiring banks and running a dedicated underwriting desk for telemedicine and online prescribing platforms — HIPAA-compliant processing with API, hosted checkout or EHR payment-module integration, tokenized refill rebilling with Account Updater, FSA/HSA acceptance, chargeback alerts and dispute defense, plus multi-MID structures for brands selling to both consumers and businesses. The same desk handles other restricted verticals, including peptides and prediction markets. Nothing is approved without underwriting, and no honest partner will claim otherwise — but a file presented by someone who speaks the acquirer’s language gets a real reading instead of a reflexive no.
The founder’s checklist
Treat payments as core infrastructure, not a plugin. That means: apply as a merchant, not a hitchhiker on someone else’s master account; keep a second approved MID warm before you need it; own your tokens; document consent obsessively; reconcile refunds and chargebacks weekly rather than monthly; and read reserve, rolling-reserve and termination clauses as carefully as you read your clinical protocols. The platforms that survive their first dispute spike are rarely the ones with the cleverest funnel. They are the ones whose payment stack was built to be examined.
