Why Telehealth Startups Get Declined by Stripe — and What Underwriters Actually Want

A founder builds an online prescribing platform, wires up a payment aggregator in an afternoon, takes forty orders, and then gets an email: account under review. Payouts frozen. Three days later, closure notice, funds held 90 days. Nothing illegal happened. No chargebacks came in. The business simply ended up in a category the aggregator never intended to carry, and the automated risk system did what it was built to do.

This happens constantly to telehealth platforms, supplement and peptide sellers, prediction market operators, and anyone billing recurring subscriptions for something a bank considers medically or legally sensitive. Understanding why it happens — and what the alternative onboarding process actually involves — saves months of lost revenue.

Aggregators vs. dedicated merchant accounts

Mainstream platforms like Stripe, Square, and PayPal are payment facilitators. They hold one master merchant account with an acquiring bank and place thousands of sub-merchants underneath it. That architecture is what makes five-minute signup possible — nobody underwrites you individually at the start. It is also why a single risky vertical can be shut down wholesale: the facilitator carries the liability for every sub-merchant, so the cheapest risk control is a prohibited-business list and an algorithm that enforces it.

A dedicated merchant account flips the model. You are underwritten as yourself, you get your own merchant identification number, and the acquiring bank knows exactly what you sell, how you sell it, and what your refund behavior looks like. Onboarding is slower and the paperwork is real. In exchange, you are not one keyword away from a frozen balance.

What makes a business “high risk” in practice

  • Card-not-present sales with no physical fulfillment signature
  • Recurring billing, free trials, or auto-refill models that generate “I didn’t authorize this” disputes
  • Products whose legality or labeling varies by state or country
  • Prescription or practitioner involvement, which pulls in licensing questions
  • Future delivery or outcome-contingent payouts, as with prediction markets
  • Historic chargeback ratios above card-brand thresholds, often cited around the 0.9–1% range

None of these are disqualifying on their own. They simply move you from a self-serve lane into one where a human reads your file.

The underwriting file: what gets asked for

Underwriting for a restricted vertical is a document exercise, and the merchants who get approved are the ones who assemble the file before they apply rather than scrambling after the first information request. Expect to produce:

  1. Corporate identity — incorporation documents, ownership breakdown for anyone holding a meaningful stake, government ID, and beneficial ownership disclosure.
  2. Banking history — typically three to six months of business bank statements, plus prior processing statements if you have them. Previous chargeback data helps more than it hurts; concealing it is what kills applications.
  3. Website and checkout review — visible terms of service, refund and cancellation policy, shipping timelines, contact details, and clear recurring-billing disclosure with the amount and interval stated before the customer confirms.
  4. Product and claims documentation — certificates of analysis, supplier information, labeling, and any disclaimers. For telehealth, the clinical workflow: who prescribes, under what license, with what intake and identity verification.
  5. Volume projections — monthly processing estimate, average ticket, and highest expected single transaction. Underestimating to look safe backfires when real volume triggers a review.

The reason the file runs this deep is that acquiring banks sit inside a supervised framework where third-party relationships are the bank’s responsibility, not just the merchant’s. Regulators have long expected institutions to apply formal due diligence and ongoing monitoring to the parties they process for — the OCC’s guidance on payment processor relationships is one of the clearest statements of that expectation, and it explains why your refund policy wording is a bank-level concern rather than a marketing detail.

Recurring refill billing without the dispute spiral

Subscription prescription and refill models are where otherwise healthy telehealth businesses quietly damage their own risk profile. The mechanics matter:

  • Tokenized card storage. Cards live in a vault as tokens, not in your database. Rebills reference the token, which keeps PCI scope narrow and lets you move gateways without re-collecting payment details.
  • Account Updater. Reissued and expired cards are refreshed automatically. Without it, involuntary churn from declined rebills can run into double-digit percentages of a subscriber base over a year.
  • Pre-billing notice. An email three to five days before each refill charge converts a surprise into an expected one. Surprise is the single biggest driver of friendly fraud.
  • Descriptor clarity. The statement descriptor should match the brand the patient recognizes, with a working phone number. Unrecognizable descriptors generate disputes that never needed to exist.
  • Frictionless cancellation. If cancelling takes more clicks than subscribing, customers cancel through their bank instead — and that counts against you.

Chargeback defense then becomes winnable. A representment package with the signed consent record, intake timestamp, IP and device data, shipment tracking, pre-bill notice, and the cancellation policy the customer accepted is a genuinely different document from a one-line rebuttal. Prevention alerts help too, letting you refund before a dispute is formally filed and keeping your ratio below card-brand action thresholds.

Where a specialist provider changes the outcome

Most founders in restricted verticals do not need advice on how to be compliant so much as access to banks that will read their file at all. That is the gap a dedicated provider fills: packaging the application, placing it with an acquiring bank whose appetite matches the vertical, and wiring the gateway into whatever the business already runs. 2Accept works specifically with merchants that Stripe, Square, and PayPal decline — telemedicine and online prescribing platforms, peptide sellers, prediction market operators and other restricted categories — handling underwriting review, acquiring bank placement, gateway integration and chargeback defense as one process rather than four separate problems. Its high risk payment processing setup covers HIPAA-conscious telehealth checkout, EHR-native and hosted checkout options alongside a gateway API, subscription and refill billing with a tokenized vault plus Account Updater, and FSA/HSA card acceptance — with compliance documentation expected as part of the file, not waved through.

Payment rails, redundancy, and the crypto question

Merchants who have been shut off once rarely rely on a single rail again. Practical redundancy means a second approved merchant account, ACH or bank-debit as a secondary method for higher-ticket or recurring plans, and a gateway layer that can route between processors without a replatform. Some operators also look at digital assets as a fallback, and it is worth understanding how cryptocurrency works at a settlement level before adding it — the chargeback exposure disappears, but so does the dispute mechanism patients expect, and healthcare-adjacent businesses still need refund paths and clean accounting. Crypto complements card acceptance; it does not replace an underwritten merchant account.

A short pre-application checklist

  • Publish terms, refund, shipping and cancellation policies before you apply, not after a request
  • Disclose prior account closures and chargeback history up front
  • State recurring amounts and intervals at the point of consent and store the record
  • Match projected volume to reality, including your largest expected ticket
  • Keep product documentation — COAs, labels, supplier details — in one folder
  • Document the clinical or verification workflow in plain language a non-specialist reviewer can follow

Approval in a restricted vertical is less about persuasion than about being legible. The businesses that get placed quickly are the ones whose file answers the bank’s questions before they are asked — and whose billing design gives customers no reason to call their issuer instead of calling you.