telehealth · regulated markets
What Telehealth Medical Claims Language Triggers FDA Letters?
The Dr. John Delony Show, a 1.55M-subscriber BetterHelp partner, has run 18 BetterHelp reads since January 2026. BetterHelp is the virtual therapy company. One slot every two weeks, every script cleared, zero FDA letters on file. A head of growth at a TRT clinic messaged me last Tuesday asking why her creator's last post got pulled inside 48 hours. The 90-second answer was that the creator said the word "replaced" on camera. That one verb flipped the post from testimonial into a medical claim her brand was not registered to make.
Across the 1,700 telehealth creators we track running BetterHelp, BlueChew (men's ED telehealth), Marek Health (a TRT and peptide clinic), Keeps (hair-loss telehealth), and Talkiatry (telehealth psychiatry), the reads that survive review share a five-word vocabulary. The reads that draw warning letters share four banned adjectives.
Testimonial vs claim
How does a creator turn a story about themselves into a regulated medical claim the brand owns?
"I feel better since I started" is a testimonial. The creator owns it. "This drug cured my anxiety" is a claim. The brand owns it. The FDA reads the second as off-label promotion if the cure language is not in the approved label.
The safe pattern looks like "I started the program in March and by May I noticed a real change." No mechanism claim. No comparative claim. Crystal Park, a 53.9K-subscriber BetterHelp partner, has shipped 30 BetterHelp reads since October 2024 on that frame. Simonesimmo (362K subs) has shipped 29 since August 2023. Neither has drawn a takedown.
Most teams under-spend on script review and over-spend on production. Here is the line-by-line script-review pass we run before any telehealth post ships.
Four risk archetypes
What are the exact four adjective patterns that flip a creator post from testimonial into actionable claim?
The four archetypes are cured, cured-faster, outperformed, and replaced. Cured says the product made a condition go away. Cured-faster says it worked in less time than the standard of care. Outperformed compares the product to a named competitor or to "what the doctor prescribed." Replaced says the product substituted for a clinical step the brand is not licensed to substitute.
A creator who says "BlueChew worked faster than the Viagra my doctor gave me" hits cured-faster and outperformed in one sentence. The brand owns both claims the moment the post goes live. Steve-O's Wild Ride Podcast (1.96M subs) has run 37 BlueChew slots since September 2024. None of those reads use any of the four archetypes. That vocabulary is the reason the channel gets rebooked while the freestylers get dropped after one cycle.
Worried your next telehealth read might hit one of the four banned adjectives? The cured / cured-faster / outperformed / replaced split is the same on every regulated-category post we ship. We can flag the risky phrasing on your current script in 48 hours, no pitch attached, just the marked-up read back to you.
Send us your script for a free review →Compounded vs branded language
When can a creator say the branded name versus when must they say "compounded" or "generic"?
Level 1 is the creator who says "I take Ozempic" when the script called out a compounded formulation. That mismatch is the cleanest setup for an FDA letter. The brand is shipping a compounded version it cannot legally market under the branded name.
Level 2 is the creator who says "I take a compounded version of Ozempic." Correct on substance. Still risky. The comparison to the branded drug invites the same FDA scrutiny.
Level 3 is the creator who says "I take a prescription from my telehealth provider." No drug name. Boring on camera. Bulletproof on review.
The GLP-1 wave through 2024 and 2025 ran the same Level 1 to Level 3 progression on hundreds of paid posts. The brands that survived the FDA's September 2025 sweep of 30 compounded-GLP-1 telehealth marketers were the ones whose creators landed at Level 3 before the first warning letter went out.
Mark Bell's Power Project, a 384K-subscriber show, is running this shape for Marek Health right now. 20 Marek Health deals have closed since December 2025. Every read uses the Level 3 frame.
- Reads pulled inside 48 hours for a single cure-language sentence
- FDA letters that follow the brand through every renewal cycle
- State medical-board inquiries that freeze ad spend for a quarter
Our main goal is to drive traffic to the quiz and convert users to purchase the blood work, but past efforts didn't translate into sales.— Marcella Taranto, Feel30 · discovery callGet the five-word vocabulary, free →
Before-and-after photos and age floors
What visual and demographic rules cut a post even when the script passes legal review?
Before-and-after pacing is banned in most medical verticals on the platforms that matter. Meta, TikTok, and YouTube ads all ban it. The ban runs at the platform-policy layer before FDA ever fires. A side-by-side photo of a creator at week 0 and week 12 reads as a results-promise in automated review. The post gets pulled in the time it takes the model to score the frame.
Age-restricted claims add a second floor. TRT targets 30+ men. Weight-loss GLP-1s require 18+ or 21+ depending on state. Mental-health Rx targeting under-18 viewers triggers a separate FDA review track. Anne of All Trades, a 669K-subscriber BetterHelp partner, has shipped 22 BetterHelp reads since February 2024. Every read keeps the audience disclosure in the description and avoids any under-18 framing in the read.
The "I am not a doctor, please talk to yours" sentence is doing more legal work than the #ad disclosure itself. Here is the disclosure-line library we keep for regulated briefs.
What survives review
What is the five-word working vocabulary that lets a creator talk about a telehealth product without triggering review?
The five-word working vocabulary is noticed, started, routine, convenient, and for me. "I noticed a real change after I started the program" is a testimonial. "This is part of my morning routine now" is a behavior claim, not a medical one. "It was convenient, I did the whole intake in 15 minutes" is a service claim about the platform, not the drug. "It worked for me" is the personal-experience hedge that the FTC's Endorsements, Influencers, and Reviews guidance blesses.
Would the brand lose sign-ups if the creator stopped saying "cured" and started saying "noticed"? Brands in our deal log that ran both versions saw the conversion gap land inside the noise floor. The cure language adds risk without adding sign-ups.
Tyler and Todd (271K subs) have shipped 29 BetterHelp reads since August 2023 on this vocabulary. Max & Occy (308K subs) have shipped 32. Eamon & Bec (1.28M subs) have shipped 19 since June 2023. The ones who freestyled a "cured" or "replaced" line on read one quietly drop off the roster.
Where We Come In
We do the line-by-line script review for you. The four risk archetypes and the compounded-versus-branded ladder repeat on every paid post we ship. We already track which creators across the 1,700 telehealth names in our database have shipped hundreds of reads without one takedown or FDA letter. The bounded downside is one extra review pass on a script that was probably close. The unbounded upside is a paid post that runs the full 90-day pilot instead of getting pulled in week one.
Send us your next telehealth script and we will mark up the risk lines before the camera turns on.
Compliance is the moat.
FAQ
What words trigger an FDA warning letter on a telehealth post?
Four adjective patterns flip a creator post from testimonial into actionable claim. Cured. Cured-faster. Outperformed. Replaced. Each one moves the line from a personal story to a medical claim the brand owns. Safe substitutes include noticed, started, routine, convenient, and for me.
Can a creator say a telehealth Rx 'cured' their condition?
No. "I feel better since I started" is a testimonial the creator owns. "This drug cured my anxiety" is a claim the brand is responsible for. The brand owns every cure-language sentence on a paid post the moment it goes live.
Are before-and-after photos allowed in telehealth creator posts?
Most categories ban them at the platform-policy layer before FDA review fires. Meta, TikTok, and YouTube all read a side-by-side photo as a results-promise. What works instead: directional language, time-window framing, no split-screen image.
Reading loop
Frequently asked
What words trigger an FDA warning letter on a telehealth post?
Four adjective patterns flip a creator post from testimonial into actionable claim. Cured. Cured-faster. Outperformed. Replaced. Each one moves the line from a personal story the creator owns to a medical claim the brand is legally responsible for. Safe substitutes the creator can say on camera: noticed, started, routine, convenient, for me.
Can a creator say a telehealth Rx 'cured' their condition?
No. A first-person 'I feel better since I started' line is a testimonial the creator owns. 'This drug cured my anxiety' is a claim the brand is responsible for under FDA rules. The brand owns every cure-language sentence on a paid post the moment it goes live, whether the script approved it or not.
Are before-and-after photos allowed in telehealth creator posts?
Most categories ban them at the platform-policy layer before FDA review even fires. Meta, TikTok, and YouTube ads all read a side-by-side photo as a results-promise. What works instead: directional language, time-window framing, no split-screen image. The 'I noticed after I started' frame survives. The week-0 vs week-12 photo does not.