After the letter: 32 of 87 warned telehealth sites still carry the kind of wording FDA quoted.

By letter batch

The other 26 companies could not be read: the site was gone, it blocked automated reading, or it drew its text in a way a plain reader cannot see. They are not counted either way.

No company is named in this index. FDA publishes every letter in its warning-letter index, where the letters in this series carry “(Telehealth)” in the subject.

What FDA quoted

The 113 letters quote 255 lines of marketing. By our check’s wording patterns, 150 of the 255 tie a compounded drug to a brand-name one: “the same active ingredient as” a named brand, a “generic” version of a brand, “similar to” a brand. Others give a pharmacy an FDA status (“FDA-approved pharmacy”) or borrow clinical proof from the approved product (“clinically proven”).

FDA’s position in the letters: compounded drugs are not FDA-approved, and wording that presents them as the approved product, or as evaluated by FDA, is false or misleading under sections 502(a) and 502(bb) of the Federal Food, Drug, and Cosmetic Act.

How we counted

• On 7 October 2026 we fetched each company’s home page and up to 14 pages it links to, treatment and pricing pages first.
• Software split each page into sentences and checked every sentence against the kinds of wording in FDA’s 255 quoted lines. No AI model made the call. The check is a fixed set of wording patterns taken from the letters.
• A site counts if it had at least one strong match: a sentence that ties a compounded drug to a named brand, calls it a generic or a version of a brand, or gives a pharmacy or facility an FDA status.
• Not counted: sentences that say the opposite (“not the same as”, “not FDA-approved”), safety and label text, questions, and “generic” used for drugs that do have approved generics.
• Tested on FDA’s own 255 lines, the check recognises 88% of them, and 64% as a strong match.

What this does and does not show

• It shows wording, not law. A match means a sentence of the kind FDA has quoted to other firms. It is not a finding that a company’s marketing breaks the law, and only FDA can say what it acts on.
• It is a floor. Text inside images, video, ads, emails and logged-in pages was not read. Companies that reworded a line well enough to leave the patterns are not counted.
• It is one day. Sites change. The count is for 7 October 2026.
• A person read a sample of the matches, not all of them. In the samples read, roughly seven to nine in ten held up, depending on the site.
• A site that still carries the wording may have answered its letter in full. The letters themselves say the lines they quote are not an exhaustive list.

Why it matters

FDA said on 9 September 2025 that it “is already implementing AI and other tech-enabled tools to proactively surveil and review drug ads.” It posted these telehealth letters in three batches, between 9 September 2025 and 8 June 2026. The wording it quotes is short, common and easy to find by machine, which is how we found it.

Sources

• FDA warning-letter index, letters with “(Telehealth)” in the subject, 9 September 2025 to 8 June 2026.
• FDA news release, 9 September 2025.
• Our reading of 7 October 2026. The list of sites names the companies, so it is not published.

Cite as “Claims Verified, The telehealth letter index, issue 1, read 7 October 2026”. Free to reproduce with attribution. Questions about the method: audit@claimsverified.org.