This post is for marketing leads and web managers at hospitals and clinics whose site still carries FAQPage or HowTo markup added years ago, often by a plugin, on the promise of expandable answers in Google results. Those results are gone. The question now is what to do with the markup, and with the FAQ blocks it describes.
It answers three things: what changed and when, why "gone" does not automatically mean "delete everything", and how to audit a sample of pages and decide, page by page, whether to keep, rewrite or remove. We describe the audit method. We have not published findings from any hospital site here, and nothing below should be read as a result. Where an example appears, it is invented and says so.
What actually changed
The dates are worth getting right, because they are often misquoted.
- HowTo. In August 2023 Google announced it was changing how FAQ and HowTo rich results appear. How-To results were first limited to desktop and then dropped. Google's Search Central changelog records the documentation removal on 14 September 2023, saying the result is no longer shown on desktop or mobile.
- FAQ, 2023. The same announcement said FAQ rich results would be shown only for well-known, authoritative government and health websites. Hospitals were, in principle, in the group still eligible.
- FAQ, 2026. The changelog records a deprecation notice on 8 May 2026 saying the feature would no longer appear in Google Search from 7 May 2026, and the documentation was removed on 15 June 2026.
So the FAQ result that health sites alone kept after 2023 has now gone too. No site, whatever its sector, gets the expandable question-and-answer result from FAQPage markup.
"Gone" does not mean "harmful"
Google's 2023 announcement made a point that still holds: you can drop the structured data, but there is no need to proactively remove it, because unused structured data does not cause problems for Search and has no visible effect.
That should shape the audit. The reason to touch legacy markup is not penalty risk. It is one of these.
- Accuracy. Markup that no longer matches the visible page is a liability. Google's structured data guidelines say not to mark up content that is not visible to readers.
- Maintenance weight. A template that emits twenty identical questions on every page is code and data someone must keep correct.
- Clinical risk. Every answer in a health FAQ is a clinical statement. A block that nobody has reviewed since 2021 is an editorial risk whether or not it carries markup.
- Focus. Time spent on dead markup is time not spent on content readers use.
The FAQ block that readers see is a different asset from the markup. A good set of questions with direct answers helps a patient who is deciding whether to book, and nothing in Google's change touches that. Our FAQ and People Also Ask optimisation service is about the content, and it is why the question is "what should the block say?", not "is the schema still valid?".
One caution on the other side. Do not assume FAQ markup now helps AI answer tools. We have not seen documentation from any major search or AI provider that says it does, so treat that as unproven and do not use it as a reason to keep boilerplate.
The audit method
The aim is a defensible decision for each template and each page, using a sample.
Step 1: Sample
For a hospital group, take a stratified sample of 100 pages: specialty hubs, treatment pages, condition pages, doctor pages and blog posts, in proportion to how many of each exist, with extra weight on pages that earn clicks. If you are studying several sites, sample the same page types from each so results are comparable. A sample of 100 does not give you statistics about the whole site; it gives you enough to see which templates are the problem.
Step 2: Extract
Crawl the sample and extract, for each page:
- every JSON-LD, Microdata or RDFa node of type
FAQPage,Question,AnswerandHowTo; - the number of questions in the markup and the number of questions visible in the rendered page;
- the answer text in the markup and the answer text on the page;
- whether the FAQ is in the raw HTML, or only after JavaScript runs;
- the template name and the CMS or plugin that emits it.
Step 3: Classify each page
Apply the rules in order.
| Observation | Classification | Action |
|---|---|---|
| Markup contains questions not visible on the page | Mismatch | Remove the markup now, or add the content if it is valuable |
| Same question set appears on many pages | Boilerplate | Remove from pages where it does not apply |
| Answers are promotional or make unreviewed clinical claims | Editorial risk | Rewrite with clinical review, or delete |
| Answers are short, specific, accurate and visible | Sound | Keep the content; keeping the markup is optional |
HowTo markup on a page that is not step-by-step instructions | Misuse | Remove |
HowTo markup on genuine instructions, such as a preparation guide | Dead feature | Remove or keep, but the result will not appear |
| FAQ only appears after a click that loads from a script | Fragile | Move into the page body if the answers matter |
Step 4: Record, count and report
For each template, report the number of pages sampled, how many fall into each class, and the proposed action. Keep the counts as plain numbers and say they are for the sample. A report that says "template X emits twelve generic questions on every service page" is far more useful to a developer than a percentage.




