Why patient education videos are harder to caption than they look
A caption error in a board meeting video is embarrassing. A caption error in a video on how to use an inhaler, prepare for a colonoscopy or start a new medication can change what a patient does at home. A patient who relies on captions may follow that video step by step, pausing and replaying it.
These videos pack the words speech recognition handles worst into a few minutes: drug names, doses, units, anatomy terms and more than one speaker. Auto-captions and quick edits tend to slip on exactly these words. Here are the kinds of errors to look for, with made-up examples:
- Sound-alike numbers. For example, “fifteen milligrams” and “fifty milligrams” sound close. A caption that reads 50 mg when the pharmacist said 15 mg is a wrong dose.
- Drug names. Generic names are long and unfamiliar to speech recognition, and brand names can come out as ordinary words. Sound-alike pairs, such as hydroxyzine and hydralazine, can be swapped without anyone noticing.
- Units. Milligrams, micrograms, milliliters and units are not interchangeable. A caption that drops or changes the unit changes the instruction.
- Frequency and timing. “Twice a day” heard as “two a day,” or “before meals” lost in background music, changes how a patient takes a drug.
- Anatomy and procedure terms. Words like “bronchodilator” or “sigmoidoscopy” often get split into common words that make no sense.
- Several speakers. A nurse, a pharmacist and a patient may all talk. Without labels, a viewer can’t tell who gave the instruction and who asked the question.
If your channel relies on YouTube’s automatic captions today, read whether YouTube auto-captions are enough and our post on caption quality standards.
Applying the DCMP Captioning Key to medical content
The Captioning Key from the Described and Captioned Media Program (DCMP) is a free captioning style guide. Four of its conventions matter most for health content.
Check technical terms against an authoritative source
The Captioning Key’s Language Mechanics section says a general dictionary is not enough: proper names, technical terms and specialized language must be verified through specialty references or an authoritative source. Drug names belong in that group. Check each one against your formulary, the product labeling, or the clinician who appears in the video, and spell it the way the label does.
Write numbers, doses and units on purpose
The Captioning Key’s Numbers section sets a default: spell out one to ten and use numerals above ten, unless one of its listed exceptions applies. Read those exceptions before you caption doses, times and measurements, then pick one style and use it across your whole library. Add a rule of your own: every dose and unit gets a second look from someone clinical, and the unit stays next to the number, as the speaker said it.
Identify the speaker
The Captioning Key prefers to identify an on-screen speaker by placing the caption under them. When placement can’t do that and the speaker’s name is known, the name goes in parentheses. YouTube Help’s sample caption file uses another common style: “>>” and the speaker’s name in capitals. Either works if you use it the same way every time. When a video doesn’t give names, a role such as (Pharmacist) tells the viewer whose instruction it is.
Caption the sounds that carry instructions
In a device video, some sounds are the instruction: the click of an inhaler, the click of a pen injector, a timer beep. The same YouTube sample file shows a sound written in brackets, [intro music]. Use that form for the sounds a patient needs to hear, such as [inhaler clicks], and skip background music that tells them nothing.
| What happens in the video | Weak caption | Better caption |
|---|---|---|
| Pharmacist says “fifteen milligrams once a day” | 50 mg once a day | 15 mg once a day, confirmed by a clinician |
| Generic drug name said quickly | An ordinary word that sounds similar | The name spelled as it appears on the label |
| Off-screen nurse says “Shake it first” | Shake it first | (Nurse) Shake it first. |
| Inhaler clicks, no speech | Nothing | [inhaler clicks] |
| Patient asks a question, pharmacist answers | One unlabeled block of text | Each line labeled with the speaker |
Demonstrations that likely need audio description
Patient education videos often show instead of tell. A nurse presses a pen injector against the thigh and holds it while music plays. The camera zooms in on a prescription label. A prep calendar appears on screen and nobody reads it. Captions cover the soundtrack. They do nothing for a viewer who can’t see the step.
WCAG 1.2.5, a Level AA criterion, asks for audio description of recorded video when the picture shows something the soundtrack doesn’t. Our guide on when audio description is required goes through the test. Signs a patient video likely needs it:
- A step is shown but not spoken, such as “now do this” with a gesture.
- A dose, phone number, date or prep timeline appears only as on-screen text.
- A device indicator matters: a dose counter, the window on a pen, a color change.
- A demonstration runs under music with no narration.
The cheapest fix is in the script: have the presenter say what they are doing as they do it. For videos you already posted, you add a described track. YouTube Help says you can add an audio description track in YouTube Studio once the original audio in the same language is uploaded, and the feature is currently available only to creators with access to Advanced features. Our guide on adding audio description to YouTube covers the steps.
What Section 504 asks of health centers, and when
HHS’s Section 504 rule covers organizations that receive HHS funding, including federally funded health centers, and adopts WCAG 2.1 AA as the standard for web content. After the 2026 extension (HHS, 91 FR 25496), the deadline is May 11, 2027 if you have 15 or more employees and May 10, 2028 if you have fewer. The earlier dates in the 2024 rule no longer apply. Check your deadline, and see our guide to Section 504 video requirements.
For recorded patient education video, that means accurate captions (WCAG 1.2.2) and audio description where the picture shows something the soundtrack doesn’t (WCAG 1.2.5). This post explains how the rules read; it isn’t legal advice.
Older videos
The archived content exception applies only when all four conditions hold: the video was made before the deadline, it is kept only for reference, it hasn’t changed since, and it sits in an area clearly labeled as an archive. A video is not archived just because it is old. Social media posts made before the deadline are exempt, but HHS didn’t say whether YouTube counts, so pre-deadline YouTube uploads are probably exempt (unclear). Patient videos embedded on current pages of your website are in scope. And any video a patient asks for has to be made accessible either way. Our guide to the archived content exception has the details.
What is not required
Only Level A and AA criteria are required. Sign language interpretation (1.2.6), extended audio description (1.2.7) and a full media alternative (1.2.8) are Level AAA. You can still offer them where they help your patients, but the rule doesn’t ask for them.
A small-team workflow
- 1
Find the videos patients are sent to
List the videos your clinicians link in after-visit summaries, patient portal messages and discharge instructions, plus those embedded on your website. Run the free website video check and scan your channel to see what you have.
- 2
Build a term glossary
Pull the drug names, devices and procedures from your formulary and your most-used videos. Spell each one the way the label or an authoritative reference does, and give the list to whoever edits captions.
- 3
Fix the top videos first
Start with the videos patients see most and the ones about medication and procedure prep. Edit or replace the captions, using our guide on adding captions to YouTube videos. YouTube Help lists Scenarist (.scc) files as its preferred format if you upload a caption file.
- 4
Have a clinician review names and doses
Send the caption file or transcript to a nurse or pharmacist with every drug name, number and unit highlighted. Ask them to confirm each one, not to proofread the whole thing.
- 5
Mark demonstrations for audio description
Watch each video with the screen covered. Note every place where you can’t follow the step from sound alone.
- 6
Set up a request page
Give patients one place to ask for an accessible version of any video. Start from our accessibility request form template and decide who answers and how fast.
- 7
Script new videos to say what they show
Ask presenters to read on-screen numbers aloud and narrate each step. It shrinks both the caption review and the audio description work.
Where to start this week
- Check your deadline so you know whether you are working toward May 11, 2027 or May 10, 2028.
- Scan your channel to sort your patient videos and see which have real captions.
- Run the website video check on the pages patients visit most.
- Run one device or medication video through the audio description check.
Sources
- Described and Captioned Media Program (DCMP) Captioning Key
- Described and Captioned Media Program (DCMP) Captioning Key
- Described and Captioned Media Program (DCMP) Captioning Key
- YouTube Help (Google)
- YouTube Help (Google)
This is a reading of the rules, not legal advice. Something out of date? Tell us and we'll fix it.