Accessibility
Accessibility statement.
Last reviewed: September 16, 2026. Reviewed at least every six months, and again whenever the design system changes.
Conformance target
DebriefMD targets WCAG 2.2 Level AA (Web Content Accessibility Guidelines, published by the W3C). This is the standard every new feature and page is built to, and the one we measure against when we review.
We target 2.2 rather than 2.1 because 2.2 is the current recommendation and its additions are the ones that matter to a resident on a phone in a hospital corridor: target sizes, focus that is not hidden behind something else, and signing in without having to remember or transcribe anything.
Not independently audited yet: we have not completed a formal third-party accessibility audit. What follows describes the engineering practices we hold ourselves to, not a certified conformance claim.
What we have done
Semantic HTML
Pages use landmark elements (nav, main, section, footer), a heading hierarchy, and ARIA attributes only where native semantics fall short. Where we cannot honour an ARIA pattern in full, we use the plain element instead of claiming the pattern.
Keyboard navigation
Every interactive element is reachable and operable with a keyboard, and a visible focus ring follows you. Menus and dialogs close on Escape. The focus ring is a dark gold measuring 5.5:1 against the page background. WCAG asks 3:1 of a visual indicator.
Colour contrast
Body text measures 13.8:1 against the page background and 5.5:1 for secondary text; the green accent used for text and links measures 6.1:1, and the gold used for warnings and focus rings measures 5.5:1. WCAG AA asks 4.5:1 for normal text. Gold fills otherwise carry dark ink and are never used as thin text on a light background, where the fill tone measures about 2.1:1.
Screen-reader labels
Interactive elements without visible text carry an accessible name, and purely decorative marks are hidden from assistive technology.
What WCAG 2.2 adds, and where we stand
Two of the new criteria we have checked and can state plainly. A third, that help sits in a consistent place, we have checked and do not meet; it is in the gaps below rather than quietly omitted.
Accessible authentication
Signing in is a link sent to your email. There is no password to remember, no puzzle, and nothing to transcribe, which is what this criterion asks for.
Target size
Standard controls are 44px tall, and the compact ones used inside dense cards are 36px. This criterion asks for 24px.
Known gaps
We have not been audited by a third-party accessibility specialist. Nothing here is a certified conformance claim.
The recording screen has not been tested with screen readers on mobile, and that is the screen this product exists for. It is the first thing an audit should look at.
Our automated checks cover structure and contrast, not the WCAG criteria that need a human to judge, such as whether an error message actually explains how to fix the problem.
We have not verified the product against a screen magnifier at 400% zoom, or with Windows High Contrast Mode.
Help is not in a fixed place once you are signed in. It appears next to whatever has gone wrong, which is useful when something fails and no help at all when you are simply stuck. WCAG 2.2 asks for a consistent location, and we do not yet meet it.
Most of the interface is in English. You can record and have a debrief transcribed in French, and the consent disclosure and the wording around a French note follow the language you chose, but the rest of the screens (buttons, errors, help) are not translated.
Ongoing work
Each new feature is built with keyboard operation and screen-reader compatibility from the start, and we would rather use a plain element than claim an ARIA pattern we have not fully implemented. A control that announces itself as something it is not is worse than one that announces nothing.
This statement is owned by the founder, who is accountable for keeping it true, and is reviewed at least every six months (next by March 2027) and again whenever the design system changes, since a recolour can invalidate every contrast figure on this page at once. The measurements above were taken on September 16, 2026 against the current palette. A third-party audit is planned as the product matures; the recording screen on mobile is where we would point it first.
Report an accessibility issue
If you hit a barrier using DebriefMD, tell us. A person reads these. Describe what you were trying to do and what got in the way, and you will get a reply. We would rather hear about it than have you work around it silently.
hello@debriefmd.ca