Hospital Website Accessibility Statistics: 2026
We ran automated WCAG 2.1 checks on the homepages of 60 of the largest US hospitals and health systems in November 2026, using the axe-core engine. The scan found 250 detectable errors across those homepages, with counts ranging from none to 39 on a single site.
Key Findings
|
Federal rules now put a date on this work. The US Department of Health and Human Services requires recipients of its funding with 15 or more employees to meet WCAG 2.1 Level AA, and in May 2026 it moved that compliance date from May 11, 2026, to May 11, 2027. Recipients with fewer than 15 employees have until May 10, 2028. A hospital homepage is the front door for appointments, billing, and locations, so it is a fair place to begin.
Our analysts took the largest hospitals and health systems from the 2026 Becker's ranking, opened each homepage in a desktop browser, and ran the WCAG 2.0 and 2.1 Level A and AA rules in axe-core. We counted an error only when axe-core marked a definite violation, and we excluded ten homepages that blocked automated browsers.
In the following sections, we compare how many errors each homepage showed, which error types came up most often, how health systems and individual hospitals differ, and how these results line up with the WebAIM Million scan of the wider web.
Nearly Half of Large Hospital Homepages Show No Automated Accessibility Errors
A count of detectable errors is the simplest way to see the spread across these sites, and that spread turned out to be wide.
Hospital Homepages by Number of Automated WCAG Errors (November 2026)
Errors on homepage | Homepages | Share |
0 | 29 | 48.3% |
1 to 2 | 8 | 13.3% |
3 to 9 | 13 | 21.7% |
10 to 24 | 9 | 15.0% |
25 or more | 1 | 1.7% |
Key Findings
Twenty-nine homepages showed no error that automated rules can detect. That does not make those sites fully accessible, since automated tools cover only part of WCAG.
The ten homepages with the most errors held 68.4% of all the errors we found. The problem is concentrated, and the typical homepage carried only a handful.
For a patient using a screen reader, one unlabeled link in the main menu can block the path to a page, so a low count can still cause harm. With the HHS date set for May 2027, homepages that still show errors have a clear deadline to work against.
Low Contrast and Unnamed Links Lead Hospital Homepage Errors
The types of errors say more than the totals, because each one affects a different group of patients in a different way.
Most Common Automated Errors on Hospital Homepages (November 2026)
Error | Homepages | Share of homepages | Errors found | Share of errors |
Low color contrast | 16 | 26.7% | 112 | 44.8% |
Links with no name | 10 | 16.7% | 27 | 10.8% |
ARIA role missing child items | 6 | 10.0% | 13 | 5.2% |
Hidden content that can take focus | 4 | 6.7% | 16 | 6.4% |
ARIA item missing a parent role | 4 | 6.7% | 13 | 5.2% |
List items outside a list | 4 | 6.7% | 13 | 5.2% |
Buttons with no name | 3 | 5.0% | 3 | 1.2% |
Images with no alt text | 3 | 5.0% | 5 | 2.0% |
Key Findings
Low color contrast appeared on more homepages than any other error and made up close to half of all errors found. Patients with low vision have the most to lose when text blends into a background.
Links with no name came second. A screen reader announces such a link with no description, which leaves a patient to guess what it opens.
Every error in the scan was rated serious or critical by axe-core, at 75.2% and 24.8% of the total. Twenty homepages had at least one critical error.
Health Systems Showed Errors Somewhat More Often Than Individual Hospitals
Hospitals and multi-hospital systems often run their websites differently, so we split the homepages by the two lists in the Becker's ranking.
Automated Errors by Organization Type (November 2026)
Organization type | Homepages | With an error | Share | Median errors | Average errors |
Health systems | 29 | 17 | 58.6% | 2 | 3.6 |
Large hospitals | 31 | 14 | 45.2% | 0 | 4.7 |
All homepages | 60 | 31 | 51.7% | 1 | 4.2 |
Key Findings
Health system homepages were more likely to show at least one error than individual hospitals. These groups are small, so the gap is a directional signal and not a firm ranking.
Individual hospitals had a lower median but a higher average. A few very heavy homepages pulled the hospital average up, which is why the median tells a different story.
The spread within each group was far wider than the difference between the groups. A patient comparing two nearby hospitals may find a real difference even when the national figures look similar.
Hospital Homepages Show Fewer Detected Errors Than the Wider Web
The WebAIM Million offers the most widely used yardstick for home page accessibility, with the caution that it relies on a different testing tool.
Automated Errors on Hospital Homepages and Across the Web (2026)
Measure | Large hospital homepages | WebAIM Million |
Home pages with detected errors | 51.7% | 95.9% |
Average errors per home page | 4.2 | 56.1 |
Low contrast text | 26.7% | 83.9% |
Empty or unnamed links | 16.7% | 46.3% |
Missing image alt text | 5.0% | 53.1% |
Empty or unnamed buttons | 5.0% | 30.6% |
Missing form labels | 3.3% | 51% |
Missing document language | 0.0% | 13.5% |
Key Findings
Hospital homepages came in below the wider web on every error type we compared. WAVE and axe-core apply different rules, so the direction is more reliable than the size of each gap.
Low contrast led both lists, which suggests the same design habits show up on hospital sites and everywhere else.
None of the 60 homepages was missing a document language, and only two were missing form labels. Human review is where the next round of work sits, since axe-core could not settle color contrast on nearly every homepage.
For patients who rely on screen readers, magnification, or keyboard navigation, these results describe a mixed starting point. Many large hospital homepages clear the automated checks, but a clean automated result covers only part of what a person needs from a page.
For hospital web teams, the HHS date gives the work a fixed deadline, and color contrast offers the most errors to fix in a single category.
The broader trend is a hospital web that has handled the most easily detected problems and still faces a long list of checks that only people can make. So if you run a hospital website, start with contrast and unnamed links on your homepage, then have a person review what automated tools cannot settle.
Methodology
We selected homepages from the health system and hospital lists in the 2026 Becker's ranking and assigned each to its list as a health system or large hospital. Each homepage was opened in a desktop Chromium browser on November 8, 2026, given six seconds to render, and scanned with axe-core 4.10 using the WCAG 2.0 and 2.1 Level A and AA rule sets.
Of 70 homepages attempted, 10 could not be scanned because the sites returned access-denied or bot-check pages or too little content, leaving 60 homepages. Those 10 are not counted as passing or failing.
An error is one element flagged as a definite axe-core violation. Items axe-core marked as incomplete, including color contrast it could not settle, are reported separately as needing human review. The study covers homepages only, at desktop width, on one day, and does not test keyboard use, screen readers, or any interior pages.
The WebAIM Million uses the WAVE engine on the top one million home pages and was collected in February 2026, so its figures are shown for context and are not directly comparable. Derived figures, including shares, medians, and averages, were calculated from these data.
Comments