We conducted an automated web accessibility audit of the homepages of public hospitals with 200 or more licensed beds that operate a dedicated hospital website. Of the 367 hospitals meeting these criteria, the survey used the automated web accessibility audit tool "Lente Scan."
As a result, accessibility issues detectable through automated testing were found on the homepages of all 354 hospitals surveyed. The median number of issues per page was 30, and the hospital with the most had 642.
Key Findings
- Issues were detected at all 354 hospitals surveyed. No hospital was free of issues in the automated audit. Issues were found across every type of hospital operator, including prefectural, municipal, and local independent administrative institution hospitals. The median number of issue types per page was 6.
- At 90% of hospitals, issues were detected where UI roles and information structure are not correctly conveyed to assistive technologies. Success Criterion 4.1.2 (Name, Role, Value) addresses whether the names, roles, and states of UI elements such as buttons and links are properly communicated to assistive technologies, including screen readers. Issues were detected at 90.7% of hospitals. Success Criterion 1.3.1 (Info and Relationships), which covers structures such as headings and lists, was flagged at 80.5% of hospitals. The most common finding under this criterion was the absence of a main landmark, which marks the primary content area of a page (161 hospitals, 45.5%). Note that these two success criteria are linked to a large number of detection items, which tends to result in higher detection rates.
- The most frequent individual issue was insufficient contrast between text and background. It was detected at 258 hospitals (72.9%), totaling 4,696 instances. Typical examples include light-colored brand text and white text placed over photographs. Note that where backgrounds are images or gradients, the automated audit cannot determine the actual background color, so the 72.9% figure is limited to cases the automated audit could evaluate.
Limitations of This Survey
Automated auditing can evaluate roughly 30 to 40% of WCAG success criteria. The figures in this survey are limited to that scope and do not assess the overall accessibility of individual hospital websites. Only aggregate results are published; audit results for individual hospitals are not disclosed. The remaining areas and items that cannot be determined through automated testing alone require manual review.
Survey Methodology
- Scope: Public hospitals listed in the member facility database of the National Association of Municipal Hospitals that have 200 or more licensed beds and a dedicated hospital website. Homepages of 367 hospitals were reviewed (homepage only, one page per site). Thirteen hospitals could not be audited because their websites were under maintenance at the time of the survey, leaving 354 hospitals in the analysis.
- Survey date: November 4, 2026
- Evaluation criteria: WCAG 2.2 Level AA
- Audit engines: axe-core 4.11.4, Lente Detect 1.6.7 (tested at 1280px, equivalent to desktop, and 320px, equivalent to smartphone)
Survey Report
The following report details the survey results, including totals by success criterion and by issue type, as well as the survey conditions.
https://www.liberogic.jp/survey/hospital/2026-11-04/report-public.html
The Lente Scan tool used in this survey automatically checks web accessibility simply by entering a URL. It is free and requires no registration, and lets you check your own website using the same testing engines and evaluation criteria as this survey. (The free version tests only the desktop-equivalent screen width; testing that includes the smartphone-equivalent screen width is available on the Pro plan.)
https://lente-scan.liberogic.jp/
Areas that require human review can be assessed through expert manual testing.
For areas that automated testing cannot evaluate, we provide manual web accessibility testing services led by a team centered on IAAP Certified Professionals in Web Accessibility (WAS). Testing is conducted in Lente Audit, our proprietary dedicated workspace, which handles issue sharing, correspondence on fixes, and report delivery in one place.
