Hello NHS e-RS API team,
We are testing A039 – Upload file to document store (FHIR R4) in the integration environment and have found that accented characters in NHSD-eRS-File-Name are not preserved.
POST https://int.api.service.nhs.uk/referrals/FHIR/R4/Binary
We reproduced this independently of our application using backend CURL. For the UTF-8 test, the outgoing hexadecimal trace confirms that é was transmitted correctly as bytes C3 A9.
Our findings are:
| Filename header supplied | Returned filename |
|---|---|
filenamé-test.txt, using UTF-8 bytes C3 A9 |
filenam??-test.txt |
filenamé-test.txt, using Latin-1 byte E9 |
filenam?-test.txt |
filenam%C3%A9-test.txt, using percent-encoded UTF-8 |
Literal filenam%C3%A9-test.txt |
For example, the UTF-8 request returned:
Content-Disposition: attachment; filename="=?UTF-8?Q?filenam=3F=3F-test-3.txt?="; filename*=UTF-8''filenam%3F%3F-test-3.txt
The replacement characters are already present in the A039 response, before the storage PUT. We have also observed corrupted filenames in the e-RS Professional Web App after uploading and associating attachments.
Could you please confirm the supported encoding for non-ASCII filenames in NHSD-eRS-File-Name?
Request details for tracing:
Environment: Integration
Request timestamp, including timezone: Tue, 03 Nov 2026 09:01:34 GMT
X-Correlation-ID: 7f87b8f5-7d5f-4f79-9c32-616f64361a58
Returned X-Request-ID: 61dc20e7-91d6-44d4-b59e-e728d2e68c74-1
Application ID: 6a0392f2-ba38-413f-b83c-449ec1168b5a
Many thanks.
Jose