What the Signature Certifies: the consultation paper is published as record twenty.

NHS England’s public consultation on the revision of DCB0129 and DCB0160 — the first since 2018 — closed on 11 September. A response was submitted that day (reference ANON-4Z5E-5X5H-J). Consultation responses are anonymised, so the argument is published here as What the Signature Certifies, the estate’s twentieth Zenodo record, with the requirement-level drafting reproduced in full in Appendix A so that it can be cited.

In a typical NHS trust, three of every four digital tools influencing patient care have no documented safety assurance. A mandatory standard nobody verifies is an optional one.

The review’s own focus groups had already named three gaps: AI governance, post-implementation monitoring and compliance. The paper puts a mechanism behind each. It adds a fourth ask the focus groups did not reach, and the one the estate exists to make: the signature should certify the environment receiving the system, not only the system. That is the Cultural Safety Case written into the standard itself, at DCB0160 requirement 3.4.2.

Two findings in the paper come from live deployment work rather than the literature, and both are named with the written consent of the people who contributed them. The first is a structural limit of deviation-from-baseline monitoring: a person already at high risk at first contact has no prior state to deviate from, so a system that watches only change-from-self cannot flag them. The second is that a signature is a control only when the second person independently verifies and their registration is genuinely exposed if they did not — the bedside evidence on primed double-checking, read at organisational scale.

Sources