Spaces

Patient and visitor privacy checklist for hospital chats

This checklist helps hospitals protect patients and visitors when an AI agent answers chat messages. It covers what may be said, what must be handed to staff and how to keep personal details safe.

Runs in your browser. Nothing you enter is sent anywhere or saved.

Before you start

A hospital holds the most private facts a person has: their health, their family, sometimes their worst day. Even small admissions can hurt. Saying that a person is on a ward may expose someone who is hiding from an abuser, or reveal a diagnosis that was meant to stay secret. A chat agent must be built around that fact.

Use this checklist with your privacy officer, not alone. Read each point, decide how it applies to your wards and your local law, and write the result into the agent's limits. Revisit it whenever you add a new service, a new site or a new way for people to write in.

Tick each one as you go. 0 of 10 done.

[ ] The agent does not confirm or deny that anyone is a patient.
[ ] The agent shares no condition, result, treatment or discharge detail.
[ ] Staff check identity and permission before any patient detail is shared.
[ ] Psychiatric and substance abuse enquiries all go to a person.
[ ] People are asked not to send full records or ID numbers in chat.
[ ] Only the minimum details are asked for, such as a first name.
[ ] Handover notes include only what staff need to continue.
[ ] Access to chat history is limited to named staff.
[ ] A clear route exists for people who want their details removed.
[ ] Privacy rules are reviewed after any near miss or change.

Wording that protects people

When someone asks about a named patient, a safe reply is warm and clear: I am sorry, I cannot share information about patients in this chat. The ward desk can help you, and I am passing your message to them now. It does not say the person is here and it does not say they are not.

Use the same careful reply for every name, every time. If the agent answers one family and not another, people will guess what the difference means, and the guess itself reveals something.

Questions to put to your privacy officer

These decisions are yours and your advisers' to make, not the agent's.

  • How long are chats kept, and who may read them?
  • Which wards need a special rule of silence?
  • What should staff do if a chat contains health details?

FAQ

Can the agent tell a visitor which ward a patient is on?

No. Ward and presence details are private. The agent should explain that only staff can help and pass the chat to the ward desk or patient information team to handle.

What if someone sends medical details in a chat?

The agent should pass the chat to a person and avoid repeating the details. Staff can then decide whether to move the conversation to a safer channel such as the clinic phone line.

Who should decide our privacy limits?

Your privacy officer, working with a senior clinician and legal advice where needed. Local law and hospital policy set the rules, and the agent only follows what you write.

More free tools for hospitals and hospital admin teams

Read the guide: An AI agent for hospital enquiries and administration.