# Patient Records and Identifiers

URL: https://docs.clinia.com/docs/patients/patient-records-and-identifiers

> How the Context Engine identifies patients.

The Context Engine stores everything about a patient under a patient ID that you choose: the records and sessions you submit, and the facts, narratives and learned memories derived from them. Each patient has its own patient graph, as described in [How the Engine Thinks About Data](https://docs.clinia.com/docs/get-started/how-the-engine-thinks-about-data). You need to [create a patient](https://docs.clinia.com/docs/patients/create-and-manage-patients#create-a-patient) before you submit anything for it.

## Choosing a patient ID [#choosing-a-patient-id]

Most integrations use a key that their source system already has for the patient, such as an MRN or a namespaced key like `hospital-a:MRN-4471` or `urn:oid:1.2.840.113619`.

Patient IDs are case sensitive, so `abc` and `ABC` are two different patients. They are also scoped to a workspace, so the same patient ID can safely be reused across workspaces. The characters a patient ID can contain are listed in [Identifiers](https://docs.clinia.com/api/identifiers#identifiers-you-assign).

## Identifiers inside a record [#identifiers-inside-a-record]

The identifier a record uses for its patient, such as `Patient.id` in a FHIR bundle, does not need to match the patient ID. The Context Engine reads the patient's name, birth date and gender from the record, and addresses the patient by its patient ID in REST paths, VFS paths and MCP tools.

## What deleting a patient removes [#what-deleting-a-patient-removes]

Deleting a patient removes the patient and all of its data, including its facts, narratives, clinical events, sessions, learned memories, search index and ingestion receipts. No data about the patient remains once the delete call returns. If a submission is still in progress, the delete cancels the rest of it.

A deleted patient ID can be created again and starts with no data.