# US 2020/0394333 A1 · De-identification of electronic records

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0394333 A1
- Document ID: US20200394333A1
- Status: Published application
- Kind: A1
- Application number: 16/897,153
- Filed: 2020-06-09
- Published: 2020-12-17
- Earliest claimed: 2019-06-14
- Assignees: The Regents of the University of California
- Applicants: The Regents of the University of California
- Inventors: Beau Norgeot; Atul Butte; Gundolf Schenk; Eugenia Rutenberg
- Web page: https://patentgrep.com/patent/US20200394333A1

## CPC

- G06F 21/6254: {by anonymising data, e.g. decorrelating personal data from the owner's identification}
- G06F 21/33: using certificates
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

A system is provided for de-identifying electronic records. The system may be configured to tokenize an electronic record to produce a plurality of tokens including a first token. The system may determine whether the first token is part of one of a plurality of expressions known to include protected health information. In response to determining that the first token is not part of any one of the plurality of expressions, the system may determine, based on a blacklist of tokens known to include protected health information, whether the first token includes protected health information. In response to determining that the first token includes protected health information, the system may generate a de-identified electronic record by replacing the first token with a second token obfuscating the protected health information. Related methods and computer program products are also provided.

## Claims

### Claim 1

A system, comprising:
at least one data processor; and
at least one memory storing instructions, which when executed by the at least one data processor, result in operations comprising:
tokenizing an electronic record to produce a plurality of tokens including a first token;
determining whether a protected health information is included in the electronic record by at least:
determining whether the first token is part of one of a first plurality of expressions, each of the first plurality of expressions known to include the protected health information, and
in response to determining that the first token is not part of any one of the first plurality of expressions, determining, based on a blacklist of tokens known to comprise the protected health information, whether the first token comprises the protected health information; and
in response to determining that the first token comprises the protected health information, generating a de-identified electronic record by at least replacing the first token with a second token obfuscating the protected health information.

### Claim 2 (depends on 1)

The system of claim 1, further comprising:
responding to an incorrect identification of the protected health information by at least updating the first plurality of expressions, the first plurality of expressions being updated by at least adding, to the first plurality of expressions, an expression including the first token and a third token adjacent to the first token in the electronic record.

### Claim 3 (depends on 1)

The system of claim 1, further comprising:
responding to an incorrect identification of the protected health information by at least updating the blacklist of tokens, the blacklist of tokens being updated by at least adding the first token to the black list of tokens or removing the first token from the blacklist of tokens.

### Claim 4 (depends on 1)

The system of claim 1, wherein determining whether the first token comprises the protected health information further comprises assigning a part-of-speech to the first token.

### Claim 5 (depends on 4)

The system of claim 4, further comprising:
responding to an incorrect identification of the protected health information by at least modifying the part-of-speech assigned to the first token, the part-of-speech assigned to the first token being modified by at least modifying a first part-of-speech tagging algorithm applied to assign the part-of-speech to the first token and/or changing the first part-of-speech tagging algorithm to a second part-of-speech tagging algorithm.

### Claim 6 (depends on 4)

The system of claim 4, further comprising:
in response to the first token being assigned a first part-of-speech, determining, based on the blacklist of tokens known to comprise the protected health information, whether the first token comprises the protected health information.

### Claim 7 (depends on 6)

The system of claim 6, further comprising:
responding to an incorrect identification of the protected health information by at least applying the blacklist of tokens in response to the first token being assigned a second part-of-speech instead of the first part-of-speech.

### Claim 8 (depends on 1)

The system of claim 1, further comprising:
in response to determining that the first token comprises neither the protected health information nor a non-protected health information, generating the de-identified electronic record by at least replacing the first token with the second token obfuscating the protected health information.

### Claim 9 (depends on 1)

The system of claim 1, further comprising:
determining whether the first token comprises a non-protected health information by at least determining whether the first token is part of one of a second plurality of expressions, each of the second plurality of expressions known to exclude the protected health information.

### Claim 10 (depends on 1)

The system of claim 1, further comprising:
determining whether the first token comprises the protected health information based at least on a notes map including one or more note-specific unsafe regular expressions, one or more note-specific blacklists, and/or one or more note-specific parts of speech.

### Claim 11

A computer-implemented method, comprising:
tokenizing an electronic record to produce a plurality of tokens including a first token;
determining whether a protected health information is included in the electronic record by at least
determining whether the first token is part of one of a first plurality of expressions, each of the first plurality of expressions known to include the protected health information, and
in response to determining that the first token is not part of any one of the first plurality of expressions, determining, based on a blacklist of tokens known to comprise the protected health information, whether the first token comprises the protected health information; and
in response to determining that the first token comprises the protected health information, generating a de-identified electronic record by at least replacing the first token with a second token obfuscating the protected health information.

### Claim 12 (depends on 11)

The method of claim 11, further comprising:
responding to an incorrect identification of the protected health information by at least updating the first plurality of expressions, the first plurality of expressions being updated by at least adding, to the first plurality of expressions, an expression including the first token and a third token adjacent to the first token in the electronic record.

### Claim 13 (depends on 11)

The method of claim 11, further comprising:
responding to an incorrect identification of the protected health information by at least updating the blacklist of tokens, the blacklist of tokens being updated by at least adding the first token to the black list of tokens or removing the first token from the blacklist of tokens.

### Claim 14 (depends on 11)

The method of claim 11, wherein determining whether the first token comprises the protected health information further comprises assigning a part-of-speech to the first token.

### Claim 15 (depends on 14)

The method of claim 14, further comprising:
responding to an incorrect identification of the protected health information by at least modifying the part-of-speech assigned to the first token, the part-of-speech assigned to the first token being modified by at least modifying a first part-of-speech tagging algorithm applied to assign the part-of-speech to the first token and/or changing the first part-of-speech tagging algorithm to a second part-of-speech tagging algorithm.

### Claim 16 (depends on 14)

The method of claim 14, further comprising:
in response to the first token being assigned a first part-of-speech, determining, based on the blacklist of tokens known to comprise the protected health information, whether the first token comprises the protected health information.

### Claim 17 (depends on 16)

The method of claim 16, further comprising:
responding to an incorrect identification of the protected health information by at least applying the blacklist of tokens in response to the first token being assigned a second part-of-speech instead of the first part-of-speech.

### Claim 18 (depends on 11)

The method of claim 11, further comprising:
in response to determining that the first token comprises neither the protected health information nor a non-protected health information, generating the de-identified electronic record by at least replacing the first token with the second token obfuscating the protected health information.

### Claim 19 (depends on 11)

The method of claim 11, further comprising:
determining whether the first token comprises a non-protected health information by at least determining whether the first token is part of one of a second plurality of expressions, each of the second plurality of expressions known to exclude the protected health information.

### Claim 20 (depends on 11)

The method of claim 11, further comprising determining whether the first token comprises the protected health information based at least on a notes map including one or more note-specific unsafe regular expressions, one or more note-specific blacklists, and/or one or more note-specific parts of speech.
