Protecting Healthcare Documents

Healthcare documents can contain multiple high-risk identifiers in the same record, including names, addresses, dates of birth, and Czech rodné číslo values. A safe default is to scan every document on ingestion and store only a redacted version outside the clinical system.

Detect common high-risk fields

from fastpii import PrivacyGuard

guard = PrivacyGuard(regions=["cz"])

medical_record = """
Patient: Jan Novak
Rodné číslo: 850101/1234
Date of birth: 1985-01-01
Address: Vodičkova 10, Praha
Diagnosis notes: follow-up required
"""

result = guard.detect(medical_record)
redacted_record = guard.redact(medical_record)

print(result.findings)

This catches the categories that usually matter most for privacy reviews: rodne_cislo, name, address, and date of birth.

Redact for compliance workflows

from fastpii import PrivacyGuard

guard = PrivacyGuard(regions=["cz"])


def prepare_record_for_external_review(text: str) -> str:
    return guard.redact(text)

Use this before sharing records with external legal counsel, processors, AI tools, or support teams that do not need direct identifiers.

Flag Article 9 sensitive context

Rodné číslo can reveal biological sex, which can trigger GDPR Article 9 concerns in some workflows.

from fastpii import PrivacyGuard

guard = PrivacyGuard(regions=["cz"])
result = guard.detect("Rodné číslo pacienta: 850101/1234")

article_9_related = []
for finding in result.findings:
    metadata = finding.metadata
    if metadata and metadata.get("article_9"):
        article_9_related.append(finding)

Use metadata-driven checks like this to send records into a stricter review path.

Practical healthcare workflow

1. Ingest the original record in the source clinical system.
2. Run detect() immediately.
3. Route records with Article 9-related metadata to compliance review.
4. Use redact() before indexing, summarization, or external sharing.
5. Retain raw data only where strictly required.