enforcing-nophi-logging

To prevent PHI leaking into logs and OpenTelemetry traces, apply a pre-emit redaction filter that scrubs structured identifiers and free-text PHI before records are sent to handlers. This fail-closed approach ensures sensitive health data never reaches centralized observability stores.

Fail-closed PHI redaction is a logging filter mechanism that blocks record emission entirely if sanitization fails, replacing clinical text with safe metadata like document identifiers and entity counts to ensure no protected health information leaks into telemetry payloads.

Redact patient names and phone numbers from Python logging handlers by adding a structured no-PHI filter that replaces clinical text with safe metadata. The filter uses fast identifier patterns to scrub sensitive data before logs are emitted to any handler.

Yes, no-PHI logging works with Sentry and crash-reporting systems by sanitizing exception messages and error-tracking payloads before emission. It provides specific guidance for scrubbing OpenTelemetry spans, events, and crash reports across OpenMed deployments processing clinical text.

Identifier pattern-based PHI redaction requires pairing with OpenMed PHI extraction testing to catch free-text health data that simple patterns miss. Limitations include balancing fast identifier matching against comprehensive clinical text scrubbing across all telemetry sinks.

Yes, testing no-PHI logging filters requires OpenMed PHI extraction alongside fast identifier patterns to validate that structured identifiers, free-text PHI, and clinical text are fully scrubbed before logs reach observability systems and centralized stores.