mirror of
https://github.com/MBombeck/HealthLog.git
synced 2026-11-12 14:16:12 +00:00
Marc-André Bombeck
e3342dc19e
The inbound-document store encrypts the raw upload (`inbound_documents.content_encrypted`) because it is sensitive clinical data, but the per-fact staging columns persisted the same class of data in the clear: `extracted_facts.data_json` held the transcribed clinical values (diagnosis text, lab values, medication names, stated codes) and `provenance_json` held the verbatim source span (up to 2000 chars copied straight out of the report). The at-rest guarantee was defeated for staged facts, which persist through PENDING and are retained after the decision. Convert both columns to the AES-256-GCM `*_encrypted` BYTEA convention shared with the note columns, going through the same Bytes codec. The upload and fact-edit paths encrypt on write; the confirm path decrypts at commit time to write into the normal structured stores, so the review-then-confirm flow is unchanged. Reads are fail-closed (a bad key id throws). Register both columns in the encrypted-columns registry and the key-rotation script so a dropped legacy key cannot strand them (the guard test enforces both). Migration 0220 swaps the plaintext JSONB columns for the encrypted BYTEA columns; pre-confirm staged rows are ephemeral and cannot be re-encrypted in SQL, so any left by a pre-release build are dropped rather than retained as plaintext (their parent document keeps its encrypted original).
43 lines
2.5 KiB
SQL
43 lines
2.5 KiB
SQL
-- v1.25 (W-DOCS-IN, security) — encrypt the staged extracted-fact payloads.
|
|
--
|
|
-- The inbound-document store already encrypts the raw upload
|
|
-- (`inbound_documents.content_encrypted`) precisely because it is sensitive
|
|
-- PHI. The per-fact staging columns `extracted_facts.data_json` /
|
|
-- `provenance_json`, however, persisted the SAME class of data in the clear:
|
|
-- the FHIR-staged clinical values (diagnosis text, lab values, medication
|
|
-- names, stated codes) and the verbatim source span (up to 2000 chars copied
|
|
-- straight out of the discharge letter / lab report). That defeated the
|
|
-- at-rest guarantee for the feature. This migration converts both columns to
|
|
-- the AES-256-GCM `*_encrypted` BYTEA convention shared with the note columns;
|
|
-- the application encrypts on write and decrypts at confirm time.
|
|
--
|
|
-- Transactional + fail-closed: encryption is application-layer, so the plaintext
|
|
-- JSONB cannot be re-encrypted in SQL. Staged facts are EPHEMERAL pre-confirm
|
|
-- (nothing committed until the user approves them) and this feature has not
|
|
-- shipped, so any rows staged by a pre-release build carry plaintext PHI we
|
|
-- will not retain — they are dropped here rather than left unencrypted, and
|
|
-- their parent document keeps its encrypted original and can be re-extracted.
|
|
-- The whole migration runs in one transaction, so a failure rolls back wholly.
|
|
--
|
|
-- Reversibility (down) — re-add the plaintext columns (data is not recoverable):
|
|
-- ALTER TABLE "extracted_facts" ADD COLUMN "data_json" JSONB;
|
|
-- ALTER TABLE "extracted_facts" ADD COLUMN "provenance_json" JSONB;
|
|
-- ALTER TABLE "extracted_facts" DROP COLUMN "data_encrypted";
|
|
-- ALTER TABLE "extracted_facts" DROP COLUMN "provenance_encrypted";
|
|
|
|
-- Add the encrypted replacements (nullable during the swap).
|
|
ALTER TABLE "extracted_facts" ADD COLUMN IF NOT EXISTS "data_encrypted" BYTEA;
|
|
ALTER TABLE "extracted_facts" ADD COLUMN IF NOT EXISTS "provenance_encrypted" BYTEA;
|
|
|
|
-- Drop any pre-release staged rows: their plaintext payload cannot be
|
|
-- re-encrypted in SQL, and a pre-confirm fact carries no committed data.
|
|
DELETE FROM "extracted_facts" WHERE "data_encrypted" IS NULL;
|
|
|
|
-- Enforce the at-rest guarantee: every staged fact now carries ciphertext.
|
|
ALTER TABLE "extracted_facts" ALTER COLUMN "data_encrypted" SET NOT NULL;
|
|
ALTER TABLE "extracted_facts" ALTER COLUMN "provenance_encrypted" SET NOT NULL;
|
|
|
|
-- Drop the plaintext columns so no verbatim clinical span survives.
|
|
ALTER TABLE "extracted_facts" DROP COLUMN IF EXISTS "data_json";
|
|
ALTER TABLE "extracted_facts" DROP COLUMN IF EXISTS "provenance_json";
|