TRAC
Enter Workspace
HIPAA COMPLIANT ARCHITECTURE

HIPAA & Data Security Policy

Last Updated: May 20, 2026 • Version 2.0 (Approach A Secure)

01. Statement of HIPAA Compliance

At TRAC (under Systemores LLC), we are committed to maintaining the highest security, privacy, and integrity standards for litigation data. The TRAC platform is fully designed to satisfy the rigorous requirements of the Health Insurance Portability and Accountability Act (HIPAA) and the Health Information Technology for Economic and Clinical Health (HITECH) Act.

Our platform handles litigation intake, medical charts, and intake forms for mass tort clinical analysis with a multi-layered security model that guarantees Patient Health Information (PHI) is protected, audited, and isolated.

02. Approach A: Pre-Processing De-Identification

To protect patient privacy without incurring the overhead of enterprise cloud agreements for external processing engines, TRAC utilizes our proprietary Approach A: In-Memory De-identification & Re-hydration protocol.

Technical Extraction Flow:

  • Local Sanitization: Before unstructured client documents are submitted to external AI extraction engines (such as the Gemini API), all direct identifiers—including patient full names, dates of birth (DOBs), social security numbers (SSNs), email addresses, phone numbers, and Medical Record Numbers (MRNs)—are scanned locally in-memory.
  • Token Substitution: PHI values are stripped and substituted in-memory with unique, context-preserving identifiers (e.g., __TRAC_TOKEN_NAME_1__).
  • Anonymized Processing: The external AI processing pipeline operates exclusively on de-identified, tokenized datasets, ensuring zero PHI leaks to external servers.
  • In-Memory Re-hydration: When the structured clinical timelines are returned in JSON format, TRAC securely re-hydrates the tokens back into their corresponding verbatim clinical quotes and date fields locally in-memory.
  • Secure Database Insertion: The fully hydrated records are immediately validated and saved into our secure database instance.

03. Strict Multi-Tenant Isolation

TRAC is built on a strict multi-tenant architecture designed to keep legal firms and healthcare organizations completely isolated:

  • Database Schema Partitioning: Every case, document, timeline event, and integrity flag is tied directly to a validated organization_id.
  • Row-Level Security (RLS): Our database instance enforces strict PostgreSQL RLS policies on all tables. This ensures that legal team members can only query data matching their own organization boundary.
  • Super Admin Controls: Access by system-wide administrators is rigorously authenticated, logged, and isolated down to standalone active filters when reviewing specific litigation pipelines.

04. Practice Management Integrations & API Connections

When Subscribers link their TRAC workspace to external Practice Management Systems (specifically Clio or Filevine), we utilize industry-standard security boundaries to govern the data pipelines:

  • De-identification Ingress: Unstructured legal matters and document summaries pulled from Clio or Filevine are instantly parsed through our in-memory Approach A de-identification engines before they are stored or processed.
  • Isolated Token Management: API keys, client OAuth credentials, and tokens are stored in the database under strict multi-tenant encryption boundaries.
  • Scope Access Limits: Integration access is governed by granular token privileges and is restricted strictly to fetching legal matters authorized by the Tenant Admin.

05. Immutable Security Audit Logging

To satisfy HIPAA auditing requirements, the platform incorporates a completely tamper-proof logging system:

Every action involving PHI reads, case ingestions, and summary report exports triggers an automatic insertion into our audit_logs database.

TAMPER-PROOF AUDIT ENFORCEMENT

PostgreSQL RLS policies on our audit trail tables explicitly block all UPDATE and DELETE commands for all users, including administrative roles. Once logged, compliance audit logs are permanently written and cannot be altered or removed.

06. PII Masking & Privacy Mode 2.0

For screen sharing, peer consultations, or print exports, TRAC offers a secure server-bound Privacy Mode. When activated:

  • All names, DOBs, SSNs, MRNs, and telephone numbers within verbatim clinical notes are redacted directly at the API payload boundary before the response leaves our secure server environment.
  • This ensures that no raw clinical PHI resides inside client-side browser caches or local states while Privacy Mode is active.

Contact Compliance Officer

For security questionnaires, audits, or questions regarding our Approach A de-identification pipeline, please contact compliance@systemores.com.