Education Center

Healthcare Data Protection Essentials

Understanding what constitutes protected information, how it can be compromised, and what your organization can do to maintain both compliance and patient trust.

Understanding Data

What Constitutes Protected Information?

Healthcare organizations handle multiple categories of sensitive data. Understanding the distinctions is critical for proper handling, storage, and transmission.

Protected Health Information (PHI)

Under HIPAA, PHI is any health information that can identify an individual and is created, received, maintained, or transmitted by a covered entity.

The 18 HIPAA Identifiers:

• Names • Geographic data (smaller than state) • Dates (birth, admission, discharge) • Phone numbers • Fax numbers • Email addresses • Social Security numbers • Medical record numbers • Health plan numbers • Account numbers • Certificate/license numbers • Vehicle identifiers • Device identifiers • Web URLs • IP addresses • Biometric identifiers • Full-face photographs • Any other unique identifier

Personally Identifiable Information (PII)

PII is broader than PHI and includes any information that can identify an individual. While HIPAA covers PHI, state laws and other regulations may protect PII more broadly.

Common PII in Healthcare:

• Employee records and payroll data • Vendor and business partner information • Financial and insurance records • Operational data with embedded identifiers • Marketing and communications lists

Key distinction: All PHI is PII, but not all PII is PHI. A patient's email address in a medical record is PHI. The same email in a newsletter list is PII but not PHI.

Safe Harbor De-identification

HIPAA provides a Safe Harbor method for de-identifying PHI. If all 18 identifiers are removed, the information is no longer considered PHI and can be used for research, quality improvement, and other secondary purposes without patient authorization.

Safe Harbor Method

Remove all 18 identifiers. No expertise required. Provides certainty that data is no longer PHI.

Expert Determination

Statistical or scientific expert certifies that the risk of re-identification is very small.

Compliance

What Violates HIPAA?

Understanding common violations helps organizations implement effective safeguards and training programs to prevent incidents before they occur.

⚠ Minimum Necessary Standard

Accessing or disclosing more PHI than necessary for a given purpose. Examples: looking up celebrity records out of curiosity, sharing full charts when only a diagnosis is needed, including unnecessary PHI in communications.

⚠ Improper Disclosure

Sharing PHI with unauthorized individuals or through insecure channels. Examples: discussing patients in public areas, sending PHI via unencrypted email, posting case details on social media, leaving records visible.

⚠ Insufficient Access Controls

Failing to implement appropriate technical and administrative controls. Examples: shared passwords, lack of role-based access, terminated employees retaining system access, unlocked workstations in clinical areas.

⚠ Lack of Business Associate Agreements

Sharing PHI with vendors or partners without proper contractual protections. Examples: using cloud storage without BAA, third-party billing without agreements, IT support accessing systems without safeguards.

⚠ Inadequate Risk Analysis

Failing to conduct regular, thorough assessments of security risks. Examples: no documented risk assessment, assessments not updated after system changes, ignoring identified vulnerabilities, lack of remediation tracking.

⚠ Insufficient Training

Workforce lacking awareness of policies and procedures. Examples: staff unaware of password requirements, no phishing awareness, failure to report suspicious activity, not understanding what constitutes PHI.

OCR Civil Penalty Structure

Tier 1
$137-$68K
Unaware of violation
Tier 2
$1.3K-$68K
Reasonable cause
Tier 3
$13K-$68K
Willful neglect, corrected
Tier 4
$68K-$1.9M
Willful neglect, not corrected

Per violation, per year. Criminal penalties may also apply for intentional wrongful disclosure.

AI Governance

Responsible AI Usage in Healthcare

Generative AI tools offer productivity benefits but introduce unprecedented data exposure risks. Organizations need clear policies that distinguish acceptable use from prohibited activities.

✓ Acceptable Use (with BAA)

  • Microsoft Copilot with Healthcare Data Protection enabled
  • OpenAI ChatGPT Enterprise with signed BAA and disabled training
  • Google Cloud Vertex AI with proper data residency controls
  • Organization-approved clinical documentation tools
  • Internal AI systems with documented security controls

Requirement: Signed Business Associate Agreement, disabled model training on your data, audit logging enabled, and IT/security approval.

✗ Prohibited Use

  • Consumer ChatGPT, Claude, or Gemini for any PHI-containing content
  • Using AI to summarize, translate, or rewrite patient notes
  • Pasting diagnostic information, lab results, or treatment plans
  • Uploading patient images, scans, or documents to AI tools
  • Using AI for coding or billing with patient identifiers
  • Any AI tool without documented BAA and security review

Risk: Data may be stored, used for model training, or accessed by the AI provider without your knowledge or control.

The "No PHI" Rule

When in doubt, apply this simple test: If you removed all patient identifiers from the content, would a reasonable person still be able to identify the individual? If yes, it's PHI and should not go into any AI tool without explicit approval.

✗ Contains PHI

"65-year-old male patient John Smith from St. Louis with Type 2 diabetes presenting with..."

✓ De-identified

"General clinical guideline: Management of Type 2 diabetes in patients over 60 with cardiovascular comorbidities..."

Training

What Staff Need to Know

Effective security awareness training transforms your workforce from a vulnerability into your first line of defense. These are the essential topics every healthcare employee should understand.

🔓

Password & Access Hygiene

Unique passwords for each system, multi-factor authentication on all accounts, never sharing credentials, locking workstations when stepping away, and reporting suspicious access immediately.

📨

Phishing Recognition

Identifying suspicious emails, verifying sender addresses, never clicking unknown links or attachments, recognizing urgency tactics, and reporting suspected phishing attempts to IT.

🗂

Physical Security

Clean desk policies, proper disposal of printed materials, securing portable devices, restricting conversations about patients to private areas, and reporting lost or stolen equipment immediately.

💻

Safe Data Handling

Using only approved applications and storage, understanding what can and cannot be stored in cloud services, proper use of encryption for transmission, and reporting accidental disclosures promptly.

🤖

AI Tool Awareness

Understanding which AI tools are approved, why consumer AI poses risks to PHI, how to de-identify content before using AI, and what to do if PHI is accidentally entered into an unauthorized tool.

🚨

Incident Reporting

Recognizing potential security incidents, understanding the reporting chain, documenting what happened, and knowing that prompt reporting is encouraged and protected — never punished.

Data Storage Guidelines

✓ Approved Storage

  • Organization-managed EHR system
  • Approved cloud storage with BAA (e.g., Google Workspace with BAA, Microsoft 365 GCC)
  • Encrypted local servers with access controls
  • HIPAA-compliant backup systems
  • Secure file transfer systems with audit logging

✗ Prohibited Storage

  • Personal email accounts (Gmail, Yahoo, etc.)
  • Consumer cloud storage (Dropbox, iCloud, Google Drive without BAA)
  • Personal USB drives or external hard drives
  • Personal devices without MDM enrollment
  • Messaging apps (SMS, WhatsApp, consumer Slack)

The Rule: When in doubt, store it in the EHR or approved organizational system. If it's not on the approved list, don't use it.

Need Help Implementing These Practices?

Velari provides customized security awareness training, policy development, and compliance program management tailored to your organization's size, specialty, and risk profile.