Back to Intelligence

NYC Health + Hospitals Breach: Detecting and Mitigating Large-Scale PHI Data Exfiltration

SA
Security Arsenal Team
May 19, 2026
6 min read

Introduction

The recent confirmation of a data breach at NYC Health + Hospitals (NYC H+H) affecting up to 1.8 million individuals is a stark reminder of the criticality of healthcare data defense. Discovered in late March, this incident involved the unauthorized access to sensitive patient information, including names, dates of birth, and medical record numbers. For defenders, this is not merely a headline; it is an indicator of a systemic failure to detect lateral movement and data staging. The exposure of Protected Health Information (PHI) of this magnitude suggests a prolonged dwell time or a high-impact extraction mechanism. Immediate action is required to audit database access controls and identify exfiltration indicators within healthcare environments.

Technical Analysis

While the specific vector in the NYC H+H case may be attributed to unauthorized network access, breaches of this scale in the healthcare sector typically follow a pattern involving Database Discovery, Credential Dumping, and Mass Data Exfiltration.

  • Affected Platforms: Electronic Health Record (EHR) systems, Microsoft SQL Server, and file servers storing unstructured PHI.
  • Attack Vector (Defensive Perspective): The attacker likely leveraged valid credentials obtained via phishing or previous compromises to access internal databases. Once inside, the chain involves:
    1. Reconnaissance: Querying database schemas (information_schema) to identify tables containing PII/PHI.
    2. Staging: Using native database utilities (e.g., sqlcmd, bcp) or PowerShell scripts to dump data into local temporary files (CSV, SQL, TXT).
    3. Exfiltration: Compressing the staged files using utilities like 7-Zip or WinRAR, followed by outbound transfer over non-standard ports (e.g., DNS tunneling or HTTP/HTTPS to cloud storage).
  • Exploitation Status: Confirmed active data theft. While no specific zero-day CVE (e.g., CVE-2024-XXXX) has been publicly attributed as the initial entry point in the summary, the TTPs align with standard Credential Access and Collection techniques observed in FIN7 and targeted ransomware operations.

Detection & Response

Given the lack of a specific CVE in the report, defense relies heavily on identifying the behavior of data theft. The following rules target database dumping mechanisms and mass compression activities commonly used during PHI exfiltration.

SIGMA Rules

YAML
---
title: Potential SQL Database Dumping via sqlcmd
id: 4a3f1c82-9e4b-4d67-bc12-3e5a8f901234
status: experimental
description: Detects the use of sqlcmd to output query results to a file, a common method for database staging prior to exfiltration.
references:
  - https://attack.mitre.org/techniques/T1005/
author: Security Arsenal
date: 2024/04/10
tags:
  - attack.collection
  - attack.t1005
logsource:
  category: process_creation
  product: windows
detection:
  selection:
    Image|endswith: '\sqlcmd.exe'
    CommandLine|contains:
      - '-o '
      - '-o"'
      - 'output '
  condition: selection
falsepositives:
  - Legitimate administrative database exports
level: high
---
title: Mass Compression in User Directories
id: 5b2f1c99-1e5c-4e78-ad12-4e5a8f901567
status: experimental
description: Detects high-volume compression processes (7z, winrar) running in user profile directories, indicative of data staging for exfiltration.
references:
  - https://attack.mitre.org/techniques/T1560/
author: Security Arsenal
date: 2024/04/10
tags:
  - attack.collection
  - attack.t1560.001
logsource:
  category: process_creation
  product: windows
detection:
  selection_img:
    Image|endswith:
      - '\7z.exe'
      - '\winrar.exe'
      - '\rar.exe'
  selection_cli:
    CommandLine|contains:
      - '-tzip'
      - '-a'
  selection_path:
    CurrentDirectory|contains:
      - 'C:\Users\'
      - 'C:\ProgramData\'
  condition: all of selection_*
falsepositives:
  - Users backing up their own files
level: medium

KQL (Microsoft Sentinel)

KQL — Microsoft Sentinel / Defender
// Hunt for large outbound data transfers or suspicious database queries
// Look for sqlcmd processes writing to disk
DeviceProcessEvents
| where Timestamp >= ago(7d)
| where FileName in~ ("sqlcmd.exe", "bcp.exe")
| where ProcessCommandLine has_any ("-o", "output", "queryout")
| project Timestamp, DeviceName, AccountName, ProcessCommandLine, InitiatingProcessFileName
| extend FileDest = extract(@'output\s*["']?(.*?)\s*["']?', 1, ProcessCommandLine)
| summarize count() by DeviceName, AccountName, FileDest
| where count_ > 10 // Threshold for bulk operations

Velociraptor VQL

VQL — Velociraptor
-- Hunt for recently created archive files in user directories
SELECT FullPath, Size, Mtime, Mode
FROM glob(globs="/*/*.zip", root="/C:/Users/")
WHERE Mtime > now() - 7d
  AND Size > 10000000 // Greater than 10MB

-- Hunt for sqlcmd execution
SELECT Pid, Name, CommandLine, Exe
FROM pslist()
WHERE Name =~ "sqlcmd.exe"
  AND CommandLine =~ "-o"

Remediation Script (PowerShell)

PowerShell
# Script to Audit Recent Database Export Activities and Compressed Files
# Requires Administrative Privileges

Write-Host "[+] Starting PHI Exfiltration Audit..." -ForegroundColor Cyan

# 1. Check for recent sqlcmd usage in Event Logs
Write-Host "[*] Checking for sqlcmd execution in Security Log..." -ForegroundColor Yellow
$Events = Get-WinEvent -LogName Security -FilterXPath "*[System[(EventID=4688)]] and *[EventData[Data[@Name='NewProcessName'] and (contains(string(), 'sqlcmd.exe'))]]" -ErrorAction SilentlyContinue

if ($Events) {
    $RecentEvents = $Events | Where-Object { $_.TimeCreated -gt (Get-Date).AddDays(-7) }
    Write-Host "[!] Found $($RecentEvents.Count) sqlcmd executions in the last 7 days." -ForegroundColor Red
    $RecentEvents | Select-Object TimeCreated, Id, Message | Format-List
} else {
    Write-Host "[-] No recent sqlcmd execution found in logs." -ForegroundColor Green
}

# 2. Scan for large ZIP files created in User Profiles
Write-Host "[*] Scanning User Profiles for large ZIP files..." -ForegroundColor Yellow
$ZipFiles = Get-ChildItem -Path "C:\Users" -Recurse -Include "*.zip", "*.rar", "*.7z" -ErrorAction SilentlyContinue | 
             Where-Object { $_.LastWriteTime -gt (Get-Date).AddDays(-7) -and $_.Length -gt 10MB }

if ($ZipFiles) {
    Write-Host "[!] Found $($ZipFiles.Count) suspicious compressed files (>10MB, created <7 days ago)." -ForegroundColor Red
    $ZipFiles | Select-Object FullName, Length, LastWriteTime
} else {
    Write-Host "[-] No suspicious compressed files found." -ForegroundColor Green
}

Remediation

  1. Audit and Revoke Access: Conduct an immediate audit of all accounts with read/write access to the compromised databases. Revoke access for any dormant or unnecessary accounts. Force a password reset for all users with database privileges, specifically enforcing complex passphrases.
  2. Network Segmentation: Ensure database servers are strictly segmented. Egress traffic from database servers should be whitelisted to specific application servers only. Block internet access from database layers.
  3. Implement Database Activity Monitoring (DAM): Deploy DAM solutions to trigger real-time alerts on bulk export operations (e.g., SELECT * queries returning > 1000 rows) or access to specific sensitive tables (SSN, Diagnosis).
  4. Review Remote Access Logs: If VPN or Remote Desktop (RDP) was used as an entry vector, audit logs for concurrent sessions or logins from anomalous geolocations.
  5. Patient Notification: In accordance with HIPAA Breach Notification Rule, ensure all affected 1.8 million individuals are notified, offering credit monitoring services where applicable.

Related Resources

Security Arsenal Healthcare Cybersecurity AlertMonitor Platform Book a SOC Assessment healthcare Intel Hub

healthcare-cybersecurityhipaa-compliancehealthcare-ransomwareehr-securitymedical-data-breachnychealth-hospitalsdata-exfiltrationphi-breach

Is your security operations ready?

Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.

NYC Health + Hospitals Breach: Detecting and Mitigating Large-Scale PHI Data Exfiltration | Security Arsenal | Security Arsenal