← Breach analysis by the Dual Vulnerabilities framework

University Urology PHI data breach (Unattributed, 2023)

2023 · United States · Q - Human health and social work activities · Data theft via network intrusion (PHI exfiltration)

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Data theft via network intrusion (PHI exfiltration)

Dual vulnerability type

Expertise to Execution

Evidence coverage: 3/6 stages, weight 18/60

Decisive stage: detection (clearest evidence) · Alternative: Strategy to Leadership

Why this classification

(1) As a HIPAA-covered urology practice holding protected health information of tens of thousands of patients on network servers, the expectation was technical safeguards that detect and contain security events (audit controls, endpoint detection, controlled remote access) and egress protection around the patient-data crown jewels. (2) In practice, an unknown actor entered internal data systems and obtained credentials, PHI and billing data of 56,816 patients; unauthorized remote access tools and persistence mechanisms sat in the environment unnoticed until suspicious activity was detected on 1 February 2023; third-party forensics needed a month (concluded 3 March) and a further three weeks of manual file review (concluded 30 March) to scope what had been accessed; data left undetected; notices went out only on 1 May. The entry vector is undisclosed and is not load-bearing. (3) This is a knowing-doing gap within the material column: every remedy the victim deployed after detection - SentinelOne agents, 24/7 SOC monitoring, password resets, removal of unauthorized remote-access tools, limiting remote access, exported backups - was standard, available expertise that was simply not executed in the environment beforehand; the classification rests on the detection stage (clearest evidence, the victim's own disclosure corroborated by the SDNY order), with the undetected exfiltration as a second material-column gap; the entry vector remains unknown. A secondary pattern sits in the response: notification 89 days after detection on a doctrinal duty (see Alternative DV-Type). RM34 (Expertise to Execution lens): available detection tooling was not configured, staffed or maintained - the practice had no continuous monitoring of its own and had to rent SentinelOne agents plus an external 24/7 SOC for 30 days only after the incident, while unauthorized remote-access tools and persistence ran unnoticed beforehand (causal role: Obscuring). RM35 (Expertise to Execution lens): egress monitoring that should have existed on a PHI repository was absent, so the records of 56,816 patients left the network without any alert or interruption (causal role: Amplifying). RM33 (Expertise to Execution lens): forensic-readiness was too weak to scope the incident from tooling alone - a month of third-party forensics plus a manual review of files took until 30 March to establish what had been accessed (causal role: Obscuring). RM08 (Expertise to Execution lens): specialists know these safeguards, but the practice's own capacity could not operationalise them - monitoring, forensics and containment cleanup all had to be bought in from outside after the fact (causal role: Enabling).

Evidence by stage

Risk markers

Sources

Notes

EuRepoC row; victim, country and year taken from the case file fields (receiver_country United States, start_date 01.02.2023). The victim's own data-incident page (universityurology.com/data-incident/) no longer serves content directly; it was read via an opened Wayback snapshot of 5 May 2023, which preserved the page shell, and its notice text via the two secondary articles that quote it in detail. The SDNY Memorandum Order of 31 March 2025 (motion to dismiss denied) states the facts as pleading-stage allegations accepted as true, so the exfiltration finding is marked primary-open; no regulator (OCR) decision on the incident was found. Actor genuinely unattributed; this does not lower confidence. The notification timing (1 Feb detected, 1 May notices) may fall within statutory deadlines depending on when discovery is dated, so the response-stage leadership reading is flagged as inference only. No related rows for this victim in the dataset; a same-named Knoxville practice in breach reporting is a different organisation and was not used. Link check: databreaches.net blocks automated access (HTTP 403); the page was opened and read during research.