← Breach analysis by the Dual Vulnerabilities framework

IZIS Republika Srpska ransomware (Unattributed, 2023)

2023 · Bosnia and Herzegovina · Q - Human health and social work activities · Ransomware attack with data theft and long system-wide outage

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Ransomware attack with data theft and long system-wide outage

Dual vulnerability type

Strategy to Execution

Evidence coverage: 4/6 stages, weight 24/60

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

Why this classification

(1) IZIS is the legally mandated Integrated Health Information System of Republika Srpska, a republic project of strategic importance holding all patients' health records and insurance status, funded with about 26 million KM, with the mandatory cutover to electronic health cards announced for 1 January 2024 - the strategic expectation was a safe, always-available, entity-wide digital health service. (2) Operationally, on 31 December 2023 ransomware reached the central database through one of the 10,500 connected computers (FZO itself assumed the attacker came through one of the special security tunnels used by external users, posing as a legitimate user), locked the base, and the whole entity's digital healthcare ran on paper for 17 days; restoration on 17 January 2024 covered only 6 of 74 institutions with laboratory services still excluded and no date for the remaining 68, the Moj karton patient app stayed withdrawn for safety reasons two years later, and reporting notes that more than a year after implementation the system never worked fully. Meanwhile citizens' data surfaced for sale on dark-web markets while the Fund insisted all data was preserved and safe. (3) The classification rests on the recovery stage: the continuity expectation is doctrinal - written into the law, the programme and the announced e-card mandate - yet it never translated into operational reality: restore and re-connect work left most institutions offline for weeks and one front-end permanently dead. That is the diagonal gap: formal strategic artefacts exist but do not become operational IT reality (Strategy to Execution). The containment reading points the same way (a designed single central database turned one endpoint compromise into an entity-wide outage), so the entry vector - officially unresolved, tunnel-vs-endpoint per the victim's own assumption - is not load-bearing. A secondary pattern is Strategy to Leadership in the response stage: leadership's public account of data safety was contradicted by the reported leak and the scale of compromise was never disclosed. (4) RM04 (Strategy to Execution lens): the architecture and operating model concentrated every institution's records and services in one central database reached over shared tunnels, so one compromised computer forced an entity-wide outage and left almost no recovery options (causal role: Amplifying). RM31 (Strategy to Execution lens): continuity requirements for the entity's crown-jewel health data were not translated into tested, staged restores - 17 days fully offline, only 6 of 74 institutions back weeks later, lab integrations and the Moj karton app never restored to date (causal role: Constraining). RM39 (Strategy to Execution lens): notification was never operationalised - no accurate account of what data was lost, no guidance to data subjects, only assurances that data was safe while it was reported for sale (causal role: Constraining).

Evidence by stage

Risk markers

Sources

Notes

EuRepoC row; victim, country (Bosnia and Herzegovina, receiver_country) and year (2023, start_date 31.12.2023) taken from the case file. Actor unattributed: the Fund and the RS ICT Agency speak of an advanced international organised hacking group; the MUP RS investigation published no forensic report, so entry remains officially unresolved (FZO assumed a compromised external security tunnel; Balkan Insight reports one of the 10,500 connected endpoints) - entry is recorded but not load-bearing. Dark-web sale of the data rests on third-party monitoring (SocSecurity) and local press, not on a victim confirmation; the Fund maintained the database was preserved. Quotation marks only for the dataset's own wording. The Fund's statements were read via regional press reproducing them in full (RTRS, BL Portal); no separate regulator or DPA decision on the incident was found in the record searched.