Hospitals are attacked by a small set of ransomware crews and state-backed groups who know that clinical downtime puts pressure on victims to pay. This guide maps the actors that government advisories name as targeting healthcare and public health, from ransomware-as-a-service operations such as Medusa, Rhysida, ALPHV and Interlock to North Korea's Andariel, which used ransomware against US healthcare to fund espionage. For each it summarises how they get in, how they move, and which controls the advisories say would have stopped them. It is written for hospital and health system CISOs, SOC teams, clinical engineering leads and the suppliers that serve them. Attribution follows the named sources; it does not repeat our broader Ransomware Groups Guide.
Key takeaways
- Healthcare is a named, repeated target: CISA, FBI and HHS advisories list healthcare among the main victim sectors for Medusa and Rhysida, and say healthcare became the most victimised sector for ALPHV affiliates after mid-December 2023.
- Identity is the usual front door: stolen VPN credentials without MFA, help desk impersonation and adversary-in-the-middle phishing recur across the advisories, as they did at Change Healthcare.
- Edge and remote access software is the other door: Medusa actors exploit ScreenConnect, Fortinet EMS, GoAnywhere and BeyondTrust flaws, often within 24 hours of a public exploit, according to the 2026 advisory update.
- State actors use healthcare for money as well as data: the July 2024 joint advisory on Andariel says the group funds espionage through ransomware against US healthcare entities.
- The controls repeat across every advisory: phishing-resistant MFA on remote access, rapid patching of known exploited vulnerabilities, allowlisting of remote access tools, segmentation and offline immutable backups.
Why attackers pick healthcare
- Downtime is dangerous: when electronic records, imaging, laboratory or pharmacy systems stop, care is diverted or delayed. That urgency is what extortion relies on.
- Data is valuable twice: patient records support extortion and fraud, and leak threats work even when backups allow recovery. See the Data Extortion Without Encryption Guide.
- Large, mixed estates: legacy clinical systems, medical devices that cannot run endpoint agents, many third-party remote support connections and shared clinical workstations widen the attack surface.
- Concentrated suppliers: one clearing house, laboratory or software supplier can serve thousands of providers, so one intrusion disrupts a whole sector.
- Retaliation and signalling: the ALPHV advisory links the surge in healthcare victims to an administrator urging affiliates to target hospitals after law enforcement action in December 2023.
The actor map
| Actor | Type | Healthcare link (per source) | Main access routes (per source) | Source |
|---|---|---|---|---|
| Medusa (ATT&CK G1051) | Ransomware-as-a-service since an affiliate model by early 2023; over 500 victims as of April 2026 | Healthcare and Public Health named as a frequent victim | Initial access brokers, phishing, unpatched CVEs on edge and remote management software | CISA, FBI and HHS, AA25-071A (updated 18 August 2026) |
| Rhysida | Ransomware-as-a-service, active since May 2023 | Healthcare among the main targeted sectors with education, manufacturing, IT and government | VPN logins with stolen credentials where MFA was not enabled, Zerologon (CVE-2020-1472), phishing | FBI, CISA and MS-ISAC, AA23-319A |
| ALPHV (BlackCat) | Ransomware-as-a-service | Most of nearly 70 leaked victims since mid-December 2023 were in healthcare | Help desk impersonation by phone or SMS, Evilginx2 adversary-in-the-middle phishing to capture MFA codes and session cookies | FBI, CISA and HHS, AA23-353A (updated 27 February 2024) |
| Interlock | Ransomware with double extortion | Advisory co-sealed by HHS with guidance for healthcare organisations; it names businesses and critical infrastructure in North America and Europe | Drive-by downloads from compromised websites, ClickFix fake CAPTCHA prompts, fake browser and security updates | FBI, CISA, HHS and MS-ISAC, AA25-203A |
| Andariel (ATT&CK G0138) | North Korean state group, RGB 3rd Bureau | Funds espionage through ransomware against US healthcare; medical sector a secondary espionage target | Exploiting public-facing servers such as Log4Shell (CVE-2021-44228), phishing with LNK or HTA files | US, South Korean and UK agencies, AA24-207A |
Deeper profiles: Medusa Ransomware Threat Actor Profile, Rhysida Ransomware Profile Guide, Interlock Ransomware Profile Guide and Andariel Threat Actor Profile Guide. In Europe, ENISA's 2026 threat landscape gives health highlight coverage and points to its health sector cyber hygiene guidance; see the ENISA Threat Landscape Guide.
How they get in
| Entry route | ATT&CK technique | Actors named for it | Control the advisories recommend |
|---|---|---|---|
| VPN or remote portal login with stolen credentials | T1078 Valid Accounts, T1133 External Remote Services | Rhysida, ALPHV affiliates | Phishing-resistant MFA on VPN, webmail and critical systems |
| Exploiting internet-facing software | T1190 Exploit Public-Facing Application | Medusa (CVE-2024-1709 ScreenConnect, CVE-2023-48788 FortiClient EMS, CVE-2025-10035 GoAnywhere MFT, CVE-2026-1731 BeyondTrust), Andariel (CVE-2021-44228 Log4j) | Patch known exploited vulnerabilities on internet-facing systems first |
| Phishing and adversary-in-the-middle | T1566 Phishing, T1557 Adversary-in-the-Middle, T1539 Steal Web Session Cookie | ALPHV affiliates, Medusa, Andariel | FIDO or PKI-based MFA, which resists relay of codes and cookies |
| Help desk impersonation | T1566 Phishing (voice and SMS lures) | ALPHV affiliates | Strong caller verification before resets; see the Help Desk Social Engineering Incident Playbook |
| Drive-by and paste-and-run lures | T1189 Drive-by Compromise, T1204.004 Malicious Copy and Paste | Interlock | DNS filtering, user training, restricting PowerShell for standard users; see the ClickFix and Fake CAPTCHA Paste-and-Run Attacks Guide |
| Bought access | Initial access brokers (Medusa offers $100 to $1 million on criminal forums) | Medusa | Monitor for leaked credentials and close unused remote access |
What they do once inside
- Credential theft: LSASS dumping (T1003.001) with Mimikatz or comsvcs.dll and theft of ntds.dit (T1003.003) are named for Medusa, Rhysida and Andariel. Interlock may use Kerberoasting (T1558.003) to reach domain admin accounts.
- Remote access tools: AnyDesk, ConnectWise, SimpleHelp, Splashtop and similar tools (T1219) are used for persistence by Medusa, Rhysida and ALPHV affiliates. Allowlisting approved tools removes this cover.
- Lateral movement: RDP (T1021.001), PsExec and PuTTY move the intrusion between servers.
- Defence impairment: Medusa disables Defender and EDR and hides tools in Defender exclusion folders; Rhysida clears event logs with wevtutil.
- Exfiltration then encryption: Rclone, Mega and Dropbox, AzCopy to Azure storage and cloud accounts (T1567.002) carry data out before encryption, so the leak threat outlives the recovery. Interlock and ALPHV can encrypt virtual machines and VMware hosts.
Suppliers, devices and the shared attack surface
Many healthcare intrusions start or spread outside the hospital's own network. The advisories point at the software that suppliers use to reach in: Medusa's CVE list is dominated by remote support, endpoint management and file transfer products, the same categories that device makers and IT providers use to service hospitals.
- Remote support platforms: ScreenConnect and BeyondTrust appear both as exploited products and as tools attackers install. Treat every supplier remote support instance as an internet-facing system with its own patch owner.
- Managed file transfer: GoAnywhere MFT and similar products move claims, imaging and laboratory files between organisations. A flaw there exposes bulk patient data at once. Restrict their administrative interfaces to internal networks.
- Clearing houses and shared services: when one supplier is hit, thousands of providers lose a function. Ask key suppliers how they enforce MFA on remote access and how fast they will notify you. See the Third-Party Incident Notification Clauses Guide.
- Medical devices: many cannot run EDR. Segment them, monitor their traffic passively and keep a list of which vendor can reach each device and how.
When a named actor hits: first decisions
- Activate clinical downtime procedures alongside the technical response, so care decisions do not wait on IT.
- Isolate, do not power off: the ALPHV advisory advises isolating affected hosts, reimaging compromised systems and issuing new credentials. Preserve memory and logs where it is safe to do so.
- Reset identities in the right order: privileged and service accounts first, including the krbtgt account where domain compromise is suspected, then remote access and supplier accounts. See the Identity Incident Response Playbook.
- Check the leak site: double extortion groups publish victim names; confirm what data has been stolen before you communicate. See the Ransomware Data Leak Site Listing Playbook.
- Report: US organisations report to CISA or the FBI, and the advisories ask for indicators to be shared. Regulatory notification duties apply in parallel.
Where identity fits
Read the advisories together and the strongest single control is identity assurance on remote access. Rhysida's VPN logins worked because MFA was not enabled. ALPHV affiliates beat code-based MFA with adversary-in-the-middle phishing and talked help desks into resets. Medusa's advisory asks for strong passwords on service and admin accounts and a review of Active Directory for unrecognised accounts.
- Remote access: require phishing-resistant MFA on VPN, Citrix and remote support portals, including for suppliers. See the Remote Access Identity Guide.
- Shared clinical workstations: fast user switching and tap-and-go badges are common; make sure the underlying accounts are not shared domain admins. See the Healthcare Identity Security Guide.
- Service accounts and integrations: interface engines, PACS and laboratory systems run on service accounts with broad rights and rarely rotated passwords. These are targets for Kerberoasting and NTDS theft. See the Group Managed Service Accounts (gMSA) Guide.
- Vendor accounts: remote support accounts for device makers should be named, time-limited and monitored. See the Privileged Access Management Guide.
Implementing the defences: a phased plan
- First 30 days, close the front doors: list every internet-facing remote access path (VPN, Citrix, RDP gateways, remote support tools, file transfer). Enforce MFA on all of them, starting with administrator and supplier accounts. Patch any listed in the CISA KEV catalogue. Remove remote support tools that are not approved.
- Days 30 to 90, harden identity: move administrators and remote users to FIDO2 or certificate-based MFA. Tighten help desk verification for password and MFA resets. Discover service accounts, rotate passwords and move suitable ones to gMSA. Pilot with one division and keep the old MFA method as a rollback until enrolment is complete.
- Days 90 to 180, contain: segment clinical devices and legacy systems from user networks, restrict RDP and SMB between servers, and allowlist remote access software. Test segmentation in monitor mode before blocking so clinical traffic is not cut by mistake.
- Ongoing, prepare to recover: keep encrypted, immutable, offline backups and test restoration of clinical systems. Run downtime procedure exercises with clinical staff, not just IT. See the Ransomware Response and Recovery Guide.
Decision guidance: large systems and small practices
| Area | Large hospital or health system | Small practice or clinic |
|---|---|---|
| Remote access | Privileged access management with session recording for suppliers | Remove direct RDP; use a managed remote access service with MFA |
| Detection | 24x7 SOC with EDR, identity threat detection and log forwarding | Managed detection service; ensure backups are offline |
| Medical devices | Dedicated segmentation and passive monitoring | Separate network for devices and vendor-managed patching |
| Supplier risk | Contractual MFA and notification clauses, assessment programme | Ask key suppliers how they protect remote access |
Small practices should read the Cyber Security for Dental and GP Practices Guide.
Detection and measurement
- What to log: VPN and remote portal sign-ins with MFA result, help desk reset tickets, new remote access tool installations, LSASS access events, shadow copy deletion, large outbound transfers to cloud storage, and changes to Defender exclusions.
- High-value alerts: successful VPN login without MFA, a password and MFA reset followed by a sign-in from a new device, Rclone or AzCopy on a server, and service accounts logging in interactively.
- Use the advisories: each #StopRansomware advisory publishes ATT&CK techniques and indicators in STIX format. Validate controls against those techniques and vet indicators before blocking, as the Medusa advisory warns some may be legitimate.
Metrics that show the programme is working:
- Percentage of remote access paths, including supplier paths, behind phishing-resistant MFA.
- Median days to patch KEV-listed vulnerabilities on internet-facing systems.
- Number of unapproved remote access tools found per month (target: falling to zero).
- Percentage of service accounts with an owner and rotated credentials.
- Time to restore a priority clinical system from offline backup in the last test.
Framework mapping
| Control area | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 |
|---|---|---|
| MFA for remote and privileged access | IA-2, IA-5, AC-17 | PR.AA (identity management, authentication and access control) |
| Least privilege and least functionality | AC-6, CM-7 | PR.AA, PR.PS |
| Segmentation | SC-7 | PR.IR |
| Backups and recovery | CP-9 | PR.DS, RC.RP |
| Monitoring | SI-4 | DE.CM |
For the regulatory side (HIPAA, NIS2 for health, UK requirements), see the Healthcare Cyber Regulation Map.
Common failure patterns
- One remote portal without MFA: a single legacy Citrix or VPN gateway outside the MFA policy undoes the rest of the programme.
- MFA that can be relayed: push and code-based MFA did not stop ALPHV affiliates using adversary-in-the-middle kits; phishing-resistant methods do.
- Help desks that reset on request: a confident caller with public information can obtain a reset unless verification is strict.
- Unmanaged remote support tools: tools installed by suppliers or staff become attacker persistence that blends in.
- Backups reachable from the domain: if backup consoles use domain credentials, stolen domain admin rights can reach and destroy them.
- Treating payment as the end: data stolen before encryption can be leaked or resold even after a ransom is paid.
Lessons from incidents
The Change Healthcare Breach 2024 shows the pattern in one case. Attackers used a stolen credential on a Citrix remote access portal that did not require MFA on 12 February 2024, spent about nine days inside, and deployed ALPHV/BlackCat ransomware on 21 February, disrupting claims and pharmacy payments across the US. UnitedHealth paid a $22 million ransom, the data was later claimed by RansomHub, and the company's August 2025 update put affected people at 192.7 million. The same weeks match the ALPHV advisory's warning that healthcare had become its most victimised sector. One remote portal with MFA would have changed the outcome.
Practitioner checklist
- Inventory every remote access path, including supplier and device vendor connections, and enforce MFA on all of them.
- Move administrators and remote users to phishing-resistant MFA.
- Patch KEV-listed vulnerabilities on internet-facing systems within days, not weeks.
- Allowlist approved remote access tools and alert on any others.
- Strengthen help desk verification for password and MFA resets.
- Discover service accounts, assign owners, rotate credentials and remove interactive logon rights.
- Segment clinical devices and legacy systems and restrict RDP and SMB between servers.
- Keep offline, immutable backups with separate credentials and test clinical restores.
- Validate detections against the ATT&CK techniques in the Medusa, Rhysida, ALPHV, Interlock and Andariel advisories.
Standards and references
- CISA, FBI and HHS: #StopRansomware: Medusa Ransomware (AA25-071A, updated August 2026)
- CISA, FBI and MS-ISAC: #StopRansomware: Rhysida Ransomware
- FBI, CISA and HHS: #StopRansomware: ALPHV Blackcat
- FBI, CISA, HHS and MS-ISAC: #StopRansomware: Interlock
- FBI, CISA, NSA, NCSC and partners: North Korea Cyber Group Conducts Global Espionage Campaign (AA24-207A)
- CISA: Known Exploited Vulnerabilities Catalog
- MITRE ATT&CK: Medusa Group
- NIST: SP 800-53 Rev. 5
Related NHI Mgmt Group resources: Healthcare Identity Security Guide · Ransomware Groups Guide · Healthcare Cyber Regulation Map · Medusa Ransomware Threat Actor Profile · Ransomware Prevention Guide