Insights

Cybersecurity in Healthcare: What Hospitals and Health Tech Teams Should Fix First

Engraved-style illustration of an ornate balance scale holding a sun and a crescent moon, on a red background with a torn gray paper strip, faded code text, two floating bubbles, and an Insights label.
Written by
Eldad Rudich
Published on
November 19, 2026

Cybersecurity in healthcare is the work of protecting patient data, clinical systems, and connected medical devices from attacks that can interrupt care as well as expose records. For providers and health tech companies in 2026, that work centers on closing identity and patching gaps without taking clinical systems offline.

Attackers can exploit a new vulnerability faster than many patch cycles absorb it, and clinical operations have to stay online while teams catch up. A proposed overhaul of the HIPAA Security Rule adds a second pressure. Teams are planning for controls that are not yet law, while the HHS Office for Civil Rights (OCR) enforces the rule already on the books. Priorities have to cover both the threats that get in and the obligations that bind today, because those obligations frame what OCR examines after a breach.

TL;DR:

  • The current HIPAA Security Rule is still the law, and OCR is enforcing its risk analysis and log review requirements while the proposed overhaul stalls.
  • Exploited edge devices and stolen identities are among the most common ways into healthcare networks, and help desks are a proven target.
  • Third parties and vendor-hosted EHR environments slow investigations when EHR, identity, cloud, and vendor-session logs cannot be correlated.
  • Response readiness is measured in evidence and recovery: the notification clock runs from discovery, and recovery is where healthcare security leaders rate themselves weakest.

The Threats Hitting Hospitals and Health Tech in 2026

Ransomware pressure on healthcare hit a record in 2026, when Health-ISAC tracked 317 ransomware events against the global health sector in Q3 alone.

1. Ransomware Groups With Healthcare Track Records

Affiliates move between ransomware brands and keep the same post-compromise tradecraft, so defenses keyed to one group name can age quickly. Attributions still show which groups keep returning to healthcare.

Qilin claimed more healthcare victims than any other group in 2025, including ApolloMD and Covenant Health, and Interlock claimed the attacks on DaVita and Kettering Health.

2. Edge Devices and Short Patch Windows

Exploited vulnerabilities caused more healthcare breaches than phishing or credential abuse in Verizon’s 2026 DBIR healthcare snapshot. Across all industries, the full DBIR puts the median time to fully fix a critical, known-exploited vulnerability at 43 days. Medusa, a group CISA names as a frequent attacker of healthcare, uses newly announced exploits within 24 hours.

Exposed VPNs, firewalls, and remote management tools therefore need compensating controls while patches roll out.

3. Identity and Help-Desk Social Engineering

In 2024, HHS’s Health Sector Cybersecurity Coordination Center (HC3) described two successful voice spearphishing scams that diverted payer payments. In one, the caller persuaded the help desk to enroll a new MFA device. Extortion crews use vishing and victim-branded credential sites to capture SSO credentials and MFA codes, then pull data straight out of SaaS applications.

Gaps in MFA coverage open a second route in. SaaS accounts often have MFA while VPNs, firewall admin consoles, and legacy apps do not.

4. Third Parties and Concentration Risk

In the Change Healthcare attack, the reference case for concentration risk, attackers used compromised credentials on a Citrix remote access page without MFA, moved laterally, and deployed ransomware nine days later. Change notified OCR that the incident affected about 192.7 million individuals. Business associates figured in 43% of large breaches in H1 2026. Hospital impact from Change varied with vendor redundancy and how much each hospital relied on Change’s technology.

HIPAA Security Rule Obligations: What Binds You Today

OCR enforces the Security Rule now in effect at 45 CFR Part 164 Subpart C. The proposed overhaul has no compliance date.

Where the Proposed Rule Stands

HHS published its notice of proposed rulemaking (NPRM) on January 6, 2025. It would remove the required/addressable distinction. It would also require MFA, encryption of electronic protected health information (ePHI), an asset inventory and network map reviewed at least every 12 months, written procedures to restore critical systems within 72 hours, and annual compliance audits.

Today’s rule already reaches part of this list. Addressable safeguards such as encryption are not optional: an entity that finds one not reasonable and appropriate must document why and adopt an equivalent alternative where reasonable.

The current agenda entry moves the rule to Long-Term Actions and pushes projected final action to July 2027, a date with no legal force. Provider groups have asked HHS to withdraw the proposal. HHS has neither finalized nor withdrawn it. If it is finalized as proposed, the rule would take effect 60 days after publication, with compliance generally due 180 days after that.

Congress may move first. S.3315, the Health Care Cybersecurity and Resiliency Act of 2026, passed the Senate by unanimous consent on September 30, 2026, and is pending in the House. The bill would direct HHS to update the Security Rule with minimum practices including MFA, encryption, and penetration testing.

Treat the NPRM controls as the direction of travel and build toward them now, since the proposed rule and the Senate bill point to a similar baseline.

What OCR Is Enforcing Now

Enforcement centers on risk analysis, and in 2026 OCR said it would broaden its Risk Analysis Initiative to risk management.

On October 29, 2026, Lehigh Valley Health Network agreed to pay $800,000 in OCR’s 22nd ransomware enforcement action. The potential violations OCR found included no accurate and thorough risk analysis, no regular review of audit logs and access reports, and late notice to patients and HHS. OCR Director Paula M. Stannard said the first two gaps make regulated entities attractive targets for attackers.

The same gaps recur across recent settlements: missing risk analysis, missing log review, and late notification. Audit log review is an enforced requirement today.

Other State and Federal Rules Already in Force

New York’s 10 NYCRR 405.46 has required general hospitals to fully comply since October 2, 2025, and its incident reporting provision has applied since adoption in October 2024. It requires a designated CISO, penetration testing at least annually based on the hospital’s risk assessment, and MFA for external access unless the CISO approves a compensating control. Hospitals must also report to the Department of Health within 72 hours of determining that a cybersecurity incident has occurred.

Health tech companies face two federal regimes as well. Health apps outside HIPAA fall under the FTC Health Breach Notification Rule when they qualify as vendors of personal health records or related entities. Sponsors of qualifying premarket submissions for cyber devices must meet FD&C Act Section 524B, including an SBOM and a postmarket vulnerability plan.

EHR and Cloud Risk: Where Investigations Break Down

Several recent health data incidents began in vendor-hosted infrastructure the hospital does not run, and at least one started with valid credentials. An attacker used stolen customer credentials to reach legacy Cerner servers on or around January 22, 2025, before those servers moved to Oracle Cloud. Oracle identified the activity around February 20.

Aesto Health’s December 2025 incident involved unauthorized AWS access and affected more than 9.5 million individuals across at least 39 healthcare clients.

Intrusions that use valid credentials leave few malware indicators, so detection depends on behavior and audit logs. Those logs rarely connect. EHR access logs, often reviewed through privacy-monitoring tools for snooping, sit apart from Entra ID or Okta sign-ins, cloud control-plane logs, and vendor remote access sessions. Shared responsibility adds friction, because the customer configures controls and owns HIPAA compliance for what it runs in a cloud provider’s services.

Clinical devices complicate remediation, because patching can depend on vendor certification while care continues. Claroty’s 2025 analysis of 351 healthcare organizations found that 89% had Internet of Medical Things (IoMT) devices with ransomware-linked KEVs and insecure internet connectivity, and imaging systems carried the highest risk. Segmenting critical networks and inventorying legacy clinical devices limit that exposure, and brokering vendor remote access through a control point with local approval and immediate revocation after each session closes a common way in.

Response Expectations After an Incident

The HIPAA breach clock starts at discovery, not at the end of forensics. Covered entities must notify individuals without unreasonable delay and no later than 60 calendar days after discovery. For 500 or more individuals, they must notify HHS at the same time, and they must notify media where more than 500 residents of a state or jurisdiction are affected. Smaller breaches go in a log reported to HHS within 60 days of year end.

OCR treats discovery as the first day the incident is known or should reasonably have been known. The burden of proof sits with the covered entity or business associate.

Ransomware incidents raise the evidentiary bar. Under OCR’s 2016 ransomware guidance, encryption of ePHI by ransomware is presumed to be a breach unless a four-factor risk assessment shows a low probability that the data was compromised. Proving that requires logs, EDR telemetry, and network data that show what the attacker touched. If your logging cannot show whether data was viewed or exfiltrated, you will usually end up notifying.

Recovery also means keeping care running. Plans should assume running without core systems for four weeks or longer, with downtime procedures that set out when to declare downtime and cancel elective procedures. When the University of Mississippi Medical Center lost its Epic EHR in February 2026, staff documented care on paper. In Health-ISAC’s 2026 benchmarking survey of 76 health sector security leaders, only 22% rated themselves at the top two maturity levels for recovery, against 60% for detection.

Decision Criteria for Prioritizing Healthcare Security Work

OCR findings and the threat data above point to an order of work:

  1. If your last risk analysis does not enumerate every system holding ePHI, redo it before buying anything. Connect its findings to a documented risk management plan, which the current rule already requires.
  2. Close MFA gaps on VPNs, firewall consoles, and vendor remote access, even where SaaS already has coverage, and add identity verification steps for help-desk MFA resets.
  3. If you cannot correlate EHR access logs with identity and cloud telemetry, check whether that gap also undermines the regular activity review OCR enforces, and fix the data path before tuning detections.
  4. If a single vendor carries claims, EHR hosting, or clinical workflow, map that dependency and drill downtime procedures that assume the vendor is unavailable for weeks.
  5. If you lack 24/7 investigation coverage, decide between building it and outsourcing it. In the same survey, 82% of respondents use SOC or MDR services, and outsourcing is the main route to round-the-clock coverage for security teams under 30 FTEs.
  6. Align incident runbooks with the 72-hour expectations that apply now or may soon: New York’s reporting window for New York hospitals, and the NPRM’s proposed restoration target.

Risk analysis comes first because it recurs in OCR findings. MFA follows because missing MFA opened Change Healthcare and recurs in the threat data. Log correlation comes next because it depends on knowing which systems hold ePHI.

Why Context Decides Which Healthcare Alerts Matter

A sign-in from a new device, a vendor session at an odd hour, or a bulk EHR export can each be legitimate or the start of a breach. Business context tells them apart: what the clinician’s role and approved exceptions allow, what the vendor contract permits, and what the last investigation of a similar event concluded. Teams that bring that context into an investigation early reach a conclusion faster and keep evidence they can show OCR, whether the alert proves benign or a breach.

Daylight is a MASS company, meaning it offers managed agentic security services for Security Operations. Its MDR investigates agreed-upon alerts with telemetry, organizational, and historic context, within the response permissions each customer sets.

Frequently Asked Questions About Cybersecurity in Healthcare

Should We Implement the Proposed 72-Hour Restoration Requirement Before It Is Final?

Yes, where it overlaps with obligations you already have. Building restoration tiers around a criticality analysis supports the contingency plan and disaster recovery specifications already required under 45 CFR 164.308. It also gives you a head start if the final rule keeps the 72-hour target.

Does a Business Associate’s Breach Start Our Notification Clock?

Your clock runs from your own discovery, but if the business associate acts as your agent under federal agency law, its knowledge can count as yours. A business associate must tell you promptly, within 60 days of its own discovery at the latest. In the Change Healthcare response, HHS let Change send individual notices on behalf of providers but left the decision to each covered entity. Write tighter notice windows into your business associate agreements (BAAs); the NPRM proposed 24-hour notice when a business associate activates its contingency plan.

Is Shadow AI a HIPAA Problem or a Policy Problem?

It is both. At Johns Hopkins, a staff member uploaded patient data to an unapproved AI tool with no agreement in place to protect it. Any AI vendor that handles PHI on your behalf is a business associate and needs a BAA, and an AI acceptable use policy should prohibit exposing PHI to public tools.

How Much Does a Healthcare Breach Cost Now?

IBM’s 2026 report puts the average healthcare breach at USD 6.64 million, down from USD 7.42 million in the 2025 edition but still the highest of any industry. Disruption can cost far more. UnitedHealth reported $1.7 billion in direct response costs for the Change attack through the first nine months of 2024.

Do Recognized Security Practices Reduce OCR Penalties?

They can reduce them but do not eliminate them. When setting fines, audit scope, and remedies, HHS must consider whether you had recognized security practices in place for the previous 12 months, under Public Law 116-321. Those include the Health Industry Cybersecurity Practices developed under Section 405(d) of the Cybersecurity Act of 2015. The law creates no safe harbor. Document implementation with dates, because you must show the practices were in place for that full 12-month period.

Subscribe to newsletter

Subscribe to receive the latest blog posts to your inbox every week.

Thanks for your interest.
Oops! Something went wrong while submitting the form.
By subscribing you agree to our Privacy Policy.
Two pinkish translucent bubbles of different sizes on a black background.Two pinkish translucent bubbles of different sizes on a black background.Two pinkish translucent bubbles of different sizes on a black background.
Illustration of a mythical hippocampus creature with the upper body of a horse and the lower body of a fish with scales and a curled tail.Illustration of coral colony with tube-shaped and frilled coral structures in shades of pink and red on a transparent background.

For expert judgment, scaled to your world, head for Daylight