Cybersecurity incidents are increasingly disrupting clinical operations, consuming organizational resources and affecting patient care delivery.

This fifth annual report was conducted to determine the healthcare industry’s effectiveness in reducing human-targeted cybersecurity risks and disruptions to patient care. With sponsorship from Proofpoint, Ponemon Institute surveyed 705 IT and IT security practitioners in healthcare organizations who are responsible for participating in such cybersecurity strategies as setting IT cybersecurity priorities, managing budgets and selecting vendors and contractors.

According to the research, 97 percent of organizations surveyed experienced at least one cyberattack in the past 12 months. For organizations in that group, the average number of cyberattacks was 51, an 8-point increase from 43 in 2025.

The cyberattacks analyzed that took place over a two-year period in this research are cloud/account compromises, supply chain attacks, ransomware and business email compromise (BEC)/spoofing/impersonation. Among the organizations that experienced the four types of cyberattacks, an average of 74 percent reported disruption to patient care, a 2-point increase from 72 percent in 2025.

Cyberattack costs and consequences

Cyberattacks are increasing in frequency, cost and operational impact. We asked the 97 percent of respondents that had at least one cyberattack to estimate the single most expensive cyberattack experienced in the past 12 months from a range of less than $10,000 to more than $25 million. Based on the responses, the average total cost for the most expensive cyberattack was $4.2 million, an increase from $3.9 million in 2025. This includes all direct cash outlays, direct labor expenditures, indirect labor costs, overhead costs and lost business opportunities.

Healthcare organizations report that the greatest impact comes from operational disruption, downtime and the effort required to restore patient services. These findings underscore the growing importance of cyber resilience and organizations’ ability to maintain and restore clinical operations following a cyberattack.

The following is a breakdown of the five cybersecurity cost categories for the single most expensive cyberattack as well as their average cost:

• Disruption to normal healthcare operations cost an average of $1,387,518 in 2026, an increase from $1,210,172 in 2025. Costs stemmed from system availability problems, which continue to be the most expensive consequence from the cyberattack.

• Users’ idle time and lost productivity cost an average of $1,051,150 in 2026, an increase from $858,832 in 2025. These costs were due to downtime or system performance delays.

• The cost of the time required to ensure the impact on patient care increased to $714,782 in 2026 from $702,680 in 2025.

• The damage or theft of IT assets and infrastructure is $546,598 in 2026, a decrease from $624,605 in 2025.

• Remediation and technical support activities, including forensic investigations, incident response activities, help desk and delivery of services to patients is $504,522 in 2026, a slight decrease from $507,491 in 2025.

 AI adoption is outpacing governance

Healthcare organizations are rapidly adopting AI technologies, but governance, visibility and security controls are lagging behind.

This year’s research takes a deeper look into the impact AI is having on cybersecurity in healthcare. As organizations increasingly adopt AI to improve operational efficiency, support patient care and enhance cybersecurity programs, they also are introducing new risks related to data protection, governance, visibility and third-party dependencies.

AI has become or will become a part of patient care and prevention of cyberattacks in most healthcare organizations. Forty-five percent of respondents say their organizations have embedded AI in cybersecurity (24 percent) or embedded in both cybersecurity and patient care (21 percent). Thirty percent plan to adopt AI in the future. Twenty-five percent say there are no plans to adopt AI.

Of the 45 percent of respondents currently using AI, the tasks and applications most often used are administrative automation (63 percent), medical imaging and diagnostics (62 percent) and virtual nursing/patient monitoring (48 percent).

The AI-specific risk of greatest concern is supply chain risk from AI vendors (40 percent of respondents).  Other risks of most concern are data leakage via prompts (36 percent) and AI-generated phishing/social engineering (34 percent). Only 14 percent of respondents say inadequate privacy protections are a serious risk and only 11 percent say non-compliant tools are a serious risk. 

Only 42 percent of respondents say their organizations use security measures to protect AI systems, agents and generative AI usage. The measures most often used are AI data governance and access controls (50 percent of respondents), AI data protection and data loss prevention (DLP) (45 percent of respondents) and AI usage discovery and visibility controls (33 percent of respondents).

 Most organizations (60 percent) do not have controls in place to govern sensitive data used in AI systems. Of the 40 percent that have controls, strong encryption and data access controls for AI training is the number one control (52 percent) followed by professionals monitoring all AI outputs to prevent mistakes and access controls (both 46 percent).

 Interoperability issues among AI technologies continue to be the top challenge to the effectiveness of AI-based security technologies (32 percent of respondents). However, errors and inaccuracies in AI decision rules increased significantly from 21 percent of respondents in 2025 to 30 percent in 2026. Only 29 percent of respondents say the lack of mature and/or stable AI technologies is a challenge.

 Seventy-four percent of organizations have either full visibility into employees’ use of unsanctioned AI tools (32 percent) or partial (42 percent). Forty-six percent of respondents use generative AI tools approved or unapproved with governance (23 percent) or without governance (23 percent).

 Securing clinical operations in the cloud

 For the second year, the research includes healthcare organizations’ plans to secure clinical operations in the cloud.

 More organizations are moving clinical applications to the cloud. In this year’s research, 35 percent of respondents say their organizations have moved clinical applications to the cloud, an increase from 30 percent in 2025. In 2026, 61 percent of respondents say their organizations are very or highly effective in securing clinical applications in the cloud, a significant increase from 54 percent in 2025.

Forty-four percent say their organizations will move clinical applications to the cloud in the next six months (12 percent) within the next year (9 percent), in the next one to two years (14 percent) or eventually (9 percent). This accelerating shift toward cloud-hosted clinical systems underscores the urgency of addressing cloud/account compromise risks, given the potential impact on patient care and service continuity.

The cyberattack landscape

The report analyzes four types of cyberattacks that occurred over the past two years and their impact on healthcare organizations, patient safety and patient care delivery.

For the fifth consecutive year, frequent attacks against the cloud make it the top cybersecurity threat and concern. Seventy-six percent say their organizations have experienced an average of 20 cloud/account compromises in 2026, an increase from 72 percent in 2025.

Cloud/account compromises remain healthcare’s most significant cybersecurity concern, an increase from 49 percent of respondents in 2025 to 54 percent in 2026.

Supply chain attacks increase. Fifty percent of respondents say their organizations experienced an attack against their supply chain, a significant increase from 44 percent in 2025. Of these organizations, on average they experienced four successful supply chain attacks in the past two years.

Fewer organizations are paying a ransom (31 percent vs. 33 percent in 2025) and the average amount decreased slightly. The average costliest ransom paid was $1.1 million, down from $1.2 million in 2025. In the past two years, organizations that had ransomware attacks (63 percent) experienced an average of 5 such attacks.

Awareness about BEC/spoofing/impersonation attacks is growing. Concerns about these attacks have increased significantly from 40 percent in 2025 to 48 percent in 2026. Sixty-five percent of respondents say their organizations experienced BEC attacks, up from 62 percent in 2025 and had an average of four attacks in the past two years.

From breach to bedside: the persistent link between cyberattacks and patient safety

The impact of cyberattacks increasingly extends beyond technology systems and into clinical operations, care delivery and patient outcomes.

As in the previous report, an important part of the research is the connection between cyberattacks and patient safety. Among the organizations that experienced the four types of cyberattacks in the study, an average of 74 percent report disruption to patient care, a slight increase from 72 percent in 2025.

To read the rest of this report, visit Proofpoint’s website.

Even in the age of AI hacking, the devil you know might still be the most serious threat. Depending on which data you pick, malicious insiders are responsible for between one-third and one-half of all hacks.

This is true in the world of scams, too. While international crime gangs get the headlines in scams, much elder fraud is committed by family members, neighbors, or live-in help.

I’ve long been fascinated by small-town crimes; I think they deserve more attention. After all, the vast majority of us will never have occasion to meet a Bernie Madoff. A bad actor with a nice office in your charming downtown?  That’s more likely, often because that nice man in the nice office might have banked decades’ worth of goodwill. Just like an employee who’s been onboarded.

I recently worked on a podcast about Burt Marshall, a man who financially terrorized a small town in upstate New York for decades.  A tax preparer and insurance salesman turned investment “guru,” Marshall took money from seemingly everyone — and everything — in town.  And when his Madoff-like Ponzi collapsed, even the fire department and local cemetery were victims.  You can hear the whole story at AARP’s Perfect Scam podcast, which I host.  

Marshall knew everyone in Hamilton, New York, a college town that hosts Colgate University. It seemed like he did everyone’s taxes in town. That gave him a privileged position. He knew everyone’s income, and he even knew how big their tax returns were.  “Why not invest that money with me?” Burt would suggest. He “guaranteed” investors an 8% annual return, backed by his ever-growing portfolio of rental properties stuffed with Colgate students. The returns were unsustainable; he could only pay when he got new investors, and when Covid hit, the cracks in Marshall’s empire were slowly revealed. By the time it came crashing down, Marshall had collected $90 million from nearly 1,000 local residents. After he pled guilty, they got back about 5 cents on the dollar.

Meanwhile, as you’ll hear in the episode, some victims had everything stolen from them — their retirement cash, their homes, their futures.  Even the local cemetery was nearly abandoned, showing Marshall’s crime cut so deep that victims couldn’t escape the consequences — even in death.

Burt Marshall isn’t unique; he’s far too common, really.  Last year, The Perfect Scam brought you the terrible story of Kota Youngblood. He wormed his way into a youth hockey community in Austin, Texas. After months playing the big man, buying team dinners, telling tall tales about his dinosaur head collection and his time in the Special Forces, Youngblood eventually convinced a set of families that they were at risk of execution by dangerous crime gangs. One family moved clear across the country based on his terrifying tall tales.  Others paid Youngblood for protection.  He was eventually exposed and convicted of wire fraud.

Both stories reveal the power of local connections. And they demonstrate the long game some criminals play.  Normal human beings couldn’t imagine keeping up a farce like that for a few hours, let alone several years. Often, criminals have nothing but time — at least until their time runs out.

My opinion is that these stories reveal something else about our human nature. Hope is the most lovely and vital of human characteristics.  But hope is often exploited by criminals.  The romance scam criminal who says over and over, “I’ll meet you in person in two weeks!”  The “financial manager” who promises you’ll get your money back next month.  Heck, the tech company that says its product will be ready next quarter.  All these people are counting on our natural, and wonderful, tendency to believe in the future.  Politicians have been living off promises and false hope since the dawn of — well, politics.  I would never want to tell any of you to abandon hope.  But consider this: If you lend $100, or a tool from the garage,  to a friend who promises you’ll get it back next week… and you hear “next week” for months and months…at what point do you just decide the money or the tool is gone?  That mental experiment might help you build some healthy guardrails around the promises you hear and the hope you give.

I hope you’ll listen to our two-part series, Bad Neighbor, at The Perfect Scam.