Among Internet of Things devices used in medical settings, Internet Protocol (IP) cameras are the most vulnerable to hackers. Meanwhile nurse call systems hold that troubling distinction among general medical devices.

That’s according to an analysis of data monitored via a commercial platform marketed by Armis, an asset visibility and data security company based in San Francisco.

Risky behaviors at scale. Analyzing inputs from 3 billion gadgets in its asset intelligence and security platform for healthcare, Armis labeled as the “riskiest” those that have the highest percentage with unpatched Common Vulnerabilities and Exposures (CVEs). Key results as released by the company April 17:

Vulnerable Internet of Things (IoT) devices:  

1. More than half of IP cameras turn out to have unpatched CVEs (56) and unpatched CVEs (59%) deemed to be of “critical severity.”
2. Printers were next, with 37% having unpatched CVEs and 30% having critical-severity unpatched CVEs.
3. Some 53% of voice over IP (VoIP) devices have unpatched CVEs, although only 2% are of critical severity.

Riskily connected medical devices:

1. Some 39% of nurse call systems have critical-severity unpatched CVEs and almost half (48%) having unpatched CVEs.
2. Infusion pumps are next most at-risk, with 27% having critical severity unpatched CVEs and 30% having unpatched CVEs.
3. Among drug-dispensing systems, only 4% have critical-severity unpatched CVEs, but some 86% have unpatched CVEs. What’s more, 32% run on unsupported Windows versions.

Confounding connections. Not only are medical and nonmedical appliances increasingly connected, but they’re also increasingly tapped to automatically feed patient data into EHRs, the company points out.

Growth to watch for. Armis projects that, by 2026, smart hospitals will deploy more 7 million Internet of medical things (IoMT) devices. That would be twice the 2021 tally.  

The overarching challenge is balancing risks with rewards. Mohammad Waqas, principal solutions architect for healthcare at Armis:

“Advances in technology are essential to improve the speed and quality of care delivery as [U.S. healthcare] is challenged with a shortage of care providers, but with increasingly connected care comes a bigger attack surface.”

The full Armis news item is posted here.

• Large language models have been game-changing AI ever since OpenAI unleashed its initial version of ChatGPT. Can you believe that was less than five months ago? Anyway, in healthcare, this and other forms of generative AI can be “hyper-personalized,” which makes the technology especially intriguing. So suggested Erik Barnett of Los Angeles-based IT consultancy Avanade at HIMSS23 in Chicago April 17. Barnett remarked that generative AI could serve as a “co-pilot,” helping healthcare navigate business and clinical problems while also bettering the care experience for patients (or passengers, to advance the analogy). HIMSS’s news team has more on the talk here, and the organization’s Healthcare IT News is reporting from the conference till its close on Friday.  
   
• Google Cloud is laying advance access to Med-PaLM 2 at the feet of select healthcare customers. Med-PaLM 2 is the tech giant’s own large language model for medicine. The company is allowing the invitees to test-drive the nascent technology, explore its potential use cases and offer feedback on the user experience. Google says its aim is to see how Med-PaLM 2 might be used to “facilitate rich, informative discussions, answer complex medical questions, and find insights in complicated and unstructured medical texts.” Full announcement here.  
   
• Philips and Amazon Web Services are putting heads together over generative AI. The partnership, which is not new but expanded, includes moving Philips’s picture archiving and communication system (PACS) to AWS. It also will afford Philips access to Amazon Bedrock for developing AI-based applications more quickly than if the company were to build foundation model AI from scratch. Full announcement from Philips here.
   
• An AI model has shown impressive prognostic chops for gauging the aggressiveness of colorectal cancers. Developed at Harvard Medical School and National Cheng Kung University in Taiwan, the system needs only histopathology images to guide therapy decisions and estimate the likelihood of tumor recurrence after treatment. The researchers say their model is meant to aid and not replace human pathologists. Harvard news item here, journal study here.
   
• As a practicing internist, Marc Siegel, MD, has a big problem with healthcare payers tempted to sneakily use AI for performing prior authorizations. “I can envision a future where fights for insurance coverage become even more escalated than they are already—and where personalized medicine is replaced by algorithms,” the media maven/popular doctor writes in an opinion piece published April 17 in USA Today. “What’s to stop insurance companies from replacing me with a cheaper, more predictable AI robot who practices some of the science but none of the art of medicine?” Read the whole thing.
   
• AI startup integrate.ai (Toronto) has taken a central role in the Canadian Platform for Genomics and Precision Health (CP4GPH). The software platform’s mission includes using AI to wring practical patient-care insights from mass stores of scattered data. Integrate.ai founder and CEO Steve Irvine: “Canada’s biotech industry has made a considerable impact on the global stage, and this project will put Canada at the forefront of precision health—driving innovation, enabling job creation and cultivating high-value research and discovery.” Full announcement here.

• Digital Magazine: This is Enterprise Imaging - In this digital magazine we talk about how moving from multiple PACS to a single enterprise imaging system is busting siloes and deepening integration; challenges in radiology imaging and how radiologists are getting more done—better and faster—by using enterprise imaging; skyrocketing image volume and an increased need for collaboration across multiple and geographically diverse sites has made image management far more complex and why cloud is a solution to this; our latest addition to Sectra Enterprise Imaging portfolio—ophthalmology and why it is a game-changer for ophthalmologists.

• Beyond the impression: How AI-driven clinical intelligence transforms the radiology experience - In this session, Nuance CMIO Sheela Agarwal, MD, and Senior Product Manager Luanne D’Antoni explore innovations in radiology report creation and the role of automated impression generation.

• AI quality assurance models saving lives and millions in avoided med-mal - Unrecognized imaging findings are an unfortunate, but undeniable, part of radiology. New advancements in artificial intelligence (AI) and machine learning offer a critical safety net that is improving care and saving lives — as well as avoiding millions of dollars in potential medical malpractice costs.