Google launches Co-Scientist, a Gemini 2.0-powered research LLM that researchers say is like having an expert collaborator. Users provide a plain language research goal, after which the tool provides a summary of published literature and cited suggestions for new hypotheses and possible experiments.
Healthcare accreditor URAC announces plans to release a healthcare AI accreditation program later this year.
CVS Health redesigns its app with AI-driven features to provide a personalized health concierge experience. The app allows users to manage prescriptions for their entire family across CVS Pharmacy, CVS Caremark mail order, and CVS Specialty pharmacies. It also allows them to open locked display cabinets and pick up prescriptions using a barcode. Future enhancements include conversational AI for checking refill status and tailored recommendations for chronic condition management.
Business
Innovaccer launches telephone-based AI voice agents for patient scheduling, protocol intake, referral, authorization, care gap closure, HCC coding, and patient access.
HP acquires wearable device maker Humane AI for $116 million, gaining its employees and software while discontinuing the device. The widely hyped product, criticized as a solution in search of a problem, will be retired. The year-old, $699 Humane AI Pin — which also required a $24 monthly subscription – will become dysfunctional next week when its cloud service shuts down. The startup had raised $230 million and sought a $1 billion buyout.
Ambient documentation vendor Abridge raises $250 million an a Series D funding round. The company was previously valued at $2.5 billion.
Crunchbase lists five healthcare-related companies whose new fundraising rounds suggest a Unicorn Club valuation of at least $1 billion:
Neko Health (body scanner), $1.8 billion.
Hippocratic AI (AI agents), $1.6 billion.
Aragen Life Sciences (drug discovery), $1.4 billion.
Truveta (healthcare data from its provider owners), $1 billion.
Cera (digital-first home health), $1 billion.
Research
A UK-based company develops an AI-powered “super test” for prostate cancer screening that offers greater accuracy and sensitivity than traditional tests like PSA. The multi-omics test uses AI to analyze blood and urine samples for the presence of specific genes and proteins that have been clinically associated with the disease.
University of Michigan and other organizations are using a federal grant of up to $25 million to develop an AI-equipped van that can help medical generalists deliver hospital-level services in rural areas.
Nine hospitals in the UK are using AI to prioritize high-risk patients who are stuck in the NHS’s 7.5 million-long waiting list. The C2-Ai system has flagged 1,000 patients who are at risk of deterioration during their wait to be seen, which reduced surgery-related complications and shortened inpatient stays by four days.
Mass General Brigham and IBM partner through the IBM Sustainability Accelerator to develop an AI tool that will predict extreme heat events, identify at-risk patients, and send warnings to those who are at risk. Previous research indicates that heat event days will cause 235,000 ED visits, 56,000 hospital admissions, and $1 billion in costs, much of that in urban areas.
OpenAI CEO Sam Altman predicts in a blog post that AI will soon reach human-level problem-solving ability (AGI), a transformation that will be comparable to the invention of the transistor. Key points:
AI intelligence scales predictably with investment since it correlates with the log of resources used.
AI costs drop 10x every 12 months, far outpacing Moore’s law.
Exponential AI investment will persist due to its super-exponential socioeconomic value.
AI agents will handle most tasks of mid-level professionals but won’t generate big ideas, will require human oversight, and will still fail in some areas.
AGI’s greatest impact will likely be in science.
Former FDA Commissioner Scott Gottlieb, MD urges the FDA not to classify AI as a medical device if it simply synthesizes and presents information, warning that excessive regulation could hinder AI integration into EHRs and limit its ability to generate clinical insights. He emphasizes that AI is most effective when it is embedded in physician workflows with access to EHR data.
Business
Australia-based Harrison.ai, which offers AI diagnostic solutions for radiology and pathology imaging, raises a $112 million Series C funding round. The company will use the proceeds to expand its US operations.
Cedars-Sinai nurses are piloting an AI tool that was developed by Aiva Health, which was developed through the hospital’s accelerator program. Nurses dictate into the nurse assistant app, which then populates Epic fields upon their approval.
Research
An American Medical Association survey finds that two-thirds of physicians recognize the benefit of AI. Use of AI in practice doubled in one year to 66%, while 57% see administrative burden as its biggest opportunity.
Other
A student-run media site profiles how Brown University Health physicians are using AI, including simplifying the language of surgical consent forms and using OpenAI’s Voice Engine to provide patients who have lost their voice with a text-to-speech tool that sounds like them.
Harvard Medical School psychiatrist and bioethics professor Rebecca Brendel, MD, JD ponders whether AI should be used in end-of-life decision-making, raising these issues:
Clinicians will respect the wishes of a patient as long as they are competent to make decisions and don’t want something that is medically contraindicated.
AI might help after the initial diagnosis if decisions will be made between treatment and palliation.
It might also help in situations where patients are incapacitated, have few relationships, and have avoided treatment.
It’s not enough to decide from the odds of survival alone since people who have experienced a sudden tragedy wouldn’t necessarily have the same wishes as someone who has battled a chronic illness over time.
Decisions of high consequence should always be made by humans.
Information asymmetry has always caused healthcare professionals to be held in high esteem, but AI might outperform humans in some aspects, which would shift professional emphasis to how they use information and interact with patients.
February 5, 2025Healthcare AI NewsComments Off on Healthcare AI News 2/5/25
News
China-based TikTok parent ByteDance releases OmniHuman-1, an AI tool that transforms a single photo into a lifelike video with natural singing and speech. The technology is raising concerns about more advanced and easily generated deepfakes.
Biomedical researcher Derya Unutmaz, MD uses ChatGPT’s new Deep Research tool to review two cancer cases. He reports, “Both reports were simply impeccable, like something only a specialist MD could write. There’s a reason I said this is a game-changer.” He has previously said that AI will be a standard, required tool in medical practice within two years, and electing not to use it might be considered malpractice. Deep Research is available to ChatGPT Pro subscribers and will be rolled out eventually to Plus subscribers. Users of X say that a key will be if OpenAI can figure out how to provide access to paywalled medical journals.
Britain’s NHS launches a 700,000-patient trial of using AI to screen for breast cancer. The Royal College of Radiologists has urged the government to embrace healthcare AI to help offset a clinician shortage.
Business
Beacon Health System will implement AI-driven medical necessity reviews using the generative AI tool of Xsolis. The health system began using company’s AI to manage people with chronic illness in May 2019.
ScribeAI launches an ambient documentation solution. Co-founder and board chair Kyle Robertson remains CEO of Zealthy. He was fired as CEO of behavioral telehealth startup Cerebral in May 2022 due to federal scrutiny of the company’s controlled substances prescribing practices, after which the company paid $3.6 million in November 2024 to settle charges of encouraging the unauthorized use of controlled substances.
OpenAI founder and CEO Sam Altman says that the company is considering developing a healthcare-focused version of ChatGPT that would comply with HIPAA.
Research
A JAMA Network discussion of using a secure instance of ChatGPT to draft clinician responses to patient messages makes these points:
Patients preferred the AI-generated messages.
A key factor was that the AI messages were more detailed than the quick responses that clinicians wrote in their available time.
The authors note that appointment lead times are so long that patients will be using AI tools to ask medical questions, even in specialty areas that are not underserved.
The authors say that healthcare, like other industries, will need to decide between improving clinician working conditions based on freed-up time versus just increasing their workload.
University of Cincinnati associate professor Dong-Gil Ko, PhD proposes a telehealth billing model that rewards experienced, efficient physicians instead of using time-based metrics that pay all providers equally. Ohio’s billing system compensates clinicians based on time spent. He cautions that a downside of the proposed model is that patients may avoid seeking virtual care due to uncertain costs.
Researchers find that ChatGPT effectively de-identifies clinical notes and generates synthetic data, maintaining a balance between data utility and privacy in clinical research.
A study finds that AI-analyzed ECG readings can estimate biological age, potentially linking heart health to cognitive function.
Other
A Health Affairs opinion piece urges the FDA to implement post-market surveillance for healthcare AI applications to detect unintended bias that may only emerge after widespread deployment.
A small Deloitte survey of health system executives finds that 40% report moderate to significant returns on AI investments. Over 80% expect AI to have a substantial impact on their organizations in 2025 but see government regulation as necessary.
A state representative in Oregon submits a bill that would ban the use of “nurse” and similar titles to refer to AI agents.
A woman turns to ChatGPT when doctors are stumped by the cause of fevers and rashes in her five-month-old son. ChatGPT correctly diagnosed him with Kawasaki disease.
The local TV station highlights BayCare St. Anthony’s Hospital’s AI pilot project using Aiva Health’s documentation tool. The system enables conversational EHR access, remote room control, facility-wide communication, and alerts.
China-based AI startup DeepSeek introduces a low-cost, open-source AI model that rivals leading US technologies, sending US tech stocks sharply down. DeepSeek, which is free to use and can be run locally on modest hardware, has prompted discussions about US technology export controls, the contrast in cost in AI development and training. DeepSeek was launched in July 2023 by a billionaire hedge fund operator who previously worked with using AI in investing. He stockpiled some lower-power Nvidia chips which were eventually banned from US export to China.
Meanwhile, Chinese technology company Alibaba releases a new version of its own LLM that it says outperforms DeepSeek-V3 and ChatGPT-4o.
OpenAI launches a research preview of Operator, an AI agent that can analyze a webpage and interact with it to perform tasks that involve typing, clicking, and scrolling.
A Health Affairs commentary piece – published as part of the National Academy of Medicine’s Vital Directions for Health and Health Care: Priorities for 2025 initiative — calls for the incoming administration to focus on four health AI-related priorities:
Ensure safe, effective, and trustworthy AI use.
Develop an AI-competent workforce.
Invest in AI research to support the science, practice, and delivery of health and healthcare.
Promote policies and procedures to clarify AI liability and responsibilities.
New Vatican ethical guidelines call for using AI in healthcare to enhance, rather than replace, the relationship between provider and patient.
Stat reports that the White House has indefinitely cancelled meetings of the Health Information Technology Advisory Committee and its workgroups, which operate under HHS’s ASTP/ONC. The committee, which was established under the 21st Century Cures Act, makes recommendations to ASTP on a variety of technology issues, including AI and interoperability.
Mayo Clinic President and CEO Gianrico Farrugia, MD tells the World Economic Forum that his organization has access to 320 algorithms, but their biggest challenge is that today’s technology that can’t support the tools. He says that he would not want to receive specialty care without the clinician using AI, which will require a new architecture to deliver at scale. The panel discussion was titled “Healthcare Innovation at Davos 2025: Cracking the Code of Digital Health.” The panel’s key points involved interoperability, platform-based models, health worker AI training, data liquidity and sharing, and building trust.
Also at the World Economic Forum, the CEO of AI company Anthropic predicts that AI could double human lifespan within five to 10 years. A skeptical response might be:
Attaining that goal would require the US lifespan to increase to 155 years almost immediately.
Even AI-enhanced drug research is unlikely to make a difference of that magnitude given the need to conduct clinical trials, earn FDA approval, and figure out the economics that would be required to make a solution universally available.
Raising the average lifespan would require applying the principles to nearly every American, which has never happened.
His statement is more aspirational than based on biology and medical research.
He also says that AI systems will outperform humans in nearly everything by 2026 or 2027.
His company sells AI.
Spain-based AI medical imaging company Quibim raises $50 million in a Series A funding round. The company hopes to use AI to extract information from medical imaging to identify phenotypes that can predict outcomes, an advanced imaging field known as radiomics.
AI startup Retro Biosciences is reportedly raising a $1 billion funding round. OpenAI CEO Sam Altman provided the company’s initial funding of $180 million and will participate in the latest round. The company’s goal is to add 10 years to the healthy human lifespan by using AI to target and reengineer the cellular drivers of aging.
US Rep.David Schweikert (R-AZ) introduces a bill that would allow FDA-approved AI tools to qualify as a medical practitioner, including the ability to prescribe drugs.
Business
Paragon Health IT will divest its consulting business, change the parent company’s name to Strings, and refocus on AI-driven healthcare workload management system.
AI-powered cancer drug discovery startup Manas AI launches with a $25 million seed funding round. The co-founders are LinkedIn co-founder and venture capitalist Reid Hoffman and oncologist and Pulitzer-winning author Siddhartha Mukherjee, DPhil, MD,
Equality AI CEO Maia Hightower, MD, MPH, MBA announces that the company, which focused on trustworthy AI in healthcare, has closed after four years following the White House’s rollback of a previous executive order that called for trustworthy AI development.
Research
Researchers in Scotland are developing AI software that can scan someone’s retina to detect early signs of dementia, which could be used during routine eye exams. The team explains, “The retina holds a whole wealth of information and is a biological barometer of our brain health … Something very simple like a photograph of a retina can now be harnessed to potentially predict brain change later on in life.” They hope to roll the technology out to opticians in 2026.
Other
A Twitter user finds a DeepSeek-generated dataset of 143,000 made-up encounter conversations on the Hugging Face AI website. Each item contains the disease, language, scenario, patient-provider conversation, common errors in diagnosis, differential diagnoses, related diseases, and a summary, all in multiple languages. It was developed for AI training and can be licensed from MIT. Leading the project is endocrinologist Johnson Thomas, MD.
January 22, 2025Healthcare AI NewsComments Off on Healthcare AI News 1/22/25
News
Shares of AI infrastructure companies rise sharply on the White House’s announcement of Stargate, a private sector project in which major technology companies — including OpenAI, SoftBank, and Oracle — will invest up to $500 billion to build AI data centers. Oracle CTO Larry Ellison, who was present during the announcement, said that the “most charismatic” application of AI would involve EHRs, while OpenAI CEO Sam Altman predicted that “we will see diseases get cured at an unprecedented rate.” The project’s key technology partners were announced as Arm, Microsoft, Nvidia, Oracle, and OpenAI. An OpenAI blog post says that SoftBank will run the financial side of the project and OpenAI will lead operations.
Oracle’s Larry Ellison also predicts at the Stargate announcement that AI will be used to create personalized MRNA cancer vaccines. He says that AI can diagnose cancer from a blood test, sequence the tumor’s genes, and then create a vaccine, all within two days. Shares of MRNA vaccine maker Moderna jumped the news coverage.
The White House revokes a previous executive order that required AI developers who create tools for certain users, including public health, to share product safety results with the federal government.
The UK’s NHS will establish a $250 million fund to help trusts invest an diagnostic and predictive AI tools, with an initial focus on medical image analysis and pathology.
Mayo Clinic and Cerebras Systems will collaborate to develop LLMs that can review the medical records of patients to determine how they will respond to treatments based on their genetic makeup. The models will be trained on the genomic data of 100,000 Mayo patients. Cerebras sells AI hardware, software, and AI model creation services.
Business
Precision medicine technology company Tempus AI launches a personal health concierge app that collects a patient’s medical records and applies AI to generate insights. The company says that its Olivia app can connect directly to the EHRs of 1,000 health systems, sync with health devices, and accept manual uploads. Outputs include a profile summary, summaries of physician notes, medical image sharing. and clinical trials matching.
Google DeepMind’s CEO says that the first drugs that were developed with the help of AI will start clinical trials this year.
Other
Hofstra’s medical school is using AI to make the first cut of the applications it receives, which it says is not only efficient, but also objective and consistent. Several other medical schools are using AI similarly to choose a few dozen students from thousands of applications. George Washington University’s medical school says it spends 6,000 hours of faculty time each year performing manual screening of applications.
The CEO of drug maker Sanofi says the company is using AI to decide if a given drug under development should advance to the next phase. He explains, “We’re not used to having somebody without a career at stake in the room at a senior level.”
Unionized nurses from the Albany, NY VA hospital and other hospitals nationwide rally to voice concerns about hospitals prioritizing AI investments over staff recruitment and retention. The union, which organized the January 16 national protest, has called for stricter AI regulation and greater nurse involvement. A member survey revealed that nurses believe AI often compromises patient safety by undermining their clinical judgment. One nurse highlighted concerns about discharging a post-operative patient with only a tablet for AI-guided updates, saying, “You need eyes on a patient.” The VA responded by emphasizing improved performance metrics and assured that AI would be used safely and responsibly to support, not replace, care teams. National Nurses United, an AFL-CIO affiliate, has issued guidelines for AI use in patient care. Over 100,000 of its members are entering contract negotiations.
January 15, 2025Healthcare AI NewsComments Off on Healthcare AI News 1/15/25
News
OpenAI launches Tasks, a beta feature for ChatGPT that allows users to instruct the LLM to execute tasks or reminders at specified times in a Siri-like way. Example: “Every Friday, provide a list of weekend activities based on my location and the forecasted weather.” The rollout of Tasks is the first step of OpenAI’s AI agent strategy, where ChatGPT can go beyond answering questions to actually performing actions selectively.
A survey by Gwynedd Mercy University of American adults finds that 59% expect AI to improve outcomes, 57% believe it will reduce costs, and 77% believe it can reduce healthcare disparities in the next five years.
The Peterson Health Technology Institute launches an AI task force that will assess AI’s current and potential impact on cost and efficiency, the uptake of AI scribe systems, and the need for metrics and evidence.
UK Prime Minister Keir Starmer says that AI analysis of NHS patient records can speed development of treatments and tests, saying that he doesn’t think a “defensive stance” can deliver possible breakthroughs. He cited diagnostic imaging AI and models that identify patients who are likely to miss appointments.
Business
Microsoft’s venture fund invests in RAAPID, which offers an AI-powered risk adjustment coding platform. Founder and CEO Chetan Parikh, MS previously held executive roles at EzDI and Mediscribes.
Mayo Clinic launches Mayo Clinic Digital Pathology, with assistance from Nvidia and Aignostics. The system was trained on 20 million digital slides and 10 million patient records from Mayo. Mayo apparently intends to commercialize it given its announcement at the J.P. Morgan Health Conference.
The American Hospital Association publishes “Building an Implementing an Artificial Intelligence Action Plan for Health Care.” It lists the top use cases that can generate a quick ROI for health systems: predicting claims denial, optimizing OR time, managing supply chain data, and streamlining discharge planning.
Amazon Web Services and investment firm General Catalyst will collaborate to develop and deploy AI solutions. General Catalyst’s portfolio includes Summa Health acquirer Health Assurance Transformation Company, Commure, and Aidoc.
Hippocratic AI launches an AI agent app store for clinician creators, who have developed 300 agents covering 25 specialties in the past few weeks.
UK-based digital-first home healthcare provider Cera raises $150 million in financing, valuing the company at over $1 billion. The company reduces hospitalization through AI-powered fall prediction, vital signs monitoring, and next best action.
Research
Researchers from Penn’s medical school use AI to analyze the EHR data of long COVID patients and identify those who may need care from specific departments and teams.
Other
The Military Health System recaps the work that it completed in 2024 to prepare for AI: inventorying use cases, enhancing collaboration with other government agencies, creating a digital health strategy, and completing an exercise to identify AI’s vulnerabilities, risk, and biases in a military medicine setting.
California’s attorney general reminds healthcare users and developers that AI must be tested, validated, and audited to ensure that its use is safe, ethical, lawful, free of bias, and transparent with regard to system training and provider use in making decisions.
Two Tenet hospitals delete a job fair announcement when people notice that the depicted employees are sporting extra fingers and unrealistic hair due to what appears to be an “AI fail.”
FDA issues draft guidance for supporting development and marketing of AI-enabled devices throughout the Total Product Life Cycle. It also publishes draft guidance for the use of AI to support development of drugs and biologicals.
Women’s health smart ring maker Movano announces EvieAI, a chatbot that was trained on peer-review medical journals. The company said in its CES announcement that the chatbot’s accuracy is 99%.
Law professors suggest that the Supreme Court’s recent overruling of the Chevron Doctrine — which allows federal agencies, rather than courts, to interpret and implement statutes when authorized by Congress — could impact the FDA’s ability to regulate AI, as its approach often relies on non-binding guidance documents and position papers.
Business
Hartford HealthCare implements Aidoc’s AI platform, which includes 17 FDA-cleared algorithms.
UK-based digital pathology vendor Deciphex raises $32 million in a Series C funding round.
Research
Researchers in China develop a health insurance fraud detection model for insurers and auditors.
A study finds that while AI is effective at making a diagnosis when fed exam-style questions, if fares worse when analyzing real-world conversations. The authors make these recommendations for AI developers:
Train and test systems on conversational, open-ended questions as are found in unstructured doctor-patient conversations.
Assess the model’s ability to ask the right questions.
Design models to work across multiple conversations.
Design models that can capture both textual data and images.
Incorporate non-verbal cues such as facial expressions, voice tone, and body language.
Other
A health news site lists ways that North Carolina providers are using AI:
Analyzing lung nodule scans to predict cancer risk (Atrium Health).
Electronic follow-up with patients who have received a hip or knee replacement (OrthoCarolina).
ED scanning of images to detect serious conditions (Novant Health).
AI-drafted responses to patient portal messages (Atrium Health and WakeMed).
Cognitive impairment detection (Wake Forest University School of Medicine).
Flagging patients who are due a follow-up visit or imaging (Wake Forest Baptist).
Early detection of sepsis (Duke Health, UNC Health).
Suicide risk assessment (Novant Health).
Optimizing OR use by predicting the length of surgical procedures (Duke Health).
A popular TV journalist in Israel who lost his voice due to Lou Gehrig’s disease returns to Channel 12 by using AI that was trained on his voice to narrate his stories.
Health authorities in Sudan are hoping that AI can perform some of the work of doctors who have been killed in the country’s civil war.
A bioethicist warns that healthcare AI systems that allow customization could restrict exposure to important information in catering to the user’s preferences and biases. She presents distinctions between systems that are customized for information discovery and those intended for information delivery.
December 18, 2024Healthcare AI NewsComments Off on Healthcare AI News 12/18/24
News
The House Task Force on Artificial Intelligence publishes its findings and recommendations, which include:
Challenges include data availability and quality, incomplete or inaccurate responses, non-individualized recommendations, lack of decision transparency, data privacy, interoperability with existing systems, liability for AI-contributed errors, biased decision making, and deployment for financial gain rather than patient care.
AI can help improve the percentage of drugs that are eventually approved by FDA by streamlining study design, finding study patients, and answering patient questions.
AI can assist radiologists by performing first-pass image screening, cleaning up images, and applying consistent interpretation criteria across all patients and locations.
Clinical decision support can be tailored to an individual patient’s symptoms and medical history and can compare their data to the medical literature or a “patients like this one” model.
AI can use population health data to predict response to initiatives among population groups.
Physician burnout can be reduced by deploying AI-assisted clinical documentation.
AI can reduce the administrative burden of the prior authorization process.
CMS will need to evaluate Medicare payment for algorithms and other AI tools.
AI has potential in medical management, but could create unnecessary denials.
Innovation will require standardization of EHR-specific data formats to be applicable to broader populations and for AI training.
HIPAA may require updating to meet the challenges of provider AI deployment.
Principal Deputy ASTP/National Coordinator Steven Posnack, MS, MHS says that healthcare AI should not be strictly regulated since it is constantly changing and because risk tolerance varies by setting, such as back office support versus patient care versus research.
A technology reporter shares five weeks of her Apple Health information with ChatGPT and finds that with no prompting, it analyzes trends and makes personalized suggestions for improvement, weaving seemingly unrelated statistics into a fitness journey story.
OpenAI makes the ChatGPT search function available to free users (it was previously available only to paying users) and adds voice-integrated search.
ASTP publishes a downloadable inventory of HHS AI use cases, documenting 66% more than a year ago.
Business
Company insiders say that most of the work of EvenUp — which calculates the value of a personal injury case by analyzing customer medical records and case files — is performed by humans rather than the AI that the company touts to investors. Former employees of the startup, which is valued at $1 billion, say that the AI system misses injuries, hallucinates medical conditions, and is not reliable in analyzing doctor visits, leading managers to tell employees not to use it.
Other
Google Cloud’s healthcare solutions director Aashima Gupta, MS predicts the short-term use of AI to be focused on routine tasks such as appointment scheduling and processing forms, clinical documentation, communication, claims processing, marketing outreach, and use of AI agents for member and provider communications. She expects longer-term AI uses to include enhanced access, enhanced imaging screening and early detection, use of AI as a health concierge, multilingual support, medication reminders, and care navigation advice.
The UK’s National Health Service is using AI to identify patients who are at risk for high usage of emergency services, who are then offered coaching and in-home support.
December 11, 2024Healthcare AI NewsComments Off on Healthcare AI News 12/11/24
News
OpenAI launches AI-driven video generator Sora, then disables it when it is overloaded from heavy traffic. It can generate video from text instructions and existing media, but only up to a final length of 20 seconds. The company also releases its Canvas feature to all ChatGPT users, which allows users to work on writing and coding projects with increase editing capability.
Google announces Willow, a new quantum chip that performs a benchmark computation in under five minutes versus the 10 septillion years that is required by the fastest supercomputers.
Business
Corewell Health chooses Abridge for ambient clinical documentation, with a rollout to 4,000 physicians planned.
Research
A Washington State University sociologist finds that rural clinics aren’t ready to use AI to optimize proactive text messages to patients. An AI-polished lung screening message drew fewer responses than the traditional one and more than 50% of all sent messages bounced back as undeliverable, raising the question of how the clinic could contact its patients in an emergency. The local hospital also noted that the area has a high population of Hispanics, which raised its own messaging challenges.
Other
Corpus Christi Medical Center Chief of Surgery Michael Ewing, MD credits his use of MedQuill’s newly launched SurgeOn Scribe – which he helped create — to saving an hour of clinic time each day.
Joe Raetano, the VA’s AI architect, says that the organization is using AI to extract social determinants of health from clinical notes and to manage patient data via an agentic chatbot.
In England, a woman credits AI with saving her life after her lung cancer was diagnosed by Royal Surry NHS Foundation Trust using AI imaging triage tool Annalise.ai.
December 4, 2024Healthcare AI NewsComments Off on Healthcare AI News 12/4/24
News
Non-profit healthcare safety group ECRI names AI as its top healthcare technology hazard for 2025. Organization President and CEO Marcus Schabacker, MD, PhD said in the list’s announcement, “Balancing innovation in AI with privacy and safety will be one of the most difficult, and most defining, endeavors of modern medicine.”
Withings and the government of France launch Project DEEP, which will use AI and non-invasive medical device innovation to detect and prevent cardiometabolic diseases, with a $23 million investment.
Business
AI and precision medicine technology vendor Tempus will work with Northwestern Medicine to explore the use of AI in clinical care and research. Their first project involves cardiology, where the health system has deployed the company’s algorithm that helps physicians identify patients who may have a one-year risk of atrial fibrillation / flutter. The Tempus ECG-AF algorithm received FDA’s 510(k) clearance in June 2024.
Spectral AI completes its analysis of burn center images that will be used to train its AI-powered DeepView System for predicting wound healing outcomes.
Research
New York University Langone Health analyzes use of its private instance of ChatGPT by its employees, of whom 1,000 applied for access in a six-month period. Clinical and research users represented half of those requests, with the most common uses being writing, editing, summarizing, analyzing data, searching for new information, and generating ideas. Examples include creating teaching materials, drafting email responses, generating job descriptions, assessing clinical reasoning documentation, and translating SQL queries. Some users reported that they struggled to create prompts and saw occasional hallucinations.
Other
Politico reports that leading house Republications want HHS to end participation in non-government AI oversight groups, specifically the Coalition for Health AI. The lawmakers said in a letter to HHS that they are concerned about having ASTP’s Micky Tripathi serve as a CHAI board observer, stating that, “help us understand how putting the organization directly in control of market entry for innovative technologies does not represent a significant conflict of interest.”
A Brookings Institution report says that use of AI in healthcare could reduce the annual US budget deficit by 20% while expanding access. The analysts, who point out that every US industry except healthcare has improved productivity in the past 50 years, say that AI could help by automating appointment scheduling, patient flow management, and preliminary data analysis. They also predict that AI could improve preventative care and disease detection.
A developer creates They See Your Photos, which allows anyone to upload a photo to see the private information Google can glean from it. I sent it a HIMSS conference photo that it analyzed well, including a note that “many people are looking down at their mobile devices.”
November 20, 2024Healthcare AI NewsComments Off on Healthcare AI News 11/20/24
News
Microsoft enhances Microsoft 365 Copilot with task automation, new agents for Teams meetings and employee self-service, and a Copilot Control System for IT management.
Samsung rolls out One UI 6 Watch features to older models, which adds AI-powered tools such as health recommendations, sleep analysis, and sleep apnea detection.
Vanderbilt University Medical Center’s Department of Biomedical Informatics and InterSystems will work together on biomedical informatics and AI research. VUMC-DBMI will develop FHIR and interoperability training coursework that includes hands-on labs that will use InterSystems products.
BrightHeart earns FDA 510 (k) clearance for its AI-powered analysis of fetal heart ultrasounds.
Washington University School of Medicine and BJC Health System launch the Center for Health AI, which will focus on using AI to personalize patient care.
The National Institutes of Health develops an AI algorithm called TrialGPT that matches patients to clinical trials for which they are eligible.
Business
Stepful – whose AI platform delivers training to working adults for medical jobs such as medical assistant, pharmacy technician, and surgical technician – raises $32 million in a Series B funding round. The company expects to train 30,000 students this year in programs that can be completed in as little as four months for $2,500. It matches its students to one of its 8,000 partner clinics and hospitals for hands-on training.
Research
Cedars-Sinai investigators develop an AI-powered process to automate the classification of patients by the severity of their cancer, which could help get them into clinical trials faster.
Other
The New York Times runs an article titled “Elon Musk Asked People to Upload Their Health Data. X Users Obliged.” Experts warn that posting images to any AI tool is not protected by HIPAA, also noting that X’s terms of service allows the company to share data with related companies.
Meanwhile, Elon Musk responds to a small study in which doctors who used ChatGPT to diagnose test patients using only their case histories performed only slightly better than those who didn’t use it, but ChatGPT by itself outperformed the doctors. Experts say that nobody really knows how doctors think, especially when they use their personal experience or intuition to diagnose patients.
A user of Google’s Gemini AI chatbot posts screenshots a session where it went off the rail when asked a question about households.
November 13, 2024Healthcare AI NewsComments Off on Healthcare AI News 11/13/24
News
Google launches Learn About, a conversational AI learning companion that helps users learn about any topic.
Inovalon founder and CEO Keith Dunleavy, MD donates $6 million to his alma mater, Harvard Medical School, to expand education in AI in healthcare.
Business
AI-powered doc-in-the-box manufacturer Forward, which said at the November 2023 launch of its CarePod that it hoped to deploy 3,200 of the machines in the next year, shuts down. The company had raised $650 million in funding. Former employees said key problems were lack of patient interest, skeptical commercial building landlords, blood draw technology failures that forced the company to stop offering lab tests, and machines that left patients stuck inside. Forward managed to install just five of the devices, which cost $1 million each to build. My analysis of the original announcement wasn’t optimistic.
Maverick Medical AI launches CodePilot, which offers real-time medical coding and MIPS/MACRA compliance notifications.
Apple is preparing to launch its first AI hardware device, an Echo-like wall mounted smart display for homes that will allow users to control apps, use FaceTime as an intercom, play music, and eventually to operate a robotic arm.
Research
Johns Hopkins researchers train a robot by showing it videos of surgical procedures, after which the da Vinci Surgical System robot performed as well as a human doctor in manipulating a needle, lifting tissue, and suturing. The imitation learning involved videos that were recorded on cameras that were attached to da Vinci robots all over the world. The robot even learned behaviors that weren’t contained in the videos, such as picking up a dropped needle.
Researchers use AI-analyzed computer vision to predict neurological changes in NICU babies.
Other
A UCSD hospital neurology ICU nurse who is also the nursing union rep says that he is terrified by “the creep of AI in our hospitals.” He observes:
A billionaire Qualcomm executive funded the hospital’s new construction, a technical connection that he speculates as to why “they dive headfirst into this AI thing.”
The hospital replaced an Epic patient acuity application with an AI-based one that he says “felt like magic, but not in a good way” because it eliminated nurse involvement and didn’t explain its logic.
He says that ambient documentation is like mass surveillance that will be used to “track nurses” as was done with RFID tracking tags.
He concludes that the real goal of applying AI isn’t patient safety, but to increase nurse efficiency and make them “operators of the machines.”
November 6, 2024Healthcare AI NewsComments Off on Healthcare AI News 11/6/24
News
Adventist Health Rideout uses Viz.ai to reduce the time that is required to transfer suspected stroke patients to a hospital that offers specialty care from 202 minutes to 109. The system analyzes the CT scan and alerts the care team if it detects a potential stroke. The data is preliminary, as the hospital has used the technology on just 10 patients since it launched in May 2024.
OpenAI adds web searching capability to ChatGPT. It also collects up-to-date search results from news and data providers and displays them in visual categories such as weather, news, and maps. Results include source links. Unlike Google’s search results, it does not display promoted links, advertisements, or sources that are of questionable quality.
OpenNotes and Abridge partner to research the effectiveness of patient visit summaries that are generated from ambient listening by having those patients evaluate them.
Business
Aignostics, which applies AI to digital pathology to support precision medicine, raises $34 million in a Series B funding round. Mayo Clinic participated in the round and will work with the company to develop foundation models and biopharma product offerings.
Snoop Dog-backed cannabis biotech Oxford Cannabinoid Technologies collaborates with Oxford University to use AI drug discovery to develop non-addictive, cannabinoid-based pain medications. The company also announced that its shares will be de-listed because they have lost 97% of their value, leaving it with a market cap of $3 million.
France-based insurance startup Alan, which allows customers to ask questions to doctors via a chat interface and receive an answer back in 15 minutes, adds a virtual assistant to the feature. The virtual assistant, which is called Mo, rephrases the question and then asks if the person would rather interact with a doctor or Mo. Conversations with Mo are checked within 15 minutes by a doctor who can correct its recommendations. The company says it will enhance the tool to provide personalized guidance based on context and the user’s health history.
Research
The National Cancer Institute awards Pieces Technologies a $2 million grant to develop a conversational AI agent that will allow cancer patients the ability to ask questions about their care. The system, which will also collect social determinants of health information, will be co-developed by MetroHealth.
Penn State researchers find that patients are more satisfied with using a medical AI chatbot if the system remembers their social information and makes small talk about their job and hobbies.
Other
Nvidia’s healthcare VP says that the next wave of AI will turn medical devices into robots. Kimberly Powell predicts that “This physical AI thing is coming where your whole hospital is going to turn into an AI.”
October 23, 2024Healthcare AI NewsComments Off on Healthcare AI News 10/23/24
News
Amazon One Medical launches AI tools for its 1Life proprietary EHR – ambient documentation, a summarized medical history, draft responses to patient messages, and workflow routing.
Anthropic releases a beta developer version of Claude’s API that adds “computer use” capability, in which developers can program interaction with a user’s computer such as looking at a screen, moving the cursor, clicking buttons, and typing text.
Google Cloud announces GA of Vertex AI Search for Healthcare along with new features of its Healthcare Data Engine.
GE HealthCare announces CareIntellect for Oncology, which summarizes clinical reports, flags deviations from a patient’s treatment plan, and identifies relevant clinical trials.
The Coalition for Health AI (CHAI) publishes draft frameworks of how it will certify independent quality assurance labs and standardize their test results into what it compares to a nutrition label for AI product performance and safety.
Aidoc and Nvidia will co-develop a framework for integrating AI into clinical workflows.
Business
Ozarks Healthcare will integrate Avo’s AI clinical decision support with its Meditech EHR. The tools include scribing, chart summary and care guide, and the ability for clinicians to ask clinical questions about the patient’s information and relevant medical evidence.
Zoom will integrate Suki’s ambient documentation into Zoom Workplace for Clinicians, a recently announced paid offering.
Atropos Health announces GA of ChatRWD, an AI co-pilot that generates real-world evidence to answer clinical questions.
Tennr, which offers AI-powered healthcare document processing, raises $37 million in Series B funding.
AI-driven patient preference and informed consent system vendor HealthEx raises $14 million in seed and Series A funding.
Artera announces new AI co-pilots to its Harmony patient communications platform: Staff (translation, predictive text for patient inquiries, message shortening, and conversation summaries that can be saved to the EHR) and Insights (no-show reports).
HCA Healthcare selects Commure to develop and deploy ambient AI. Commure closed its $139 million acquisition of ambient documentation technology vendor Augmedix in July 2024.
Research
Researchers develop a system of AI-powered GoPro cameras that detects when the wrong drug syringe or vial is being used to prepare a patient’s doses, which the authors note could prevent serious medication errors in ORs, ICUs, and EDs.
Other
Harvard Medical School adds a required month-long AI course for students in its MD/PhD translational medicine program and engineering program that it offers with MIT.
October 16, 2024Healthcare AI NewsComments Off on Healthcare AI News 10/16/24
News
Patient engagement startup Parakeet Health launches its generative AI voice platform for health system contact centers and raises $3 million in seed funding.
Luma Health announces AI products for healthcare call centers: a fax processing tool and a patient-facing voice AI concierge.
Business
MIT News profiles alumni-founded Ambience Healthcare, which offers an ambient documentation solution that is being used in 40 large institutions.
Creating and analyzing videos of surgeries to create “highlight reels” for training, collaboration, and support.
Create 3D maps for surgical procedures.
Analyze de-identified EHR data to identify disease targets and drugs.
Support clinical trial recruitment.
Analyze genomic and clinical data from diagnostic tests to target treatments and identify candidates for clinical trials.
Predict drug supply and demand to optimize distribution of products where they are needed most.
Research
A retrospective cohort study of Yale New Haven Health System medical-surgical inpatients finds that the accuracy of six AI-powered early warning tools for patient deterioration varies widely. ECARTv5 and the National Early Warning score were the top performers, while Epic’s Deterioration Index had the lowest positive predictive value. The authors recommend that health systems verify how the tools work and oversee their use.
Other
Angela Elena Olazaran Laureano, a 17-year-old student in rural Mexico, is chosen from 11,000 nominations to win a $100,000 student prize for her work developing an AI-powered virtual medical assistant that provides basic diagnosis for 21 conditions and warns users if they are likely to be contagious. She was previously part of robotics team that won the national Home Care Challenge. She will use the award to create a STEM classroom in her home town.
October 9, 2024Healthcare AI NewsComments Off on Healthcare AI News 10/9/24
News
WVU Medicine implements an AI-powered search system that allows employees to search for content that is contained in internal documents, such as clinical protocols, how-to guides, and tip sheets. It plans to eventually offer searching across all of its sites as well as across all Epic-using hospitals.
A newly signed California bill requires providers to include a specifically formatted disclaimer when AI was used in patient-facing clinical communication that has not been reviewed by a licensed provider. The communication must also contain instructions for the patient to contact an appropriate person.
HHS Assistant Technology for Technology Policy and AI leader Micky Tripathi, PhD, MPP says that HHS is looking at ways to create AI models using its vast data stores, but he cautions that its data is inherently biased since it excludes people who don’t have insurance or who don’t seek care from providers.
Wyoming’s Medicaid program approves coverage of Canvas DX, an online tool that assesses autism using family-provided information and videos. Families can use the tool instead of trying to find one of the few state clinicians that can administer the ADOS test.
The American College of Radiology will participate in the FDA’s program to boost innovation in breakthrough devices. FDA has expanded the program for its cardiology device origins to include neurology, ophthalmology, and radiology.
Business
The authors of a literature review on AI in healthcare suggest that its public health impact will be limited by the “politics of avoidance,” where the US health system’s treatment-focused model ignores broader social determinants of health. They express concern that the hype around AI may overshadow more effective, evidence-based interventions like using community health workers and implementing harm reduction programs, which address root causes rather than just offering suggestions for treating symptoms.
Research
A test of GPT-4’s ability to provide clinical recommendations for ED visits – including admission status, radiology requests, and antibiotic prescriptions – finds that it performs poorly compared to medical residents. Its recommendations were often overly cautious and called for medically unnecessary admissions and orders.
A review of AI-powered, FDA-approved medical devices finds that most did not include race or ethnicity, socioeconomic data, and the ages of study participants in their submitted approval documentation. The authors conclude that the lack of consistency and data transparency may exacerbate health disparities
Other
Drug maker Eli Lilly and Company names Thomas Fuchs, DrSc as its inaugural chief AI officer, where he will be tasked with leading the company’s AI initiatives in drug discovery, clinical trials, and manufacturing. He was previously chair of the AI program at Icahn School of Medicine at Mount Sinai.
Mayo Clinic will use a $25 million donation to provide funding and work time for its clinicians to purse AI-related projects.
Four high-profile cancer centers use $40 million in tech company funding to form the Cancer AI Alliance, which will study their collective patient data to find clinical insights.
US Army researchers develop SeptiBurnAlert, which uses AI to analyze the blood samples of burn patients to detect sepsis-related components. The system predicts risk within the first 24 hours of hospitalization with high accuracy. The team has applied for a patent and hopes to bring the product to the commercial market in three years, pending FDA approval. They are also working with companies that want to product a handheld device.
October 2, 2024Healthcare AI NewsComments Off on Healthcare AI News 10/2/24
News
Microsoft releases a major enhancement to Copilot that adds conversational capabilities, a virtual news presenter that will read headlines, webpage history recall, and the ability to answer questions about the text and images on a browser page.
FDA will hold the first meeting of its Digital Health Advisory Committee on November 20-21, which will address lifecycle considerations for AI-enabled medical devices. The session will be available via webcast with no registration required.
Business
Claimable launches an AI-powered appeals platform to provide documentation to help patients protest health insurer medication coverage denials for 60 major treatments. The service costs $39.95 per appeal and the company submits appeals via fax and first class mail. It claims that 80% of its appeals are accepted and most cases are resolved in less than 10 days.
WellSpan Health uses Hippocratic AI to develop an AI-powered conversational agent that contacts patients by telephone to close gaps in care.
GE HealthCare closes its $51 million cash acquisition of the ultrasound AI business of Intelligent Ultrasound Group.
Research
Clinicians compare the usefulness of OpenAI’s o1-preview, which features enhanced reasoning, to ChatGPT GPT-4 for medical AI:
The new version can perform advanced, step by step reasoning. The effect on diagnostic accuracy has not yet been studied.
o1-preview is slower.
GPT-4 is good for patient communication and medical advice, while the new version is better at complex reasoning, genetic analysis, and research. The value of those changes for specific medical specialties has not yet been studied.
GPT-4 is more prone to hallucinate.
GPT-4 is better at human-like conversation.
While transparency is limited with both versions, the new version provides chain of thought for double checking.
GPT-4 is trained to provide information that is broad but not deep, while o1-preview applies PhD-level reasoning in physics, chemistry, and biology.
A ChatGPT review of 725 websites of clinics that promote “complementary” and “alternative” medicine finds that 97% contain false or misleading claims, including some that relate to cancer treatment.
Other
A woman is told by the ED doctor who sent her home that her face pain and drooping was not concerning gets a second (and correct) opinion from ChatGPT, which advised her to seek immediate medical attention for what was possible Bell’s palsy. She want back to the ED, where the doctor agreed with the diagnosis and started treatment immediately.
A study of LinkedIn and Glassdoor job postings that mention AI finds that Mount Sinai Health System is #6.
I admire Senator Murray and appreciate her advocacy for a safe and functional EHR. However, do she and others realize…