Healthcare AI News 11/8/23

News

OpenAI announces significant new ChatGPT features at its first developer conference, most significantly the ability for users to create a tailored version of ChatGPT – called, somewhat confusingly, a GPT – to perform specific tasks. User-created GPTs can be shared or sold via a GPT Store that OpenAI will launch later this month.

Samsung introduces Gauss, its generative AI model.

Business

Smart stethoscope maker Eko Health earns UK regulatory approval for its AI-enabled tool for the evaluation of heart failure, valve disease, and atrial fibrillation, which Imperial College London will deploy to 100 GPs who will use it in conjunction with the company’s digital stethoscope. The system is sold in the US under the name Sensora.

Agamon Health  will work with Mayo Clinic to extend its offering into cardiology, where it will enhance patient adherence to follow-up.

Former Meta COO Sheryl Sandberg and other high-profile investors launch Cercle. The company uses AI to support women’s health, especially within fertility care, and organizes unstructured medical data into a standard format to help clinicians develop personalized treatment plans.

Israel-based Eleos Health, whose automated scribing solution turns behavioral health conversations into documentation and clinical insights, raises $40 million in a Series B funding round.

Research

Researchers find that ChatGPT does a reasonably good job in answering consumer questions about lifestyle-related conditions, making it appropriate for patients who are waiting on a doctor’s appointment.

Other

Experts in India write that integrating AI into India’s healthcare ecosystem could significantly enhance service delivery and patient care, though it requires overcoming challenges related to data, policy, and infrastructure.

Adweek reports that a hospital shut down its marketing team over compliance problems after it tested ChatGPT.

A Washington Post opinion piece by Leana Wen, MD, MSc describes Kaiser Permanente’s use of AI to flag patients whose condition is deteriorating using near real-time data whose patterns might otherwise be overlooked. She says that KP’s key difference is that it sends the alerts to an offsite team of nurses to decide whether the hospital’s rapid response time should be deployed. The original study she references was published in November 2020.

CNN says that MSN.com and Microsoft Start are featuring bizarre, inaccurate stories that are generated by AI instead of the editors it has fired. The company published an article about a woman’s death that it lifted from The Guardian, then added a “pathetic, disgusting” poll that invited readers to speculate on whether she died by murder, accident, or suicide. One of the laid off editors says the company features stories from small, unreliable sources that provide little information about their ownership or editorial standards. The story above that called a deceased NBA player “useless” was speculated to have resulted from AI that translated the original source’s article from English to another language and back again to hide plagiarism.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 11/1/23

News

A White House executive order on responsible use of AI assigns HHS to establish a safety program for receiving reports of AI-caused harm.

The VA launches a competition for developing AI solutions that reduce the time clinicians spend on taking notes or looking up patient medical records during appointments.

A new Microsoft Windows 11 update adds Windows Copilot, an AI chatbot that can help users open apps, summarize web pages from the Edge browser, and customize the Windows display. The company also started business sales today of Microsoft 365 Copilot, which offers Open AI-powered productivity enhancements to the former Microsoft Office 365 for $30 per user per month.

In England, Addenbrooke’s Hospital uses AI for the first time to interpret images of liver cancer patients to guide radiologists who are performing thermal ablation of a tumor.

Business

China-based technology giant Alibaba releases a healthcare-specific version of its Tongyi Qianwen AI model, which is one of the most powerful available.

AdaptX, which offers AI-powered EHR analytics tools, raises $10 million in funding. CEO Warren Ratliff, JD was a co-founder and COO of Caradigm.

Thomas Jefferson University Hospital expands its use of software from BigBear.ai for running operational planning scenarios related to resource and bed management, nurse staffing, and transfer center.

AI-enabled radiology performance measurement software vendor Covera Health raises up to $50 million in a Series C funding round and finalizes its acquisition of CoRead, a hospital AI quality assurance company.

Other

A Stoltenberg Consulting survey of CHIME members finds that interest in AI has grown fivefold since last year’s survey, taking the top spot in industry focus.

University of Iowa Health Care reminds employees that using AI to draft patient letters, or using AI-powered transcription services, involves sharing PHI that could trigger HIPAA violations.

William Hersh, MD, informatics professor at Oregon Health & Science University, recently delivered a presentation titled “Artificial Intelligence: Implications for Health Professions Education,” for which he shared the session recording.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 10/25/23

News

Washington University in St. Louis launches the AI for Health Institute, which will involve a collaboration of its engineering and medical schools.

NPR Shots covers doctors using AI for diagnosis, including a Mass General infectious disease doctor who is using the AI-enhanced version of Wolters Kluwer’s UpToDate — NPR calls it “Google for doctors” — that can conduct a conversation rather than simply looking up keywords in medical references. Other AI uses mentioned are interpreting diagnostic images, summarizing a patient’s medical history in preparation for their appointment, and allowing AI-assisted primary care physicians instead of specialists to care for some patients.

Business

Amazon begins rollout of AI-powered warehouse robots that it predicts will speed up order fulfillment by 25%. The system’s robotic arms and computer vision identify the correct item and send it to workers, which speeds throughput and places items at waist level to reduce worker injuries.

Research

A Stanford medical school study warns that AI chatbots may worsen health disparities for black patients by perpetuating race-based medical ideas that are known to be inaccurate. When asked to describe differences in skin thickness, lung capacity, and kidney function for a black man, the chatbots provided erroneous information instead of accurately stating that no differences exist. Some experts question the usefulness of the study’s conclusions since doctors shouldn’t be using chatbots to make medical decisions.

Researchers create software that uses AI to enhance CT images to produce MRI-quality information, which they say could be used to improve diagnosis in primary care.

Other

A Nature article predicts a rise in “generalist medical AI,” in which models that are trained on large data sets can mimic the broad analysis of a physician instead of performing a single function. It notes these challenges to AI in healthcare:

• AI can’t be trusted to make unreviewed decisions.
• Early medical imaging AI tools were trained on available imaging data that allowed them to perform tasks for which doctors don’t need help, such as detecting pneumonia.
• Individual tools don’t reflect the cognitive work of radiologists, and the plethora of one-trick AI tools could result in “an IT soup.”
• ChatGPT-like “Foundation models” that are trained on broad data sets of images, text, and other information can offer more capabilities than supervised learning, in which experts might analyze chest x-rays and label them as “pneumonia” or “not pneumonia” to train the system.
• Few institutions recognize that diagnostic AI models have usually been validated against the same data set that they were trained on, which in the absence of external validation makes it impossible to see if they are actually beneficial.

I hadn’t tried Google’s Bard lately, but enhancements that were release a few weeks ago give it some advantages over ChatGPT:

• It can access up-to-date information, such as news headlines and stock market information.
• It integrates with Gmail, Docs, and Drive to find or summarize content.
• Extensions are supported for Google-owned content, such as Maps and YouTube, and can be invoked specifically by using a keyword such as as @YouTube.
• Bard’s initial responses can be double-checked via a Google search by clicking a button.
• It can include images in its responses.

I also experimented with Pi, which terms itself as a “personal AI assistant” that emphasizes friendly, personal, and supportive responses. It serves as a polite sounding board, more of an AI companion than a brute force data retrieval engine, and it can conduct natural-sounding conversations. Developer Inflection AI recently raised $1.3 billion at a $4 billion valuation.

I’m also experimenting with Summarize.tech, which summarizes a YouTube video given its link. I gave the link for a recent VA EHR hearing and it did a pretty good job, complete with links to jump to individual video sections. 

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 10/18/23

News

OpenAI adds voice capabilities to the ChatGPT mobile app, which can carry on a conversation with the user through one of several available natural-sounding voices. Existing voice assistants such as Alexa, Siri, Google Assistant, and Cortana will need to follow quickly or risk becoming instantly obsolete. 

Rumors suggest that OpenAI will soon launch AI-powered autonomous agents, which can interact with other software such as email or calendars to complete tasks without supervision.

Vanderbilt University Medical Center develops an EHR-embedded AI tool to identify pediatric patients who are at risk for blood clots. Outcomes were no better than for patients in the control group, however, which the researchers say may be due to physicians rejecting the tool’s recommendation to start blood-thinning therapy only 25% of the time due to fears of causing a major bleed. Yale medical school dean F. Perry Wilson, MD, MSCE observes that the VUMC study illustrates a key point that making accurate predictions is table stakes that vendors shouldn’t launch an “accuracy arms race” since “even perfect prediction is useless if no one believes you or if they don’t change their behavior.”

Business

UAE-based M42 launches Med42, a clinical large language model that answers medical questions using synthesized knowledge. The company claims it outperforms ChatGPT 3.5. It’s free for non-commercial use and research and can be downloaded from Hugging Face.

Research

ChatGPT performs at least as well as primary care physicians in choosing treatment options for newly diagnosed depression, and unlike doctors, it does not exhibit gender or socioeconomic bias. ChatGPT nearly always recommends psychotherapy for mild cases versus the 4% of PCPs who do so. It also favors using antidepressants alone for severe cases, while doctors usually add anxiolytics or hypnotics.

A Klick Labs study finds that AI analysis of a 10-second recording of a person’s voice can determine if they have Type 2 diabetes with nearly 90% accuracy. The company says that the non-intrusive, accessible approach will allow screening large numbers of people for Type 2 diabetes and potentially other chronic conditions.

Researchers are surprised to find that ChatGPT does a poor job of reviewing urology residency application letters.

Other

Experts predict that AI will play an outsized healthcare role in Canada because of clinician shortages in rural areas, with initial use in creating clinician documentation, providing easier access to patient records, helping with staff scheduling, and reviewing digital images.

Healthcare AI experts list ways that AI can help in healthcare, which include providing patient-friendly explanations, performing back office functions, predicting next events in a patient’s journey, analyzing images, keeping patient data private, and training itself. On the downside, it can create disinformation that sounds plausible, introduce bias in agreeing with what the user says, provide a false sense of security, and cause professionals to question whether their knowledge or careers are obsolete.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 10/11/23

News

Microsoft introduces new healthcare-related tools to its Fabric analytics platform that can assemble and standardize data from multiple sources – such as EHRs, imaging systems, lab systems, medical devices, and claims data – and present it in a single view. The company also announced Azure AI Health Bot, which can answer staff questions about treatments and protocols and patient portal queries about symptoms and medical terms. Microsoft also announced a text analytics solution, along with generative AI models that create a patient history, simply medical reports into patient-friendly language, and help radiologists identify possible radiology report errors.

Google Cloud rolls out AI-powered clinician search for its Vertex AI Search platform. The company says it will speed up clinician EHR searches and to perform more complex operations such as suggesting billing codes or determining if patients meet enrollment criteria for clinical trials. It can also cite and link the source of the information that it finds to ease concerns about hallucination.

University of California health systems and other groups launch VALID AI, which will review uses, pitfalls, and best practices for using generative AI in healthcare and research. The invitation-only group’s name comes from Vision, Alignment, Learning, Implementation, and Dissemination.

Business

Permanente Medical Group will expand its pilot project with France-based autonomous scribe vendor Nabla to 10,000 physicians. The browser-based Nabla Copilot generates clinical notes from conversations between providers and patients and offers white-label API integration with EHRs.

FluidAI Medical raises $15 million in a Series A funding round. The company offers an postoperative patient monitoring tool, powered by sensors and AI, that warns surgeons of anastomotic leaks in the GI tract that has an 8% surgical incidence with 12% mortality. The company is based in Canada and its product has not been cleared for sale in the US, although it lists Cleveland Clinic and Texas Medical Center as partners.

Research

Researchers say that clinical use of AI-powered predictive models feeds data back into the EHR on which model updates are trained, which will then reduce their accuracy when their training is updated. The authors observe that retraining can actually degrade model performance, as updated EHR information disrupts the connection between presentation and outcome. They recommend that health systems document the machine learning predictions that are used on a given patient and warn that a “model-eat-model world” can render an individual model and future models worthless.

Other

Biomedical researchers tell Stat that nobody knows if a given healthcare algorithm is useful because the companies that develop them don’t share data with researchers or anyone else. The authors propose that a federal agency test algorithms against a standard test set and publish their accuracy results, including a breakout by demographic groups, that FDA would use in reviewing a product for approval. They note as a precedent NIST, which evaluates facial recognition software in publicly available reports.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 10/4/23

News

Augmedix offers early access to Go, which uses generative AI to create a draft medical note after each visit.

CloudWave releases an AI security and privacy policy template that guides AI implementation and management in healthcare, including risk assessment, audit and compliance procedures, ethical considerations, and review and revision procedures. 

Zoom extends its AI Companion summarization capabilities to healthcare customers. They include whiteboard idea generation, fast meeting review and catch-up via highlights and smart chapters, and drafting quick messages based on team chats.

Sentara Health expands its implementation of Regard’s AI clinical documentation solution to all 12 of its hospitals.

Virtual provider HealthTap launches Dr. AI, a pre-appointment patient interview that uses GPT-4 to collect the patient’s information, then present it concisely to the physician.

HeHealth launches its AI-powered sexual health app in the US, offering to “help all penis owners” who suspect they have a sexually transmitted infection. Users scan and submit photos to receive an initial AI report that is followed by physician review. The service, which costs $10, allows users to remain anonymous. Co-founder and CEO Yudara Kularathne, MBBS is a Singapore-based ED physician 

Business

Health Data Analytics Institute raises $31 million in a Series C funding round. The company’s AI-optimized platform quantities patient and population risks and identifies clinical improvement opportunities.

Research

A Spanish hospital is conducting trials of Cordio Medical’s HearO app, which hopes to detect early signs of congestive heart failure by analyzing a patient’s speech patterns, specifically those indicative of lung fluid buildup.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 9/27/23

News

Nuance announces GA of Dragon Ambient EXperience (DAX) Copilot, formerly known as DAX Express, that uses conversational, ambient, and generative AI to help physicians create medical documentation.

OpenAI updates ChatGPT to support using browser-based data, effectively bypassing the previous September 2021 training cutoff. To activate, click GPT-4 and then “Browse with Bing (beta).”

Business

A planned share sale by OpenAI values the company at up to $90 billion.

OpenAI investor and competitor Microsoft is developing smaller, cheaper alternatives to ChatGPT using existing Microsoft technologies.

Drug maker Novo Nordisk, which sells weight loss drugs Ozempic and Wegovy, partners with AI-powered drug discovery vendor Valo Health to study cardiometabolic drug programs, also acquiring three pre-clinical cardiovascular drug discovery programs from Valo. Novo Nordisk will pay $60 million upfront, up to $2.7 billion in milestone payments, fund the R&D costs, and possibly pay royalties. The 45-year-old Valo founder, CEO, and board director David Berry, MD, PhD, has founded 25 companies and holds 200 patents.

Research

Researchers at Kennesaw State University will study the use of RFID clothing tags to capture body motion for the study of health problems in the elderly, especially those involving fall risk.

Other

Students at University of Calgary’s medical school create an app that allows students to practice their interactions with patients and receive feedback on their performance along with suggestions they could have asked.

Researchers find that using AI chatbots to find an ophthalmologist returned results that were biased and sometimes wrong, choosing a disproportionate percentage of male doctors and those who practice in an academic setting.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 9/20/23

News

Oracle announces Clinical Digital Assistant, a voice-first mobile app for clinicians that responds to voice commands, uses ambient listening to generate clinical notes from conversations, allows creating notes and clinical documentation from voice narration, and creates notes and chart details.

Ochsner Health is testing the use of Epic’s new feature that creates draft responses to MyChart patient messages that physicians then review.

Drug and medical technology maker Johnson & Johnson describes how it is using AI:

• Applying AI to diagnostic tests and connected medical devices to detect disease earlier.
• Driving drug discovery.
• Analyzing de-identified patient datasets to identify clinical research research sites, to bring trials to patients outside of major medical centers, and to diversify trials.
• Analyzing demand and inventory trends to prioritize drug shipments.
• Creating a “highlight real” of surgical video that allows surgeons to perform post-case analysis that can be used for teaching.

In Japan, commercial insurer Sompo – which is the country’s largest nursing home operator with 301 facilities and 58 day centers — will install TytoCare’s remote medical examination technology in its facilities.

Business

Denmark-based Corti, whose voice assistant system listens to doctor-patient conversations and makes AI-powered suggestions based on millions of other such conversations, raises $60 million in a Series B funding round.

Research

Chan Zuckerberg Initiative will fund and build a 1,000 GPU computing cluster that will allow researchers to study how cells behave in health or disease

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 9/13/23

News

A Harrisburg, PA TV station covers UPMC’s use of Abridge to create transcripts of doctor-patient encounter conversations. UPMC VP/CMIO Salim Saiyed, MD, says he has never had a scribe, but the system is like having a real-time scribe whose accuracy is 95%. He adds that the physician can focus on the patient instead of the computer and spends less “pajama time” completing documentation after hours.

Mass General Brigham and GE HealthCare will collaborate to develop an AI algorithm that will optimize radiology patient scheduling as part of GE HealthCare’s commercially available Radiology Operations Module.

OpenAI will hold its first developer conference on November 6 in San Francisco.

Business

RCM technology vendor Aspirion acquires AI-powered intelligent document processing platform vendor Infinia ML. Aspirion says it will run Aspirion as an R&D development shop that will be run by Infinia ML executive chair Nick Giannasi, PhD,  who was previously chief AI officer of Change Healthcare.

Crunchbase News leads off its list of “failed and struggling AI-focused unicorns” with Babylon Health, which it notes mentioned “AI” 25 times in its SPAC filing, raised $600 million at a valuation of billions, then filed bankruptcy last month.

Notable launches a conversational AI tool that helps patients find doctors, schedule appointments, create directions to a location, pay bills, and request prescription refills. It is fully customizable (including tone of voice), can converse in 130 languages, and can have its training extended by uploading training manuals or other documents.

A former Google AI researcher raises $100 million from big-name Silicon Valley investors for Inceptive, his startup that will use AI to develop drugs and vaccines.

Research

Researchers note that ChatGPT has accumulated 1,000 PubMed citations in nine months, which predicts its future influence even though one-third of the cited articles were opinion pieces. Google needed 14 years to hit the 1,000-citation mark.

Other

Attorneys note the healthcare regulatory risks of using AI:

• Running afoul of privacy laws when training AI systems using patient data.
• Using AI to replace rather than support clinicians, which could be construed as the unlicensed practice of medicine.
• Providers deviating from the standard of care based on AI recommendations, creating malpractice risk.
• Reducing physician supervision of lower-level employee who instead use AI, which could fail to meet payer requirements.
• Running afoul of the Anti-Kickback Statue if using vendor-developed AI algorithms that favor use of their products.
• Inadvertently practicing unlawful discrimination by using algorithms that use factors such as race, color, national origin, sex, age, or disability.

The Today show features a mother whose son’s chronic pain and other symptoms had baffled 17 doctors who had seen him over three years. She pasted notes from his MRI into ChatGPT, which provided a diagnosis of tethered cord syndrome, a deformation that limits the movement of the spinal cord and is related to spinal bifida. He is recovering from surgery to repair the condition.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 9/6/23

News

Google Chief Health Officer Karen DeSalvo, MD, MPH, MSc says in NEJM Catalyst that large language models will give patients “a more personalized and anticipatory experience” and that medicine needs to change to meet consumer expectations for a mobile-first digital health experience.

A hospital in Israel declares that an AI app saved a patient’s life when it detected brain bleeding from their CT scan and alerted doctors, who called him to return immediately for additional scans and surgery. The app was developed by Viz.ai.

Research

A study finds that consumer voice assistants such as Alexa as well as ChatGPT delivered poor results when asked layperson questions about performing CPR, with the authors suggesting that bystanders should call 911 instead of asking Siri. The voice assistants gave “grossly inappropriate” responses that often weren’t related to CPR and almost always failed to recommend calling emergency services.

National Taiwan University Hospital develops an ultrasound device that uses AI to assess a person’s risk of sleep apnea in 15 minutes instead of an overnight sleep study. The machine diagnosed sleep apnea with 95% accuracy in requiring patients to breathe a few times while awake. At-risk patients are then referred for traditional polysomnography.

Other

OSF HealthCare develops an AI model that predicts an inpatient’s death at admission using 13 commonly available data elements, allowing clinicians to work with the identified patients to document their end-of-life wishes.

Two teens launch Vytal.ai, whose app uses the camera of a laptop or mobile device to assess brain health using a 30-second gaze tracking exercise. The high school students have developed use cases that include early detection of neurological problems, clinical trials, and concussion screening. Co-founder, CEO, and CTO Rohan Kalahasty, who is 18, has spent three years as a researcher in Harvard’s ophthalmology AI lab and has performed AI research at MIT, while co-founder, COO, and CFO Sai Mattapali, aged 17, has spent time as an intern in both neurophysiology and business growth.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 8/30/23

News

OpenAI releases an enterprise version of ChatGPT that features enhanced security, privacy, and speed. The product overlaps the offerings of the company’s investor Microsoft, which provides a similar offering as part of Azure. The ChatGPT service costs $30 per user per month.

Hackensack Meridian Health will work with Google Cloud on generative AI projects, while HCA Healthcare announces similar plans.

Business

Viz.ai will commercialize three ECG AI algorithms that were developed by UCSF. The algorithms detect cardiac amyloidosis, pulmonary hypertension, and supraventricular tachycardia.

AI-powered, message-based care provider Curai Health joins Amazon Clinic’s virtual care marketplace.

Drug maker Bayer expands its digital therapeutics business with a partnership with Mahana Therapeutics, which offers an FDA-cleared digital product for irritable bowel syndrome.

Research

A consumer research survey finds that 82% of Americans haven’t used ChatGPT, 81% of them don’t expect it to have a major impact on their jobs, and 85% don’t think it can help them do their work.

A study of EHR data in Israel finds that AI can predict food allergies in newborns, surprising researchers who found a significant correlation in the exposure of pregnant women to antibiotics. They conclude that the antibiotics may have interfered with the babies’ microbiome, which influences allergy development.

ChatGPT successfully improves the readability of patient consent forms, researchers find, creating versions that are shorter and easier to read. The required reading level dropped from college freshman to eighth grader.

ChatGPT-generated cancer treatment plans are full of mistakes, researchers find, where one-third of the tested plans contained incorrect information. ChatGPT also mixed correct and incorrect information together in a way that made it hard for even experts to detect.

Other

KFF News headlines a story about Bamboo Health’s NarxCare as “Artificial Intelligence May Influence Whether You Can Get Pain Medication.” The software predicts the likelihood that a patient will overdose based on their documented use of narcotics, sedatives, and stimulants. CDC has warned providers to make sure use of the algorithm doesn’t harm patients, such as turning them away for visits or denying them medications for documented chronic pain, and has voiced concern about “proprietary algorithms” whose methods are not transparent.

A startup sells Nextflix-type subscriptions to access the AI clones of celebrities who earn recurring earnings for doing little more than providing content from elsewhere and setting the guidelines for how their clones will operate. Experts fear the chatbot will worsen loneliness and encourage replacing human interactions with “parasocial” ones. The founders noted the success of a YouTube influencer whose “virtual girlfriend” earned her $71,000 in its first nine days.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 8/23/23

News

In China, the Beijing Municipal Health Commission proposes prohibiting the use of AI by online services to generate prescriptions, make diagnoses, and deliver treatment. The commission also plans to take a supervisory role for medical institutions that run online services.

Business

Microsoft lists its AI collaboration efforts with Epic as showcased at UGM:

• Note summarization for clinicians to speed up documentation.
• Nuance DAX Express embedded in Hyperdrive and Haiku.
• Providing medical coding staff with suggestions based on the clinical documentation.
• Adding NLP queries and interactive analysis to SlicerDicer.

Genesis Therapeutics, which hopes to advance its AI-discovered drug pipeline to clinical development, raises $200 million in a Series B funding round.

Suki integrates its AI-powered clinical documentation assistant with Cerner, adding to its previously announced integration with Epic.

Research

A Mass General Brigham study finds that ChatGPT delivered 72% accuracy in making diagnoses and care management decisions throughout a patient’s complete case, performing equally well in primary care and emergency settings. ChatGPT struggled with differential diagnosis, demonstrating the value of physician expertise at the beginning of the encounter, where a small amount of presenting information is used to consider possible diagnoses.

Other

Three physicians say in a Time opinion piece that healthcare AI could create a “nightmare scenario” where AI tools block patients from talking to humans for reassurance or to offer input in care decisions. Studies have shown that AI is perceived by some as more empathetic and compassionate than doctors, which means it’s time to hit the “reset button” on how doctors approach patient communication and teach those skills to medical students and residents.

Dermatologists describe how ChatGPT could be used to support rural dermatology practices, including creating disease-specific educational patient handouts at appropriate readability levels, generating procedural note templates, and creating drafts of prior authorization requests.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 8/16/23

News

Abridge, whose app turns doctor-patient conversations into transcripts and visit summaries, becomes the first member of Epic’s Partners and Pals integration program. The company also announced that Emory Healthcare will deploy its solution.

Piction Health launches a dermatology virtual clinic in Massachusetts in which patients complete an online intake questionnaire, attach at least three pictures of their area of concern, and then receive a dermatologist’s evaluation – supported by AI image analysis and comparison to its own image database — within two days. The company says that two-thirds of patients receive a treatment plan without an in-person visit. The service is available for patients in New Hampshire, Connecticut, Massachusetts, and Florida and is covered by some insurances or costs $80 otherwise.

Business

Morgan Stanley lists four areas in which it expects AI to create significant investment opportunities:

• Drug discovery, manufacturing, and physician-patient engagement.
• Earlier detection of disease, more efficient patient access, claims processing, supply chain management, and helping patients use their insurance and prescription benefits effectively.
• Support for advanced diagnostics that combine EHR, genomic, and imaging data to gain disease insight and to personalize treatment.
• Enhance monitoring by analyzing data streams from medical devices such as sensor-equipped implants, continuous glucose monitoring, and cardiac monitoring.

Research

Vanderbilt University Medical Center researchers develop an NLP algorithm that analyzes EHR data to identify patients who need lung cancer screening, improving existing methods that look only at EHR smoking flags. The team found that nearly half of the documented smokers are missing discrete EHR data related to pack years and quit dates, but the NLP analysis of clinical notes delivered 96% accuracy.

Taiwan’s National Cheng Kung University creates AI-powered software to detect age-related muscle loss (sarcopenia) from CT scans. The government has approved the software for sale to other hospitals, although the creators suggest that hospitals perform screening of cancer patients for free since they already have their CT images. Sarcopenia, which increase fall and fracture risk in older patients and those undergoing chemotherapy, can be improved through diet and rehabilitation if caught early.

Carnegie Mellon University is using AI, ML, and NLP to analyze YouTube videos that offer patients information on conditions or treatments for accuracy, comprehension level, trustworthiness, and delivery of actionable guidance. They hope to create filters that give doctors a short list of videos that can be reviewed and then prescribed as needed.

Newly developed, AI-powered “smart socks” will allow people with dementia to live at home, where the wearable tracks heart rate, sweat levels, and motion to determine when the wearer is in distress. The socks look like real socks, are machine washable, and do not require charging.

Other

In Canada, a report warns that successful use of healthcare AI will be limited by the healthcare system’s reliance on fax machines, handwritten notes, scanned files, and paper recordkeeping.

Veterinary chain Banfield Pet Hospital offers customers a free AI-powered device – developed by Whistle Health, which like Banfield, is owned by candy and dog food maker Mars — that attaches to a dog’s collar to track caloric intake, scratching that could indicate skin problems, and other health data, with an app that allows chatting with a veterinarian. An enhanced version allows tracking the pet’s location.

An AI expert notes that AI-created influencers – realistic, computer-generated characters – will increase mental health issues and suicides among followers who can’t function in the real, imperfect world. He says that mental health was already threatened by the false perfection of Photoshopping and plastic surgery. An AI-generated influencer who portrays a young woman from Finland has thousands of followers on TikTok and Instagram, while another influencer created an AI-powered “virtual girlfriend” version of herself using ChatGPT that she expects to generate $60 million per year in subscriptions from an audience that is 98% male.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 8/9/23

News

Senator Mark Warner (D-VA) expresses concerns to Google officials that hospitals are testing the company’s Med-PaLM 2 large language model. He asks specifically whether the LLM memorizes the full set of a patient’s data, whether patients are notified of its use or are offered the chance to opt out, and for the company to provide a list of those hospitals that are participating in testing. Warner raised questions in 2019 about whether Google’s “secretive partnerships” with hospitals that would use patient data without their consent could create privacy issues.

A hospital in Israel adds a startup’s ChatGPT-powered clinical intake tool to its ED admission process. It collects the results of a three-minute chatbot Q&A that the patient answers in their own words, after which the system generates a condition summary for the doctor.

Business

MUSC Health will use Andor Health’s ChatGPT models to create a virtual care ecosystem that will include virtual visits, virtual hospital, virtual patient monitoring, virtual team collaboration, and virtual community collaboration.

A Bain survey of health system executives finds that while 75% of them think that generative AI has reached the turning point that is necessary to change healthcare, only 6% of their organizations have established a strategy to use it. They rank the top three uses over the next 12 months as clinical documentation, analyzing patient data, and optimizing workflows, while within 2-5 years they will be looking at using AI in predictive analytics, clinical decision support, and making treatment recommendations.

Research

A University of Maryland School of Medicine article says that physicians need more training in probabilistic reasoning to productively use AI-powered clinical decision support. The authors suggest that physicians undertake training in sensitivity and specificity, to help them understand test and algorithm performance, and learn about how they should use algorithm recommendations in their decision-making.

Harvard Medical School researchers find that AI-generated narrative radiology reports aren’t yet as good as radiologist-generated ones, but they have developed two tools to evaluate them for ongoing improvement.

Other

Pharma bro and former federal prisoner Martin Shkreli takes social media umbrage with a PhD AI researcher who thinks LLMs like the one he’s selling shouldn’t give medical advice, calling her an “AI Karen.” Shkreli claims that his Dr. Gupta, which uses ChatGPT, will ease physician burdens, reduce healthcare costs, and help the economically disadvantaged. Shkreli’s claim to healthcare fame was buying rights to a old, cheap drug to treat parasitic disease and immediately jacking up its price from $13.50 to $750, after which the FTC forced him to return his $65 million in profit for suppressing competition. Shkreli has also created a veterinarian version of Dr. Gupta called Dr. McGrath. Experts note that in addition to the legal exposure of providing medical advice over the Internet, Dr. Gupta at one time identified itself as a board-certified internist, although it now answers the identity question with, “I understand that you may have questions about my credentials, but let’s focus on addressing your symptoms and concerns.”

Time magazine recaps how struggling New York City-based urgent care chain Nao Medical is apparently using AI to generate nonsensical articles to improve its search engine rankings. The above article helpfully clarifies the understandable confusion between color guard and colonoscopy (or its own failure to know the difference between color guard and Cologuard colon cancer screening test), noting the subtle difference that “Color guard is a performance art, while a colonoscopy is a medical procedure.” The young software engineer who runs the company previously developed Fake My Fact, which generated phony Google results and online evidence to use “when you knew you were wrong but wanted to be right.”

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 8/2/23

News

Duke Health will create an AI Innovation Lab and Center for Excellence in partnership with Microsoft, with one of its research topics being the reliability and safety of generative AI in healthcare. Duke Health will use Azure and its OpenAI Service.

The British Standards Institution issues guidance for using AI in healthcare that includes criteria for evaluating products for clinical benefit, performance standards, safe integration into clinical environments, ethical considerations, and equitable social outcomes.

The Australian Medical Association, which previously warned doctors to stop using ChatGPT to write medical notes because it does not protect patient confidentiality, says that Australia lags other countries in regulating AI. It calls for clinicians to always make final decisions and to obtain patient consent before using AI for their treatment or diagnosis.

Business

Using synthetic patient data for research – applying AI to convert real medical records to broader sets of realistic but artificial data that involves no privacy issues – is appealing, but the industry is challenged to create those data sets while preserving many relevant variables so that the results reflect real-life situations. The WSJ article also notes that most of the data vendor markets are young startups, as big cloud providers that might be a more comfortable choice for customers haven’t shown an interest in that market so far.

Forbes profiles OpenEvidence, which collects current medical data before answering questions, an “open book” approach that works around the expensive problem of constantly retraining AI models, potentially competing with human-edited reference databases such as Wolters Kluwer UpToDate. OpenEvidence analyzes 31,000 peer-reviewed journals from the National Library of Medicine, weighted by impact factor, and places their content into the company’s retrieval database within 24 hours. Physicians who register can test the system by asking their own questions. The founder’s previous company developed an AI-powered tool that analyzes the stock market to identify arbitrage opportunities. OpenEvidence is a Mayo Clinic Platform Accelerate company.

Healthcare talent software and staffing company Aya Healthcare acquires two AI software companies whose tools forecast patient demand and staffing needs.

Other

Venture capital firm Andreessen Horowitz says that AI’s biggest impact will be in healthcare, predicting that the industry will jump from fax machines straight to AI similar to how the developing world skipped credit cards in going from cash to mobile payments. It says that any new technology must be 10 times better than whatever it displaces since marginal improvements aren’t worth the effort, noting that while enterprise software failed to clear that 10x bar in healthcare, AI does so easily. It predicts that AI can’t fail because it can help address clinician shortages and reduce costs while saving lives.

Former FDA Commissioner Scott Gottlieb, MD says that AI will perform some of the activities of doctors “sooner than we think.” He sees four areas of AI usage: (1) processing claims or creating and analyzing medical records, such as Amazon HealthScribe; (2) analyzing medical images and ECGs; (3) providing clinical decision support using patient-specific data; and (4) AI operating as an independent tool to diagnose and prescribe, although he admits that is unlikely any time soon given regulatory caution. He says that using AI in a clinician-supporting role won’t lower costs because healthcare is a people-centric business where workers are paid increasingly more despite no change in their productivity (the Baumol effect), with no cost improvement unless AI replaces doctors completely.

A NEJM perspective piece says that medical educators should take an activist approach to integrating AI into physician training, because if they don’t, external forces that are motivated by efficiency and profits will do it for them. The authors question the role of cognitive apprenticeship when medical students will likely use AI-powered chatbots on their first day of training and how lifelong learning fits with using AI to  immediately and reliably answer questions at the point of care.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 7/26/23

News

Amazon Web Services launches AWS HealthScribe, which allows developers to create clinical documentation tools that use speech recognition and generative AI. Software companies can use a single API to create transcripts, extract terms, and create summaries of doctor-patient conversations. The service will initially support applications for general medicine and orthopedics. AWS HealthScribe is priced per second of audio transcript, with the company estimating that processing 1,000 visit transcripts per month would cost $1,500. The company also says that the service is HIPAA-eligible and the company will not use its content for machine learning training.

ChatGPT developer OpenAI sunsets its tool that it hoped would be able to identify whether text was generated by AI or humans, citing the software’s low rate of accuracy.

Sen. Elizabeth Warren (D-MA) is investigating Google’s attempts to gain access to data from the military’s Joint Pathology Center for use in building healthcare AI tools, also accusing the Defense Department of showing favoritism toward Google. ProPublica previously reported that Google offered to digitize the military’s 55 million pathology slides in return for exclusive access to the archive.

Business

Safety-focused healthcare large language model developer Hippocratic AI raises another $15 million in its seed funding round, increasing its total to $65 million.

Vanderbilt University Medical Center will use Nference’s AI-powered federated clinical analytics platform to analyze combined de-identified data from EHRs, imaging systems, and digital pathology in biomedical research. VUMC will digitize its glass slide pathology using a digital pathology solution from Nference-owned Pramana.

Korea Exchange warns investors that share prices of some of the country’s medical AI companies show signs of manipulation, noting as an JLK, whose shares are up 1,000% so far this year.

Other

A hospital in Belgium is using a smartphone app called PMcardio that can analyze an ECG in five seconds to detect 38 cardiovascular conditions, either as a second opinion for cardiologists or as a primary diagnostic tool for ED doctors and GPs. Powerful Medical says its app is available for use in 15 European countries.

Researchers confirm that Prevencio’s AI-powered lab tests can accurately determine the risk of heart attack, stroke, and cardiac death, as validated by coronary or peripheral angiography. The company’s seven tests are AI-derived and algorithmically score multi-proteomic biomarkers. President, CEO, and board member Rhonda Rhyne is a pharmacist with life sciences leadership experience who joined the company in 2013.

The VP of data science at Unity Health, the first hospital in Canada to form a dedicated applied AI team, describes how AI might be used by hospitals:

• AI will make sense of the data that is already stored in hospital systems.
• It will get better at highlighting at-risk patients.
• High quality, nationwide health data sets are needed to create algorithms for rare conditions.
• AI should be trained on peer-reviewed publications rather than incomplete or low-quality data found on the Internet.
• Unity Health will use generative AI to help generate admission notes, discharge notes, and progress notes that clinicians will review, edit, and approve.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.

Healthcare AI News 7/19/23

News

Meta and Microsoft release Llama 2, a fully open competitor to ChatGPT 3.5 that is free for research and commercial use. Unlike ChatGPT, users can add their own data to Llama 2.

Elon Musk forms XAI, an AI company that he says will seek to “understand reality” and will work with Twitter, Tesla, and other Musk-owned companies.

Teladoc Health will use Microsoft’s AI services and its Nuance DAX ambient documentation product into its virtual care platform, expecting to automate the creation of clinical documentation during virtual exams. The company’s medical group will use Nuance DAX Express for the visits it provides directly.

Engineering consortium MLCommons develops MedPerf, an open benchmarking platform that evaluates the performance of AI models on real-world medical data while preserving patient privacy.

Business

Causaly, which created an AI platform for drug development, raises $60 million in a Series B funding round.

Nividia invests $50 million in AI drug discovery vendor Recursion Pharmaceuticals, which will train AI models on Nvidia’s cloud platform. RXRX shares jumped 80% on the news, valuing the 10-year-old company at $2.6 billion.

Research

A Google paper proposes supporting medical imaging predictive AI in a “know when they don’t know” manner, where the extra AI layer can decide via a confidence score whether it should defer to a clinician. The use of such a system reduced false positives by 25% while still identifying all true positives.

A new study finds that ChatGPT’s healthcare-related output is hard to distinguish from that created by healthcare providers, but patient trust decreased as task complexity increased, suggesting that the best use of healthcare chatbots is to assist with patient-provider communication related to administrative tasks and routine management of chronic conditions.

Researchers find that Google’s PaLM large language model generated long-form answers to common medical questions that aligned with scientific consensus just 62% of the time, but system tuning improved performance to equal that of human clinicians, with 93% of its answers found to be scientifically correct. The system generated potentially harmful answers 5.8% of the time, slightly outperforming clinicians.

A Science perspective piece predicts that pairing AI applications with medical robotics will create a new era of medicine in which autonomous robots could perform diagnostic imaging and surgical procedures as well as create and optimize the use of prosthetics.

Researchers are developing an AI too that can quickly recognize the genetic features of gliomas, the most common form of brain cancer, providing a molecular diagnosis in 15 minutes versus the manual process that takes weeks. Surgeons could use the results to make immediate operating decisions in the OR.

Other

UMC Health System deploys ZeroEyes, an AI-based platform that identifies guns from live security camera video streams. The company’s monitoring center can verify threats, issue alerts, and dispatch first responders within 3-5 seconds of detection.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Contact us.