What happened in health care technology this week, and why it’s important.
Why you shouldn’t trust AI search engines
Last week was the week chatbot-powered search engines were supposed to arrive. The big idea is that these AI bots would upend our experience searching the web by generating chatty answers to our questions instead of just returning lists of links as searches do now. Only … things did not go according to plan, Melissa Heikkilä, senior reporter at MIT Technology Review, reports. Approximately two seconds after Microsoft let people poke around with its new ChatGPT-powered Bing search engine, people started finding that it responded to some questions with incorrect or nonsensical answers, such as conspiracy theories. Google had an embarrassing moment when scientists spotted a factual error in the company’s own advertisement for its chatbot Bard, which subsequently wiped $100 billion off its share price. What makes all of this all the more shocking is that it came as a surprise to precisely no one paying attention to AI language models.
“Rather than taking a careful approach to this, they’re going in a very bold fashion. Let the [AI system] make mistakes, because now the cat is out of the bag. We’re all guinea pigs at this point,”
Chirag Shah, Professor, University of Washington
Why it’s important – Here’s the problem: the technology is not ready to be used like this at this scale. The recent blunders from Big Tech don’t mean that AI-powered search is a lost cause. Search is just one of the areas where the two tech giants are battling each other. They also compete in cloud computing services, productivity software, and enterprise software. Conversational AI becomes a way to demonstrate cutting-edge tech that translates to these other areas of the business. Essentially, we—the users—are now doing the work of testing this technology for free.
Infographic of the week – The Figure below, from Dr. Eric Topol’s Ground Truths article, shows the opportunity for Hospital@Home, and other use cases, going forward. Integrating such data has been a limiting issue. Still, large language models (LLMs) that can deal with over a trillion parameters, including the interface between text, speech, and images, are well suited to actualize this capability. I recently reviewed these LLMs or foundation models, the exemplar ChatGPT is certainly the craze right now. While there’s been much interest in sensors that collect physiologic data, the progress in wearable patches for self-ultrasound imaging is also notable. He emphasizes that for HaH and any other use cases in the Figure below, this—multimodal AI—has not been done yet. It’s an aspiration that may be accelerated with LLMs, but the case hasn’t been cracked yet.
First 3D Printed Pediatric Medicine Trials to begin in Europe
A hospital in Spain is launching a study to test the efficacy, tolerability, and acceptance of a drug produced for children using a 3D printer, according to an article in 3D Printing.com. The study will be the first clinical trial with a 3D-printed drug in Europe in the pediatric field and is a result of collaboration between the Pharmacy Service at Vall d’Hebron University Hospital, the University of Santiago de Compostela, and the company FabRx. The 3D printer used for the study produces medicines in semi-solid and chewable forms, which are personalized to each child based on their weight and clinical characteristics.
Why it’s important – This new method of administering medication is much more convenient and avoids dosing errors compared to traditional syrups. The clinical trial will not only test the efficacy of the new formulation but will determine whether it increases acceptability and improves the experience of minors and their families. The 3D-printed medications have a taste that masks the active ingredient and can be customized based on each child’s preference. Additionally, they do not need to be refrigerated, increasing safety and ease of transport.
Video of the week – Mark Cuban Assails PBMs at AAM meeting – Mark Cuban made scathing, profanity-laced remarks about pharmacy benefit managers (PBMs) and heaped praise on drug manufacturers during a 26-minute fireside chat interview at the annual meeting of the Association of Accessible Medicines, the trade association of generic drug manufacturers. Currently, the company only sells generics, but Cuban said that it was getting ready to sell brand-name drugs. He was coy about sharing any specifics about the company’s plans for selling insulin and biosimilars but suggested that it would.
Neuroscientists listened in on people’s brains for a week. They found order and chaos.
Another article from MIT Technology Review, this one authored by Jessica Hamzelou, on a study that shows that our brains exist between chaos and stability—a finding that could be used to help tweak them either way. Avniel Ghuman and Maxwell Wang at the University of Pittsburgh wanted to know what happens in the longer term. After all, the symptoms of many neurological disorders can develop over hours or days, says Wang. To get a better idea of what might be going on, the pair devised an experiment to let them watch brain activity for around a week. The researchers recruited 20 individuals to volunteer in their study. Each person had 10 to 15 electrodes implanted for between three and 12 days. The pair collected recordings from the electrodes over the entire period. The volunteers were all in the hospital while they were monitored, but they still did everyday things like eating meals, talking to friends, watching TV, or reading books.
The team found some surprising patterns in brain activity over the course of the week. Specific brain networks seemed to communicate with each other in what looked like a “dance,” with one region appearing to “listen” while the other “spoke,” say the researchers, who presented their findings at the Society for Neuroscience annual meeting in San Diego last year.
Why it’s important – Understanding how our brains restore some degree of stability after chaos could help us work out how to treat disorders at either end of this spectrum. Too much chaos is probably what happens when a person has a seizure, whereas too much stability might leave a person comatose, say the neuroscientists behind the work. A better understanding of what’s going on could one day allow us to use brain stimulation to tip the brain into a sweet spot between the extremes.
Rising Costs Are Keeping Healthcare CFOs Up a Night — Can Data Help?
Katie Adams brings us this story in her article in MedCity News. The increasing costs of labor and supplies are the top issues keeping healthcare CFOs up at night, according to a new report released by Syntellis, a performance management software provider. To effectively address these concerns, healthcare providers must ensure they have access to data-driven analytics software that can help them make important decisions — such as managing their existing workforce effectively and where to shift care delivery models — the report said. Survey respondents ranked workforce optimization and productivity monitoring as their chief priority in 2023. Leaders’ top departmental goals for this year are reducing costs, managing improvement initiatives, and managing productivity.
Why it’s important – Even though providers’ margins are currently quite tight, the report recommended that healthcare organizations adopt data analytics software to remedy these concerns and help with important decisions like how to effectively manage their existing workforce and where to shift care delivery models. The report authors cited workforce optimization, effective cost management, continuous performance monitoring, decision support, and understanding market placement as key reasons why the investment is warranted.
Long COVID Now Looks like a Neurological Disease, Helping Doctors to Focus Treatments
The causes of long COVID, which disables millions, may come together in the brain and nervous system, according to an article by Stephani Sutherland, a neuroscientist and science journalist based in southern California, published in the February issue of Scientific American. The most common, persistent, and disabling symptoms of long COVID are neurological. Some are easily recognized as brain- or nerve-related: many people experience cognitive dysfunction due to difficulty with memory, attention, sleep, and mood. Others may seem rooted more in the body than the brain, such as pain and postexertional malaise (PEM), a kind of “energy crash” that people experience after even mild exercise. But those, too, result from nerve dysfunction, often in the autonomic nervous system, which directs our bodies to breathe and digest food and generally runs our organs on autopilot. This so-called dysautonomia can lead to dizziness, a racing heart, high or low blood pressure, and gut disturbances, sometimes leaving people unable to work or even function independently.
One study found that in people with neurological COVID symptoms, the immune system seems to be activated specifically in the central nervous system, creating inflammation. But brain inflammation is probably not caused by the virus infecting that organ directly. Isolated pockets of SARS-CoV-2 or even pieces of viral proteins may remain in the body well after the initial infection and continue to elicit an immune attack. The first solid evidence for “viral persistence” outside the lungs came in 2021 from researchers in Singapore who found viral proteins throughout the gut in five patients who had recovered from COVID as much as six months earlier.
Why it’s important – Although 16 million U.S. sufferers is a reasonable estimate of the condition’s toll, there are other, more dire assessments. A meta-analysis of 41 studies conducted in 2021 concluded that worldwide, 43 percent of people infected with SARS-CoV-2 might develop long COVID, with about 30 percent—translating to approximately 30 million people—affected in the U.S. Some studies have offered more conservative numbers. Determining which patients have ongoing inflammation could help inform treatments. The syndrome is still mired in a lot of medical uncertainty. Patients might have one or a combination of the problems investigated so far: Long COVID might be caused by viral particles that persist in the brain or other parts of the nervous system. Or it might be an autoimmune disorder that lasts long after the virus has disappeared. Maybe overactive immune cells continue to perturb the nervous system and nearby blood vessels. Fortunately, the increasing ability to recognize specific problems is helping clinicians hone treatments that give patients the best chance of recovery.
UnitedHealthcare Launches Rewards Program With Wearable Devices
By simply completing a health survey or meeting step goals, some UnitedHealthcare members can now receive rewards of up to $1,000 per year. Marissa Plescia reports that UnitedHealthcare Rewards members can earn up to $1,000 per year by meeting daily activity goals, tracking sleep, and fulfilling other one-time health activities in her article in MedCity News. The program is available to select employers with fully insured plans, with more employers to be added later in 2023. Those who are enrolled can use an activity tracker, smartphone, or smartwatch. To earn the full $1,000, members must complete all movement goals each day, as well as the one-time activities. The rewards can be added to a prepaid debit card or put into a health savings account. UnitedHealthcare hopes to make the rewards program available to 3 million members by the end of the year and also plans to expand it to self-funded plans in 2024
Why it’s important – While 90% of employees have access to well-being programs from their employer, only 23% are taking advantage of them, a recent study found. This program is designed to change that by spurring engagement through a combination of daily incentives and gamification and by using a broad definition of what it means to pursue health. For employers offering UnitedHealthcare Rewards, they expect the program to help cultivate a culture of wellness, bolster productivity, and help achieve overall better health for their members.
Scientists develop five minute AI respiratory disease test
Oxford University scientists and co-founders of health-tech company Pictura Bio have developed what they say is a world-first diagnostic test, powered by artificial intelligence, that can identify known respiratory viruses such as influenza and COVID-19 within five minutes from just one nasal or throat swab. Their findings, published in the peer-reviewed scientific journal ACS Nano, demonstrate how machine learning can improve the efficiency, accuracy, and time taken to identify different types of viruses and differentiate between strains. This could help better control the spread of respiratory infections and alleviate pressure on the NHS and healthcare staff while reducing medical waste. The technology is now licensed by Pictura Bio, which aims to turn the method into a diagnostic test by creating a dedicated imager and single-use cartridge for use in point-of-care testing, with limited input from the user. The team will also expand the number of viruses the models are trained on and will eventually start looking at other pathogens, such as bacteria and fungi, in respiratory samples, blood, and urine.
Why it’s important – Current tests are either lab-based and time-consuming or fast and less accurate. They are also limited. For instance, a lateral flow test only tests for one infection. This means illnesses are spread as infected people wait for results or because they are unaware they are infected.
5 Best Practices To Build A Future-Ready Digital Health Company
This week’s final article is from Dr. Bertalan Mesko and his team at The Medical Futurist Institute. Transparent communication, clinically-validated technologies, and addressing real-life clinical needs might sound like no-brainer components to a digital health company since they are working in the healthcare sector. It is also a rapidly expanding one where investments were heavily channeled during the pandemic. However, some estimate that 90% of digital health startups will go bust or be ‘acquired’ within a few years of being founded. If digital health represents the future of medicine and healthcare, it is crucial to understand why this is the case. He presents the five features of an excellent digital health company.
Why it’s important – Dr. Mesko and his team have covered this ground before. But this update features examples of companies he believes are doing it right (e.g., AliveCor) and companies that haven’t (e.g., Theranos and Proteus Digital Health).