What happened in health care technology this week, and why it’s important.
Experimental insulin implant uses electricity to control genes
Genetically engineered human cells that produce insulin when stimulated by a small electric current could one day be used to develop better treatments for type 1 diabetes. Lilly Tozer reports on the development in Nature. Researchers generated cells that undergo a chain reaction in response to reactive oxygen species (ROS) — unstable oxygen-containing radicals produced when a current is applied — that ultimately switch on the gene needed to make insulin. In a proof-of-concept experiment, they implanted the engineered cells into mice and showed that the cells released insulin when a current was applied using electrified acupuncture needles.
Why it’s important – The findings, published in Nature Metabolism on July 31st, offer hope that this technology could one day be incorporated into medical implants. The team hopes that one day this system could be adapted into wearable medical devices controlled by a computer or smartphone. But the technology is still at a very early stage, and more work is needed before it can be tested in people.
Infographic of the week – Amazon’s Healthcare Flywheel: Amazon’s activity and movement toward critical mass in healthcare is picking up steam. The retail e-commerce giant has made several plays across AWS and cloud, but also, interestingly, in care delivery (both virtual and in-person) and pharmacy. If they’re not already, Amazon is headed toward being a consumer-focused healthcare platform, making long-term investments and product launches into the space (Clinic, PillPack → Pharmacy, One Medical). On the other side of the coin, AWS for Health, Amazon’s enterprise cloud play, has been increasingly focused on providing the picks, shovels, and Lego bricks for healthcare software developers.
Mass General Brigham bets big on hospital-at-home
Mass General Brigham sees hospital-at-home care as a big part of its long-term future. As Diane Eastbrook reports in Modern Healthcare online, The Boston-based provider says it is on track to shift 10% of inpatient care to hospital-at-home—through which acute care is delivered in-home and virtually, and patients are connected to remote monitoring—within five years. In the coming weeks, Mass General Brigham anticipates word from CMS about regulatory waivers enabling that expansion by OK’ing Medicare reimbursements for these services that match payments for inpatient care. Mass General Brigham would not estimate how much its hospital-at-home development will cost or outline its profit expectations. The current program is “breaking even,” said Heather O’Sullivan, president of Mass General Brigham Healthcare at Home.
Why it’s important – Escalating demand from an aging population with greater healthcare needs is driving the initiative. The hospital-at-home program is one component of a larger strategy to meet that demand. Medicare reimbursement remains the wild card. There are no guarantees CMS will continue to reimburse hospital-at-home services at parity with inpatient care. CMS is still collecting data from the more than 400 hospitals in the waiver program to determine whether Medicare will continue its current payment policy or devise a new one. Whatever CMS decides will likely provide a roadmap for private health insurance companies.
Podcast episode of the week – From Creating a New Healthcare podcast series: “A novel virtual care platform supporting patient access and population health – with Lyle Berkowitz MD, CEO & Founder, KeyCare” In this interview, we’re going to discover how this novel platform is attempting to solve the issues of access to care, capacity, cost-effectiveness, and burnout amongst providers and their teams. Throughout the interview, Dr. Berkowitz illustrates the numerous ways that healthcare systems, as well as other provider groups, can utilize and leverage KeyCare. You can listen to the podcast here.
AI model can help determine where a patient’s cancer arose
Anne Trafton from MIT News reviews a new approach developed by researchers at MIT and Dana-Farber Cancer Institute that may make it easier to identify the sites of origin for certain enigmatic cancers. Using machine learning, the researchers created a computational model to analyze the sequence of about 400 genes and use that information to predict where a given tumor originated in the body. Using this model, the researchers showed that they could accurately classify at least 40 percent of tumors of unknown origin with high confidence in a dataset of about 900 patients. This approach enabled a 2.2-fold increase in the number of patients who could have been eligible for a genomically guided, targeted treatment based on where their cancer originated.
Why it’s important – For a small percentage of cancer patients, doctors cannot determine where their cancer originated. This makes it much more challenging to choose a treatment for those patients because many cancer drugs are typically developed for specific cancer types. In 3 to 5 percent of cancer patients, particularly in cases where tumors have metastasized throughout the body, oncologists don’t have an easy way to determine where the cancer originated. These tumors are classified as cancers of unknown primary (CUP). The researchers now hope to expand their model to include other data types, such as pathology images and radiology images, to provide a more comprehensive prediction using multiple data modalities. This would also provide the model with a comprehensive perspective of tumors, enabling it to predict not just the type of tumor and patient outcome but potentially even the optimal treatment.
YouTube video of the week – From The Medical Futurist Institute: “One Ring to Care for All”: At The Medical Futurist, we’ve had the opportunity to review countless digital health technologies, wearables, sensors, and services. But believe it or not, we never had a smart ring – until now. Now, we’ve added a new trophy to that list: RingConn, a smart ring that promises seamless health tracking.
From windows to wall art, hospitals use virtual reality to design more inclusive rooms for kids
For many young patients, harsh lights, bare walls, and windows facing parking lots or brick buildings make painful hospital visits more unpleasant, stoking fear and uncertainty instead of hope. Often, those patients say, it makes recovery more challenging. As Mohana Ravindranath reports in Stat, their perspectives — historically overlooked in hospital design — are at the heart of a budding movement to make architecture more inclusive for the people who actually spend time there. Hospital groups like UCSF Benioff Children’s and Boston Children’s are exploring ways to fold young patients’ feedback into hospital design, like the color of walls and the placement of windows, art, and couches.
COURTESY XRLAB BERKELEY
In the basement of a gray brutalist campus building, Berkeley’s “extended reality” or XR lab is stocked with virtual reality headsets and a pristine white hospital bed. The study has recruited roughly 30 previously hospitalized children to explore mocked-up hospital rooms in virtual reality. Participants use handheld controllers to flip through virtual hospital rooms, using a button to toggle between viewpoints from the bed or beside it. By tilting their heads, they can get a 360-degree view. In one room, a window faces into the hallway — a feature some like and others find invasive. Some rooms have different-sized windows facing the sky; some have warm-colored paintings hanging on different walls.
Why it’s important – The project could inform UCSF Benioff Children’s plans for a new inpatient wing. It’s a slightly higher-tech version of a similar undertaking at Boston Children’s Hospital, where architects designing a new building made a cardboard model of an exam room that children with developmental diagnoses, their parents, and staff could walk into. Their feedback was sometimes surprising: a bright, cheerful color palette was actually too intense, risking overstimulation. Lightweight, smooth chairs without sharp corners meant children wouldn’t get hurt, but they were also easier to pick up and throw, so staff suggested weighing them down with sand. In addition to easing stress, inclusive design makes care more efficient: A quiet waiting room for kids who find the usual one too overwhelming means they’re more likely to come to their appointments; making scans less frightening means kids won’t want to avoid them.
It’s high time for more AI transparency
Tech companies are rushing to release their AI models into the wild, and we’re seeing generative AI embedded in more and more products. But the most powerful models out there, such as OpenAI’s GPT-4, are tightly guarded by their creators. Developers and researchers pay for limited access to such models through a website and don’t know the details of their inner workings. This opacity could lead to problems down the line, as is highlighted in a new, non-peer-reviewed paper that caused some buzz last week. Researchers at Stanford University and UC Berkeley found that GPT-3.5 and GPT-4 performed worse at solving math problems, answering sensitive questions, generating code, and doing visual reasoning than they had a few months earlier.
Why it’s important – This has some serious implications. Companies that have built and optimized their products to work with a specific iteration of OpenAI’s models could “100%” see them suddenly glitch and break. When OpenAI fine-tunes its models this way, products that have been built using particular prompts, for example, might stop working like they did before. Ultimately, the open vs. closed debate around AI boils down to who calls the shots. With open models, users have more power and control. With closed models, you’re at the mercy of their creator.