Sleep apnea retrofit designed by doctors and engineers could help address ventilator shortage

The FDA has been working to adapt its policies and restrictions to respond to the growing need for unconventional solutions like shortages of medical equipment needed for treating COVID-19 patients. A group of doctors, engineers and medical researchers from UC Berkeley, UCSF and working hospitals has devised a creative solution to the ventilator shortage they’re hoping will meet FDA standards for emergency use authorization (EUA), working with readily available hardware and a stockpile of medical breathing equipment that’s resting mostly unused under our noses.

The group, which includes pulmonary care physicians, medical and engineering professors, and many more, is calling themselves the COVID-19 Ventilator Rapid Response Team, and together they’ve figured out a way to modify existing CPAP machines typically used to treat sleep apnea to act as the kinds of ventilators needed for intubation to keep severe COVID-19 patients breathing in the ICU.

Sleep apnea machines are not designed for continuous use with patients who can’t breathe on their own – they basically just ensure that a patient’s airway doesn’t become blocked during sleep, which maintains oxygen levels, and prevents unwanted wake-ups and snoring. The group behind this new CPAP modification has adapted the hardware using a tube that can be used for intubation, led by Dr. Ajay Dharia, a critical care physician focused on pulmonary issues in the ICUs at three Bay Area hospitals as well as an engineering graduate from UC Berkeley.

Already, the FDA has issued guidance stating that healthcare facilities and professionals should consider use of breathing devices not designed for use as ventilators in case of urgent need, so the Ventilator Rapid Response Team already has some leeway in its approach. It’s still seeking an emergency authorization from the agency, however, because it would like to work with suppliers and manufacturers at scale to start producing large quantifies of the modifications required.

It’s also enlisting the help of any individuals or organizations that are looking to donate CPAP or sleep apnea machines that aren’t currently in use to assist with the supply of the base hardware needed to make the modified ventilators. Anyone interested in that can check out their website at https://www.ventilatorsos.com for more info.

NASA details how it plans to establish a sustained human presence on the Moon

NASA’s Artemis program aims to bring humans back to the Moon, with the goal of staying there for good in the interest of pursuing additional science and exploration missions, including to Mars. But how will the agency actually make it possible for people to remain on the Moon for longer-term science missions? NASA has provided some more detail about its plans with a sustainability concept it released describing some core components of the infrastructure it plans to put in place on the lunar surface.

NASA’s plans focus on three key elements that would enable sustained presence and research work on the Moon’s surface, including:

A lunar terrain vehicle (LTV) that would be used by crew to get around on the Moon. Essentially, this is a rover but that is piloted instead of being robotic. This wouldn’t have an enclosed cockpit, so astronauts would be wearing full protective extra-vehicular activity (EVA) spacesuits while using it for short trips.

A habitable mobility platform, which would be a larger rover that is fully contained and pressurized, enabling longer trips further afield from the spacecraft landing site of up to 45 days at a time.

A lunar foundation surface habitat that could act as a more permanent, fixed location home for crew during shorter stays on the surface. this could house up to four astronauts at once, though the habitable mobility platform would be the primary active residence for surface missions, while the Gateway space station orbiting the Moon would be the main base of operations for crew not engaged in active surface exploration and science.

Like the International Space Station before it, the Gateway is designed to be scaled up over time, with new models attached to add more crew habitation capabilities, as well as additional work and experimentation space. This will be important as it becomes the jumping off point not just for Moon surface missions, but also as a way station for exploration of Mars and beyond.

NASA also says that robotic rovers will be a key component of its Moon infrastructure, to be used for purpose including gathering data and materials for research, as well as helping to spur along the development of production of key resources for sustained presence, like water, fuel and oxygen.

The agency also includes some details about its Mars plans, including how it will send a four-person crew to the Gateway for a “multi-month stay to simulate the outbound trip to Mars.” If it goes ahead as planned, this would be longest continuous human stay in deep space environs, and a key step in understanding how a human trip to Mars would work.

The full NASA “Plan for Sustained Lunar Exploration and Development” is available here for more granular detail on the broad outline listed above. Artemis and its timelines are bound to feel the impact of the global coronavirus crisis, but the goals of the program aren’t likely to change too much, even if the targets for accomplishing them do.

Pinterest CEO and a team of leading scientists launch a self-reporting COVID-19 tracking app

There have been a few scattered efforts to leverage crowd-sourced self-reporting of symptoms as a way to potentially predict and chart the progress of COVID-19 across the U.S., and around the world. A new effort looks like the most comprehensive, well-organized and credibly backed yet — and it has been developed in part by Pinterest co-founder and CEO Ben Silbermann.

Silbermann and a team from Pinterest enlisted the help of high school friend, and CRISPR gene-editing pioneer / MIT and Harvard Broad Institute member, Dr. Feng Zhang to build what Silbermann termed in a press release a “bridge between citizens and scientists.” The result is the How We Feel app that Silbermann developed along with input from Zhang and a long list of well-regarded public health, computer science, therapeutics, social science and medical professors from Harvard, Stanford, MIT, Weill Cornell and more.

How We Feel is a mobile app available for both iOS and Android, which is free to download, and which is designed to make it very easy to self-report whether or not they feel well — and if they’re feeling unwell, what symptoms they’re experiencing. It also asks for information about whether or not you’ve been tested for COVID-19, and whether you’re in self-isolation, and for how long. The amount of interaction required is purposely streamlined to make it easy for anyone to contribute daily, and to do so in a minute or less.

The app doesn’t ask for or collect info like name, phone number or email information. It includes an up-front request that users agree to donate their information, and the data collected will be aggregated and then shared with researchers, public health professionals and doctors, including those who are signed on as collaborators with the project, as well as others (and the project is encouraging collaborators to reach out if interested). Part of the team working on the project are experts in the field of differential privacy, and a goal of the endeavor is to ensure that people’s information is used responsibly.

The How We Feel app is, as mentioned, one of a number of similar efforts out there, but this approach has a number of advantages when compared to existing projects. First, it’s a mobile app, whereas some rely on web-based portals that are less convenient for the average consumer, especially when you want continued use over time. Second, they’re motivating use through positive means — Silbermann and his wife Divya will be providing a donated meal to nonprofit Feeding America for every time a person downloads and uses the app for the first time, up to a maximum of 10 million meals. Finally, it’s already designed in partnership with, and backed by, world-class academic institutions and researchers, and seems best-positioned to be able to get the information it gathers to the greatest number of those in a position to help.

How We Feel is organized as an entirely independent, nonprofit organization, and it’s hoping to expand its availability and scientific collaboration globally. It’s an ambitious project, but also one that could be critically important in supplementing testing efforts and other means of tracking the progress and course of the spread of SARS-CoV-2 and COVID-19. While self-reported information on its own is far from a 100% accurate or reliable source, taken in aggregate at scale, it could be a very effective leading indicator of new or emerging viral hotspots, or provide scientific researches with other valuable insights when used in combination with other signals.

Henry Ford Health System to conduct first large US study of hydroxychloroquine’s ability to prevent COVID-19

Despite false assertions by the president to the contrary, any potential treatments to counter or prevent COVID-19 are still only at the stage of early investigations, which include one-off treatment with special individual case authorizations, and small-scale clinical examinations. Nothing so far has approached the level of scrutiny needed to actually say anything definitively about their actual ability to treat COVID-19 or the SARS-CoV-2 virus that causes it, but the first large-scale U.S. clinical study for one treatment candidate is seeking volunteers and looking to get underway.

The study will be conducted by the Henry Ford Health System, which is seeking 3,000 volunteers from healthcare and first responder working environments. Depending on response, the researchers behind the study are looking to begin as early as next week. Study lead researcher Dr. William W. O’Neil said in a press release announcing the study that the goal is to seek a more definitive scientific answer to the question of whether or not hydroxychloroquine might work as a preventative medicine to help protect medical front-line workers with greater risk exposure from contracting the coronavirus.

Hydroxychloroquine (as well as chloroquine) has been in the spotlight as a potential COVID-19 treatment due mostly to repeated name-check that President Trump has given the drug during his daily White House coronavirus task force press briefings. Trump has gone too far in suggesting that the drug, which is commonly used both as an anti-malarial and in the treatment of rheumatoid arthritis and lupus, could be an effective treatment and should be thrust into use. At one point, he claimed that he FDA had granted an emergency approval for its use as a COVID-19 treatment, but Dr. Anthony Fauci clarified that it was not approved for that use, and that clinical studies still need to be performed to evaluate how it works in addressing COVID-19.

Studies thus far around hydroxychloroquine have been small-scale, as mentioned. One, conducted by researchers in France, produced results that indicated the drug was effective in treating those already infected, particularly when paired with a specific antibiotic. Another, more recent study from China, showed that there was no difference in terms of viral duration or symptoms when comparing treatment with hydroxychloroquine with treatment using standard anti-viral drugs, already a common practice in addressing cases of the disease.

This Henry Ford study looks like it could provide better answers to some of these questions around the drug, though the specific approach of seeking to validate prophylactic (preventative) use will mean treatment-oriented applications will still have to be studied separately. The design of the study will be a true blind study, with participants split into three groups that receive “unidentified, specific pills” (possibly anti-virals or some equivalent); hydroxychloroquine; or placebo pills, respectively. They won’t know which they’ve received, and they’ll be contacted weekly by researchers running the study, then in-person both at week four and week eight to determine if they have any symptoms of COVID-19, or any side effects from the medication. They’ll get regular blood draws, and the results will be compared to see if there’s any difference between each cohort in terms of how many contracted COVID-19.

These are front-line healthcare workers, so in theory they should unfortunately be at high risk of contracting the disease. That, plus the large sample size, should provide results that provide much clearer answers about hydroxychloroquine’s potential preventative effects. Even after the study is complete, other competing large-scale trials would ideally be run to prove out or cast doubt on these results, but we’ll be in a better position than we are now to say anything scientifically valid about the drug and its use.

Google and USCF collaborate on machine learning tool to help prevent harmful prescription errors

Machine learning experts working at Google Health have published a new study in tandem with the University of California San Francisco (UCSF)’s computational health sciences department that describes a machine learning model the researchers built that can anticipate normal physician drug prescribing patterns, using a patient’s electronic health records (EHR) as input. That’s useful because around 2 percent of patients who end up hospitalized are affected by preventable mistakes in medication prescriptions, some instances of which can even lead to death.

The researchers describe the system as working in a similar manner to automated, machine learning-based fraud detection tools that are commonly used by credit card companies to alert customers of possible fraudulent transactions: They essentially build a baseline of what’s normal consumer behavior based on past transactions, and then alert your bank’s fraud department or freeze access when they detect a behavior that is not in line with and individual’s baseline behavior.

Similarly, the model trained by Google and UCSF worked by identifying any prescriptions that “looked abnormal for the patient and their current situation.” That’s a much more challenging proposition in the case of prescription drugs, vs. consumer activity – because courses of medication, their interactions with one another, and the specific needs, sensitivities and conditions of any given patient all present an incredibly complex web to untangle.

To make it possible, the researchers used electronic health records from de-identified patient that include vital signs, lab results, prior medications and medical procedures, as well as diagnoses and changes over time. They paired this historical data with current state information, and came up with various models to attempt to output an accurate prediction of a course of prescription for a given patient.

Their best-performing model was accurate “three quarters of the time,” Google says, which means that it matched up with what a physician actually decided to prescribe in a large majority of cases. It was also even more accurate (93%) in terms of predicting at least one medication that would fall within a top ten list of a physician’s most likely medicine choices for a patient – even if its top choice didn’t match the doctor’s.

The researchers are quick to note that though the model thus far has been fairly accurate in predicting a normal course of prescription, that doesn’t mean it’s able to successfully detect deviations from that yet with any high degree of accuracy. Still, it’s a good first step upon which to build that kind of flagging system.

Estimote launches wearables for workplace-level contact tracing for COVID-19

Bluetooth location beacon startup Estimote has adapted its technological expertise to develop a new product designed specifically at curbing the spread of COVID-19. The company created a new range of wearable devices that co-founder Steve Cheney believes can enhance workplace safety for those who have to be colocated at a physical workplace even while social distancing and physical isolation measures are in place.

The devices, called simply the “Proof of Health” wearables, aim to provide contact tracing – in other words, monitoring the potential spread of the coronavirus from person-to-person – at the level of a local workplace facility. The intention is to give employers a way to hopefully maintain a pulse on any possible transmission among their workforces and provide them with the ability to hopefully curtail any local spread before it becomes an outsized risk.

The hardware includes passive GPS location-tracking, as well as proximity sensors powered by Bluetooth and ultra-wide band radio connectivity, a rechargeable battery, and built-in LTE. It also includes a manual control to change a wearer’s health status, recording states like certified health, symptomatic, and verified infected. When a user updates their state to indicate possible or verified infection, that updates others they’ve been in contact with based on proximity and location-data history. This information is also stored in a health dashboard that provides detailed logs of possible contacts for centralized management. That’s designed for internal use within an organization for now, but Cheney tells me he’s working now to see if there might be a way to collaborate with WHO or other external health organizations to potentially leverage the information for tracing across enterprises and populations, too.

These are intended to come in a number of different form factors: the pebble-like version that exists today, which can be clipped to a lanyard for wearing and displaying around a person’s neck; a wrist-worn version with an integrated adjustable strap; and a card format that’s more compact for carrying and could work alongside traditional security badges often used for facility access control. The pebble-like design is already in production and 2,000 will be deployed now, with a plan to ramp production for as many as 10,000 more in the near future using the company’s Poland-based manufacturing resources.

Estimote has been building programmable sensor tech for enterprises for nearly a decade and has worked with large global companies, including Apple and Amazon . Cheney tells me that he quickly recognized the need for the application of this technology to the unique problems presented by the pandemic, but Estimote was already 18 months into developing it for other uses, including in hospitality industries for employee safety/panic button deployment.

“This stack has been in full production for 18 months,” he said via message. “We can program all wearables remotely (they’re LTE connected). Say a factory deploys this – we write an app to the wearable remotely. This is programmable IoT.

“Who knew the virus would require proof of health vis-a-vis location diagnostics tech,” he added.

Many have proposed technology-based solutions for contact tracing, including leveraging existing data gathered by smartphones and consumer applications to chart transmission. But those efforts also have considerable privacy implications, and require use of a smartphone – something that Cheney says isn’t really viable for accurate workplace tracking in high-traffic environments. By creating a dedicated wearable, Cheney says that Estimote can help employers avoid doing something “invasive” with their workforce, since it’s instead tied to a fit-for-purpose device with data shared only with their employers, and it’s in a form factor they can remove and have some control over. Mobile devices also can’t do nearly as fine-grained tracking with indoor environments as dedicated hardware can manage, he says.

And contact tracing at this hyperlocal level won’t necessarily just provide employers with early warning signs for curbing the spread earlier and more thoroughly than they would otherwise. In fact, larger-scale contact tracing fed by sensor data could inform new and improved strategies for COVID-19 response.

“Typically, contact tracing relies on the memory of individuals, or some high-level assumptions (for example, the shift someone worked),” said Brianna Vechhio-Pagán of John Hopkins University’s Applied Physics Lab via a statement. “New technologies can now track interactions within a transmissible, or ~6-foot range, thus reducing the error introduced by other methods. By combining very dense contact tracing data from Bluetooth and UWB signals with information about infection status and symptoms, we may discover new and improved ways to keep patients and staff safe.”

With the ultimate duration of measures like physical distancing essentially up-in-the-air, and some predictions indicating they’ll continue for many months, even if they vary in terms of severity, solutions like Estimote’s could become essential to keeping essential services and businesses operating while also doing the utmost to protect the health and safety of the workers incurring those risks. More far-reaching measures might be needed, too, including general-public-connected, contact-tracing programs, and efforts like this one should help inform the design and development of those.

Virgin Orbit announces new plans for first Asian spaceport in Oita, Japan

Virgin Orbit may be focusing its production efforts right now on making ventilators to support healthcare workers battling COVID-19, but it’s also still making moves to build out the infrastructure that will underpin its small satellite launch business. To that end, the new space company unveiled a new partnership with Oita Prefecture in Japan to build a new spaceport there from which to launch and land its horizontal take-off launch vehicle carrier aircraft.

Working in collaboration with ANA Holdings and the Space Port Japan Association, Virgin Orbit says it is currently targeting Oita Airport as the site for its next launch site – the first in Asia – with a plan to start flying missions from the new location as early as 2022.

There are still a number of steps that have to take place before the Oita airport becomes official – including performing a technical study in partnership with local government to determine the feasibility of using the proposed site. Already, Oita is home to facilities from a number of corporations including Toshiba, Nippon Steel, Canon, Sony, Daihatsu and more, but this would marks its first entry into the space industry, an area where Oita is hoping to encourage in future.

“We are eager to host the first horizontal takeoff and landing spaceport in Japan. We are also honored to be able to collaborate with brave technology companies solving global-level problems through their small satellites,” said Katsusada Hirose, Governor for the Oita Prefectural Government, in a press release. “We hope to foster a cluster of space industry in our prefecture, starting with our collaboration with Virgin Orbit.”

Virgin Orbit is looking to scale its efforts globally in a number of ways, even as it gears up for a first demonstration launch of its orbital small satellite delivery capabilities sometime later this year. The company announced plans to provide launch services from a forthcoming spaceport facility in Cornwall for the UK market, and it’s also looking at standing up a site in Guam.

The horizontal launch model that Virgin Orbit uses means that it can much more easily leverage traditional airport infrastructure and processes to set up launch sites, and doing so can provide domestic launch capabilities essentially on-demand for countries looking to add small satellite flight to their in-country housed services. That’s a big selling point, and Oita securing should be a considerable win and for Japan as the site of a first Virgin Orbit port across the whole continent.

NASA issues agency-wide crowdsourcing call for ideas around COVID-19 response

There’s crowdsourcing a problem, and then there’s crowdsourcing a problem within NASA, where some of the smartest, most creative and resourceful problem-solvers in the world solve real-world challenges daily as part of their job. That’s why it’s uplifting to hear that NASA has issued a call to its entire workforce to come up with potential ways the agency and its resources can contribute to the ongoing effort to fight the current coronavirus pandemic.

NASA is using its crowdsourcing platform NASA @ WORK, which it uses to internally source creative solutions to persistent problems, in order to collect creative ideas about new ways to address the COVID-19 crisis and the various problems it presents. Already, NASA is engaged in a few different ways, including offering supercomputing resources for treatment research, and working on developing AI solutions that can help provide insight into key scientific investigations that are ongoing around the virus.

There is a degree of specificity in the open call NASA put to its workforce: It identified key areas where solutions are most urgently needed, working together with the White House and other government agencies involved in the response, and determined that NASA staff efforts should focus on addressing shortfalls and gaps in the availability of personal protective equipment, ventilation hardware and ways to monitor and track the coronavirus spread and transmission. That’s not to say NASA doesn’t want to hear solutions about other COVID-19 issues, just that these are the areas where they’ve identified the most current need.

To add some productive time-pressure to this endeavor, NASA is looking for submissions from staff on all the areas above to be made via NASA @ WORK by April 15. Then there’ll be a process of assessing what’s most viable, and allocating resources to make those a reality. Any products or designs that result will be made “open source for any business or country to use,” the agency says — with the caveat that this might not be strictly possible in all cases depending on the specific technologies involved.

Researchers to study if startup’s wrist-worn wearable can detect early COVID-19 respiratory issues

It’s highly unlikely that the current coronavirus crisis will neatly and fully “solved” by any one endeavor or solution, which makes news studies like one involving startup WHOOP’s wrist-worn fitness and health tracking wearable all the more important. The study, conducted by the Central Queensland University Australia (CQUniversity), in partnership with the Cleveland Clinic, will employ data collected WHOOP’s hardware with hundreds of volunteers who have self-identified as having contracted COVID-19 to study changes in their respiratory behavior over time.

The data to be used for this study has been collected from WHOOP’s 3.0 hardware, which has also recently been validated by a University of Arizona external study conducted specifically to determine the accuracy of its measurement of respiratory rates during sleep, which the device uses to provide quality of sleep scores to its users. That study showed it to be among the most accurate measurement tools for respiratory rate short of invasive procedures, which is what has led researchers behind this new study to hypothesize that it could be valuable as a sort of early-warning system for detecting signs of abnormal respiratory behavior in COVID-19 patients before those symptoms are detectable by other means.

The WHOOP team says that the respiratory rate its hardware reports very rarely deviates from an established individual baseline, and that when it does so, it’s usually due to either one of two causes: environmental factors, like unusually high temperatures or significant differences in oxygen concentration, or something happening within the body, like a lower-respiratory tract infection.

COVID-19 is specifically a lower-respiratory tract infection, unlike the flue or the cold, which are upper-respiratory issues. That means there’s a strong correlation between rate changes due to lower-respiratory tract issues not accounted by environmental problems (which are relatively easy to cancel out) and instances of COVID-19. And since the WHOOP wearable is designed to look for deviations as a sign of distress, among the other sings it monitors, it could notice changes to respiratory rates relative to baselines before an individual becomes aware of any significant shortness of breath themselves.

This is a study, so at this point that’s just a hypothesis, and will need to be backed up by data. The team behind it says it should take around six weeks, and there are an “initial several hundred self-reported COVID-19 cases” already present in the app from which it will begin, with a target of enrolling at least 500 individuals with positive COVID-19 test results. There are also other investigations underway to see if wearables that monitor a user’s health and fitness can provide early warning systems for potential COVID-19 cases, including a study being conducted by UCSF using the Oura Ring.

Unlike with previous pandemics, the current coronavirus crisis comes at a time when we’re increasingly used to taking data-driven approaches to solving challenges, and when we also have a lot of self-quantifying health devices in circulation. Those could help us get a better grip on assessing the spread, as well as trends related to how it circulates and ebbs/grows within a population.

DeepMind’s Agent57 AI agent can best human players across a suite of 57 Atari games

Development of artificial intelligence agents tends to frequently be measured by their performance in games, but there’s a good reason for that: Games tend to offer a wide proficiency curve, in terms of being relatively simple to grasp the basics, but difficult to master, and they almost always have a built-in scoring system to evaluate performance. DeepMind’s agents have tackled board game Go, as well as real-time strategy video game StarCraft – but the Alphabet company’s most recent feat is Agent57, a learning agent that can beat the average human on each of 57 Atari games with a wide range of difficulty, characteristics and gameplay styles.

Being better than humans at 57 Atari games may seem like an odd benchmark against which to measure the performance of a deep learning agent, but it’s actually a standard that goes all the way back to 2012, with a selection of Atari classics including Pitfall, Solaris, Montezuma’s Revenge and many others. Taken together, these games represent a broad range of difficulty levels, as well as requiring a range of different strategies in order to achieve success.

That’s a great type of challenge for creating a deep learning agent because the goal is not to build something that can determine one effective strategy that maximizes your chances of success every time you play a game – instead, the reason researchers build these agents and set them to these tasks at all is to develop something that can learn across multiple and shifting scenarios and conditions, with the long-term aim of building a learning agent that approaches general AI – or AI that is more human in terms of being able to apply its intelligence to any problem put before it, including challenges it’s never encountered before.

DeepMind’s Agent57 is remarkable because it performs better than human players on each of the 57 games in the Atari57 set – previous agents have been able to be better than human players on average – but that’s because they were extremely good at some of the simpler games that basically just worked via a simple action-reward loop, but terrible at games that required more advanced play, including long-term exploration and memory, like Montezuma’s Revenge.

The DeepMind team addressed this by building a distributed agent with different computers tackling different aspects of the problem, with some tuned to focus on novelty rewards (encountering things they haven’t encountered before), with both short- and long-term time horizons for when the novelty value resets. Others sought out more simple exploits, figuring out which repeated pattern provided the biggest reward, and then all the results are combined and managed by an agent equipped with a meta-controller that allows it to weight the costs and benefits of different approaches based on which game it encounters.

In the end, Agent57 is an accomplishment, but the team says it can stand to be improved in a few different ways. First, it’s incredibly computationally expensive to run, so they will seek to streamline that. Second, it’s actually not as good at some of the simpler games as some simpler agents – even though it excels at the the top 5 games in terms of challenge to previous intelligent agents. The team says it has ideas for how to make it even better at the simpler games that other, less sophisticated agents, are even better at.