News roundup: Proteus may be no-teous, DOJ leads on Google-Fitbit, HHS’ mud fight, Leeds leading in health tech, malware miseries, comings and goings

Proteus stumbles hard, cuts back. The original ‘tattle-tale pill’ company, Proteus Digital Health, plans to lay off 292 people in the San Francisco Bay Area and to permanently close its three Redwood City and Hayward locations, starting 18 January, according to notices sent to California state and local offices, including the state employment development department. It is unclear where Proteus will be located after the closures.

This followed after Proteus failed to launch a twelfth funding round of $100 million. According to reports, they furloughed most of their employees for two weeks in November and are reorganizing. This is after a substantial number of investors have put in about $487M in funding through a Series H (Crunchbase), including a game-changing investment by Novartis dating back to 2010.  Proteus achieved unicorn status about three years ago, but its high-priced pill tracking technology with a pill sensor tracked by a skin-worn monitor reporting into a smartphone has a built-in limited market to expensive medication. Otsuka Pharmaceutical in 2017 partnered with Proteus for an FDA-cleared digital medicine system called Abilify MyCite that basically put an off-patent behavioral drug back into a more expensive tracking methodology. But Proteus remains a great idea on tracking compliance in search of a real market, and may not have much of a future. San Jose Mercury News, CNBC

But ingestible detectable pills are still being tested. On Monday, as Proteus’ bad news broke, eTectRx announced its FDA clearance of the ID-Cap System and its testing at Brigham and Women’s Hospital and Fenway Health, focusing on HIV medication when used for treatment and prevention. Release, HISTalk

Department of Justice taking the lead on scrutinizing Google’s Fitbit acquisition. The Federal Trade Commission also sought jurisdiction over the transaction. According to the New York Post, “both agencies are concerned that a Google-owned Fitbit would give the search giant an even bigger window into people’s private data, including sensitive health information, sources said. Under the Hart-Scott-Rodino Act, all large mergers must file proposals with both the DOJ and the FTC, but only one antitrust agency reviews the merger.”

Coal from stockings being thrown about at HHS. According to POLITICO and the New York Times, the disagreements between Seema Verma, the head of the Centers for Medicare and Medicaid Services (CMS), and the Cabinet-level Secretary of Health and Human Services (HHS), Alex Azar, have boiled over, enough to have to be settled by the President’s acting chief of staff, Mick Mulvaney. According to the Times, both President Trump and VP Mike Pence have told them to find a way to work together. Both are administration appointees, but President Trump has not been reluctant to cut a mis-performing or overly contrary appointee loose. The latest salvo from those obviously not on Ms. Verma’s side was the revelation that she requested compensation for jewelry stolen on a business trip, contrary to government policy of course. She was compensated for other items which is standard. (Isn’t that what homeowners’ insurance is for? And what sensible person actually travels with valuable jewelry?) Under Ms. Verma, CMS has been quite progressive in developing new business models in Medicare fee-for-service, moving providers to two-sided risk, and innovating in both Medicare and Medicaid. It will either be settled, or one or both will be gone. Pass the popcorn.

Leeds picks up another health tech company. Mindwave Ventures is opening an office there, as well as appointing Dr Victoria Betton and Dr Janak Gunatilleke to the roles of chief innovation officer and chief operating officer. Mindwave develops technologies around digital products and services in healthcare and health research. Leeds reportedly is home to over 250 health tech companies and holds an annual Leeds Digital Festival in the spring [TTA 11 April].

Ransomware attack hits Hackensack Meridian. Systems were down for about a week. While this large New Jersey health system hasn’t admitted it, sources told the Asbury Park Press that it was ransomware. And if it’s not ransomware, its Emotet and Trickbot. Read ZDNet and be very apprehensive for 2020, indeed, as apparently healthcare is just one big target.

Comings and Goings: There may be some end of year bombshells, but after last week’s big news about John Halamka, it’s been fairly quiet. Paul Walker, whom this Editor knew at New York eHealth Collaborative, has joined CommonWell Health Alliance as executive director. Mr. Walker was most recently Philips Interoperability Solutions’ vice president of strategy and business development. CommonWell’s goal is improving healthcare interoperability and its services are used by more than 15,000 care provider sites nationwide. Blog release, Healthcare Innovation ….Dr. Jacqueline Shreibati, the chief medical officer for AliveCor, is joining Google Health in the health research area. Mum’s the word when it comes to Fitbit (see above). CNBC ….Peter Knight has pleaded guilty to falsifying educational credentials to gain his position as chief information and digital office at Oxford University Hospitals NHS Foundation Trust. He held that position from August 2016 until September 2018. BBC News

News roundup: Philips allies with Humana for pop health, Dexcom’s outage outrage, Halamka ankles Lahey for Mayo, Google and NHS Wales changes, Agfa’s health sale, Victrix/WhatsApp, more

Insurer Humana is identifying high-acuity and chronic CHF Medicare Advantage members and deploying two support programs utilizing Philips PERS and remote patient monitoring (RPM) systems. The first program identifies at-risk older people with chronic conditions and offering them Philips Lifeline with AutoAlert, Lifeline’s fall detection technology, and their CareSage predictive analytics. Philips Lifeline is already offered in select Humana Medicare Advantage plans. The second is a pilot with telehealth RPM to monitor a select group of CHF patients. This will use a Philips interactive tablet and connected measurement devices for care teams to actively monitor congestive heart failure patients. The rationale in the press release is centered on population health management, quality of care, and positively influencing patient outcomes, with “more efficient resource utilization” a/k/a lowering cost of care. Philips release.

Health tech is great, when it works–and Dexcom found out how serious it can get when it doesn’t. Dexcom, a continuous glucose monitoring system, experienced a server outage over the US Thanksgiving holiday weekend into Monday. It knocked out its updates in the Follow feature, frequently used by parents to monitor Type 1 diabetic children, and those with artificial pancreas devices that adjust insulin based on monitored BG levels. Dexcom was not only blasted by users on the server outage, which they attributed to ‘overload’, but also on its communications of the problem to users which depended on Facebook postings and not on real-time direct contacts or messaging. It was a ‘big surprise’ to their CEO, who also dismissed the possibility of a data breach, which seems a bit premature. Both Google and Microsoft provide cloud and tech services to Dexcom. CNBC 12/2, 12/3

Comings and goings: HIT pioneer, strategist, and general guru John Halamka is following the AI Star, leaving Boston’s Beth Israel Lahey Health to head up a machine learning/AI initiative at the Mayo Clinic in Minneapolis. Mayo this fall announced a 10-year high-level partnership with Google Cloud to store patient data and analysis. Modern Healthcare  According to the Healthcare IT News article, he’ll be returning on weekends to the Bay State to his 250-acre working farm….Also moving on to Google Health is Facebook’s Hema Budaraju, a product management director. Business Insider has annoyingly hid the news behind its paywall, leading to speculation in Mobihealthnews that she will be engaged in Google’s “social and environmental impact” efforts as she was at FB…And speaking of Google, founders Larry Page and Sergey Brin are stepping down at long last from active management. Google CEO Sundar Pichai will now be running Google and its corporate parent, Alphabet. See their letter on the GoogleBlog.DigitalHealth reports on changes at the NHS Wales Informatics Service. Helen Thomas is now interim director as NWIS director Andrew Griffiths is departing this month. NWIS is also transitioning to a new Special Health Authority….Agfa’s Healthcare Information Solutions and Integrated Care, plus their imaging division, are definitely going to Italy’s Dedalus Holdings S.p.A. for €975 million. It awaits approval from various authorities, their employee groups, and the usual closing conditions. Release, DigitalHealth.

UK healthcare analytics company Victrix Socsan has signed a licensing agreement last month with WhatsApp. Victrix will use Whats App for communications with beneficiaries as part of their furnishing proactive preventive care services and provide secure information. Release.

Google’s ‘Project Nightingale’–a de facto breach of 10 million health records, off a bridge too far?

Breaking News. Has this finally blown the lid off Google’s quest for data on everyone? This week’s uncovering, whistleblowing, and general backlash on Google’s agreement with Ascension Health, the largest non-profit health system in the US and the largest Catholic health system on the Planet Earth, revealed by the Wall Street Journal (paywalled) has put a bright light exactly where Google (and Apple, Facebook, and Amazon), do not want it.

Why do these giants want your health data? It’s all about where it can be used and sold. For instance, it can be used in research studies. It can be sold for use in EHR integration. But their services and predictive data is ‘where it’s at’. With enough accumulated data on both your health records and personal life (e.g. not enough exercise, food consumption), their AI and machine learning modeling can predict your health progression (or deterioration), along with probable diagnosis, outcomes, treatment options, and your cost curve. Advertising clicks and merchandising products (baby monitors, PERS, exercise equipment) are only the beginning–health systems and insurers are the main chance. In a worst-case and misuse scenario, the data modeling can make you look like a liability to an employer or an insurer, making you both unemployable and expensively/uninsurable in a private insurance system.

In Google’s latest, their Project Nightingale business associate agreement (BAA) with Ascension Health, permissible under HIPAA, allowed them apparently to access in the initial phase at least 10 million identified health records which were transmitted to Google without patient or physician consent or knowledge, including patient name, lab results, diagnoses, hospital records, patient names and dates of birth. This transfer and the Google agreement were announced by Ascension on 11 November. Ultimately, 50 million records are planned to be transferred from Ascension in 21 states. According to a whistleblower on the project quoted in The Guardian, there are real concerns about individuals handling identified data, the depth of the records, how it’s being handled, and how Google will be using the data. Ascension doesn’t seem to share that concern, stating that their goal is to “optimize the health and wellness of individuals and communities, and deliver a comprehensive portfolio of digital capabilities that enhance the experience of Ascension consumers, patients and clinical providers across the continuum of care” which is a bit of word salad that leads right to Google’s Cloud and G Suite capabilities.

This was enough to kick off an inquiry by Health and Human Services (HHS). A spokesperson confirmed to Healthcare Dive that “HHS’ Office of Civil Rights is opening an investigation into “Project Nightingale.” The agency “would like to learn more information about this mass collection of individuals’ medical records with respect to the implications for patient privacy under HIPAA,” OCR Director Roger Severino said in an emailed statement.”

Project Nightingale cannot help but aggravate existing antitrust concerns by Congress and state attorneys general on these companies and their safeguards on privacy. An example is the pushback around Google’s $2.1 bn acquisition of Fitbit, which one observer dubbed ‘extraordinary’ given Fitbit’s recent business challenges, and data analytics company Looker. DOJ’s antitrust division has been looking into how Google’s personalized advertising transactions work and increasingly there are calls from both ends of the US political spectrum to ‘break them up.’ Yahoo News

Google and Ascension Health may very well be the ‘bridge too far’ that curbs the relentless and largely hidden appetite for personal information by Google, Amazon, Apple, and Facebook that is making their very consumers very, very nervous. Transparency, which seems to be a theme in many of these articles, isn’t a solution. Scrutiny, oversight with teeth, and restrictions are.

Also STAT News , The Verge on Google’s real ambitions in healthcare, and a tart take on Google’s recent lack of success with acquisitions in ZDNet, ‘Why everything Google touches turns to garbage’. Healthcare IT News tries to be reassuring, but the devil may be in Google’s tools not being compliant with HIPAA standards.  Further down in the article, Readers will see that HIPAA states that the agreement covers access to the PHI of the covered entity (Ascension) only to have it carry out its healthcare functions, not for the business associate’s (Google’s) independent use or purposes. 

International news roundup: Verily ‘eyes’ India, Oxford VR trials in Hong Kong, Israel-Finland collaborate, Blue Cedar-Blackberry partner, NuvoAir inhales $3M

Verily rolls along in India with eye health research. Spending its fresh $1bn in funding [TTA 17 Jan], Verily and Alphabet sibling Google are researching machine learning in screening for diabetic retinopathy and diabetic macular edema. The retinal diagnostic program is at Aravind Eye Hospital, a network of eye hospitals headquartered in Tamil Nadu. Early research at Madurai has demonstrated that the algorithm performs on par with general ophthalmologists and retinal specialists assessing the images for disease. The estimated shortage of eye doctors in India is estimated at over 100,000. The stated objective of the research is to detect retinal disease earlier and expand access to screening. Further testing is verifying safety and effectiveness. Verily has received the CE Mark for the algorithm. Verily blog

Oxford VR in pilot to manage mental health in Hong Kong. Oxford VR, AXA Hong Kong and The Chinese University of Hong Kong (CUHK) have partnered for an “immersive therapy” trial using VR. This will focus on managing common mental health issues, such as social avoidance, anxiety and depressive symptoms. The scenarios will be in both English and Cantonese. Details will be announced in June. Mobihealthnews, Oxford VR press release

Israel and Finland are launching a joint collaboration announced last December, led by the Israel Innovation Authority (IIA) and the Helsinki Business Hub. The pilot program is starting with a Call for Proposals for joint projects. The two countries will co-develop, test, improve and pilot technologies, products and services in digital health, smart mobility and information and communication technologies. According to Mobihealthnews, the countries both have strong commonalities in public health systems with coordinated data sets, digital health, genomic projects, and advanced interoperability. Also The Jewish Voice

App security innovator Blue Cedar partners with Blackberry. Blue Cedar’s securing data from the app to the provider location will work with BlackBerry Dynamics-enabled mobile apps so that development teams will not be required to manually change source code or add libraries. Both Blue Cedar and Blackberry target international high-security fields such as healthcare, government, and financial. Blue Cedar blog

Swedish respiratory device/app NuvoAir inhales $3M in new funding. NuvoAir targets those managing asthma, COPD, and cystic fibrosis who need to monitor their breathing and lung health in conjunction with their overall health. Last year, the company developed the Air Next system, a portable device that connects via Bluetooth LE to users’ smartphones or tablets. Air Next has a CE Mark. NuvoAir plans to expand into Europe, North America, and Asia. The digital respiratory therapy area is booming, with an estimated 210 million asthma and COPD patients with smartphones worldwide a strong factor. One sign: Propeller Health’s premium acquisition by ResMed.  Mobihealthnews

Verily, Google’s life sciences arm, gathers in another billion to go…where? (updated for Study Watch FDA clearance)

Biotech/device company Verily added to its 2016 $800 million stake from Singapore’s Temasek a fresh $1 bn from Silver Lake Partners. with reported participation from Ontario Teacher’s Pension Plan. Verily is majority-owned by Google parent Alphabet, which has added a new member to the Verily board, CFO Ruth Porat, and Egon Durbat from Silver Lake.

CEO Andrew Conrad, who is still there despite a brace of bad press two years ago [TTA 6 Apr 16], stated that “We are taking external funding to increase flexibility and optionality as we expand on our core strategic focus areas. Adding a well-rounded group of seasoned investors, led by Silver Lake, will further prepare us to execute as healthcare continues the shift towards evidence generation and value-based reimbursement models.”

One is tempted to say, ‘whatever that means’. They have had multiple ventures from contact lenses with Novartis’ subsidiary Alcon (reportedly discontinued but dating back with Google to 2014), diabetes with Sanofi, to sleep apnea with ResMed. VentureBeat reports they are cash-profitable and even venturing into areas such as small exploding needles that can extract blood through a wearable device–not precisely for the needle-phobic. There seem to be multiple projects in multiple directions that are primarily research. Certainly their finding at $1.8 bn is an outlier even at 2018’s big scale–but with Alphabet/Google as a parent and A-list partners, the risk is minimal. Mobihealthnews, Crunchbase

FDA clearance of Verily’s Study Watch. Late last week, Verily announced that their Study Watch was given a 510(k) FDA clearance. It records, stores, transfers and displays single-channel ECG. To date, there are no plans to use it beyond a handful of research studies primarily on cardiac disease. Mobihealthnews. Meanwhile, Google, not Verily, paid Fossil $40 million for a still under development smartwatch technology to fit into Google’s Ware OS area. It’s not known whether it is health related, but their CEO admitted that it was based on tech from the Misfit acquisition–and Misfit was focused on health tech. After the sale closing, it is predicted that some Fossil R&D staff will move over to Google. Back in 2015, Fossil paid $260 million for Misfit and their fitness tech but generally has stayed in the conventional smartwatch area. The story broke in Wareable. Also Mobihealthnews.

UK’s DeepMind loses Streams, health projects to Google Health

DeepMind loses its Health to Google. DeepMind, the London-based AI developer acquired by Alphabet (Google) in 2014, no longer has a Health division. This group will be absorbed by Google Health, now headed by ex-Geisinger CEO David Feinberg. The former DeepMind health team will continue to be headed by former NHS surgeon Dr Dominic King, who will remain in London along with about 100 reported staffers, at least for now.

DeepMind’s major health initiative is Streams, an AI-powered mobile app that analyzes potential deterioration in patients and alerts nurses and doctors, saving time. It also monitors vital signs and integrates different types of data and test results from existing hospital IT systems. Streams is currently deployed at Royal Free NHS Foundation Trust Hospital in north London for acute kidney injury. The rollout is expected to be made at Imperial College Healthcare NHS Trust, Taunton and Somerset NHS Foundation Trust and Yeovil District Hospital NHS Foundation Trust. It is expected that test partners will be found outside of the UK.

DeepMind’s other health initatives and research include fast eye disease detection, planning cancer radiotherapy treatment in seconds rather than hours; and detecting patient deterioration from electronic records.

Google Health is now expanding into products and research into digital technologies which was to be expected with Dr Feinberg on board. Currently, its revenue stream consists of advertising and search.

The remainder of DeepMind not engaged with health will remain independent. CNBC, DeepMind blog

News roundup: Walmart and Microsoft AI, are derm apps endangering public with 88% skin cancer diagnosis?

[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2017/12/Lasso.jpg” thumb_width=”150″ /]Walmart and Microsoft partner to change the retail experience via AI. The five-year agreement will switch over applications to the cloud and will affect shipping and supply chain. It’s projected in Healthcare Dive that the impact will be in healthcare as well. Microsoft announced last month that it is forming a unit to advance AI and cloud-based healthcare tools. The landscape is under extreme pressure in retail and healthcare delivery, and Walmart needs to ready for future moves which will certainly happen. Walmart is rumored to be interested in acquiring Humana and is currently working with Emory Healthcare in Atlanta. Then there is CVS-Aetna, Cigna-Express Scripts, Google, and (looming above all) Amazon. (Though you can tuck all the years of Amazon’s profits into one year of Walmart’s.)

The ITV News headline grabs attention — but are dermatology apps really endangering the public when teledermatology can help diagnose 88 percent of people with skin cancer and 97 percent of those with benign lesions? A University of Birmingham-led research team did a metastudy of the literature and found three failings: “a lack of rigorous published trials to show they work and are safe, a lack of input during the app development from specialists to identify which lesions are suspicious and flaws in how the technology analyses photos” particularly for scaly or non-pigmented melanomas. But did access to these apps encourage early diagnosis which can lead to up to 100 percent five-year survival? Of course review is required as recommended by the study, but this last factor was not really examined at the British Association of Dermatologists’ annual meeting in Edinburgh. University of Birmingham release with study abstract

Google’s ‘Medical Brain’ tests clinical speech recognition, patient outcome prediction, death risk

Google’s AI division is eager to break into healthcare, and with ‘Medical Brain’ they might be successful. First is harnessing the voice recognition used in their Home, Assistant, and Translate products. Last year they started to test a digital scribe with Stanford Medicine to help doctors automatically fill out EHRs from patient visits, which will conclude in August. Next up, and staffing up, is a “next gen clinical visit experience” which uses audio and touch technologies to improve the accuracy and availability of care.

The third is research Google published last month on using neural networks to predict how long people may stay in hospitals, their odds of re-admission and chances they will soon die. The neural net gathers up the previously ungatherable–old charts, PDF–and transforms it into useful information. They are currently working with the University of California, San Francisco, and the University of Chicago with 46 billion pieces of anonymous patient data. 

A successful test of the approach involved a woman with late-stage breast cancer. Based on her vital signs–for instance, her lungs were filling with fluid–the hospital’s own analytics indicated that there was a 9.3 percent chance she would die during her stay. Google used over 175,000 data points they saw about her and came up with a far higher risk: 19.9 percent. She died shortly after.

Using AI to crunch massive amounts of data is an approach that has been tried by IBM Watson in healthcare with limited success. Augmedix, Microsoft, and Amazon are also attempting AI-assisted systems for scribing and voice recognition in offices. CNBC, Bloomberg

Contact lenses as a drug delivery system take home MIT Sloan Healthcare prize

This Editor has been covering contact lenses in health tech since at least 2013–contact lenses that detect glucose for diabetics (Google/Novartis/Alcon), eye pressure (Sensimed), and even detect multiple diseases (Oregon State University). None to date have made it into commercial release.

Here’s another try, this time from this year’s winner of the MIT Sloan Healthcare Innovation Prize competition. Theraoptix won the $25,000 grand prize, sponsored by Optum. The lenses are designed to deliver eye medication on a time release basis using a thin polymer film formed into a tiny circular strip sandwiched into the lens material. They can be worn for up to two weeks to slowly but constantly deliver drugs in the treatment of diseases like glaucoma or after surgery. It can also deliver drugs effectively for back of the eye treatment of macular degeneration, diabetic retinopathy, retinal vein occlusion, and similar diseases that today require in-office injections.

Theraoptix was developed by Lokendra Bengani Ph.D. of the Schepens Eye Research Institute of the Massachusetts Eye and Ear Infirmary. It was based on core technology by ophthalmologist Joseph B. Ciolino MD, who is Dr. Bengani’s mentor. We wrote about Dr. Ciolino’s research previously [TTA 7 Sept 16] including a look back at contact lens research. There were seven other finalists, of which the most interesting to this Editor was Kinematics shoe insole sensors for gait detection analysis (and fall prevention).  MIT News

Is Uber fit to deliver healthcare transport? Healthcare organizations may want to check.

Healthcare-related organizations have codes of conduct pertaining to suppliers. Does Uber meet compliance standards? As we reported a few days ago in our article on the burgeoning area of non-emergency medical transport (NEMT) [TTA 9 Mar], Uber Health’s debut with a reputed 100 healthcare organizations has led this Editor to a further examination of Uber, the organization. Uber has had a hard time staying out of the headlines–and the courts–in the past two years, in matters which might give healthcare partners pause.

  • On 21 Nov, Uber reported that the personal data of 57 million users, including 600,000 US drivers, were breached and stolen in October 2016–a full year prior. Not only was the breach announcement delayed by over a year, but also in that year it was made to go away by Uber’s paying off the hacker. Reuters on 6 December: “A 20-year-old Florida man was responsible for the large data breach at Uber Technologies Inc [UBER.UL] last year and was paid by Uber to destroy the data through a so-called “bug bounty” program normally used to identify small code vulnerabilities, three people familiar with the events have told Reuters.” The payment was an extraordinary $100,000. “The sources said then-CEO Travis Kalanick was aware of the breach and bug bounty payment in November of last year.” The Reuters article goes further into the mechanism of the hack. It eventually led to the resignation of their chief security officer, former Facebook/eBay/PayPal security head Joe Sullivan, who ‘investigated’ it using encrypted, disappearing messaging apps. Atlantic.
  • CEO and co-founder Travis Kalanick was forced to resign last June after losing the confidence of the company’s investors, in contrails of financial mismanagement, sexual harassment, driver harassment, and ‘bro culture’. This included legal action over Uber’s 2016 acquisition of self-driving truck startup Otto, started by former Googlers who may or may not have lifted proprietary tech from Google before ankling. These are lavishly outlined in Bloomberg and in an over-the-top article in Engadget (with the usual slams at libertarianism). Mr. Kalanick remains on the board and is now a private investor.
  • The plain fact is that Uber is still burning through funds (2017: $1bn) after raising $21.1bn and its valuation has suffered. The new CEO Dara Khosrowshahi, who earlier righted travel site Expedia, has a tough pull with investors such as SoftBank and Saudi Arabia’s Public Investment Fund. Also Mashable.

Healthcare and NEMT, as noted in our earlier article, are a strong source of potential steady revenue through reimbursement in Medicare Advantage and state Medicaid programs, which is why both Uber and Lyft are targeting it. The benefits for all sides–patients, practices, these companies, sub-contractors, and drivers–can be substantial and positive in this social determinant of health (SDOH).  

Healthcare organizations, especially payers, have strict codes of compliance not only for employees and business practices but also for their suppliers’ practices. Payers in Medicare Advantage and Medicaid are Federal and state contractors. While Uber under its new CEO has shown contriteness in acknowledging an organization in need of righting its moral compass (CNBC), there remains the track record and the aftermath. Both deserve a closer look and review.

Google ‘deep learning’ model more accurately predicts in-hospital mortality, readmissions, length of stay in seven-year study

A Google/Stanford/University of California San Francisco/University of Chicago Medicine study has developed a better predictive model for in-hospital admissions using ‘deep learning’ a/k/a machine learning or AI. Using a single data structure and the FHIR standard (Fast Healthcare Interoperability Resources) for each patient’s EHR record, they used de-identified EHR derived data from over 216,000 patients hospitalized for over 24 hours from 2009 to 2016 at UCSF and UCM. Over 47bn data points were utilized.

The researchers then looked at four areas to develop predictive models for mortality, unplanned readmissions (quality of care), length of stay (resource utilization), and diagnoses (understanding of a patient’s problems). The models outperformed traditional predictive models in all cases and because they used a single data structure, are projected to be highly scalable. For instance, the accuracy of the model for mortality was achieved 24-48 hours earlier (page 11). The second part of the study concerned a neural-network attribution system where clinicians can gain transparency into the predictions. Available through Cornell University Library. AbstractPDF.

The MarketWatch article rhapsodizes about these models and neural networks’ potential for cutting healthcare costs but also illustrates the drawbacks of large-scale machine learning and AI: what’s in the EHR including those troublesome clinical notes (the study used three additional deep neural networks to discern which bits of the clinical data within the notes were relevant), lack of uniformity in the data sets, and most patient data not being static (e.g. temperature). 

And Google will make the chips which will get you there. Google’s Tensor Processing Units (TPUs), developed for its own services like Google Assistant and Translate, as well as powering identification systems for driverless cars, can now be accessed through their own cloud computing services. Kind of like Amazon Web Services, but even more powerful. New York Times

Verily’s million points of BYO health data to take to your next doctor visit

Verily‘s visit to last week’s Health 2.0 conference had an odd-but-fun tack, comparing the data received from human bodies to the billions of data points generated by an average late-model automobile in normal operations. We generate a lot less (ten orders of magnitude difference, according to Verily Chief Technology Officer Brian Otis), but Verily wants to maximize the output by wiring us to multiple sensors and to use the data in a predictive health model. Some of the Verily devices this Editor predicts will be non-starters (the sensor contact lens developed with Alcon) but others like the Dexcom partnership to develop a smaller, cheaper continuous blood glucose monitor and Liftware, the tremor-canceling silverware company Google acquired in 2014, appear promising. Key to predictive health is the Study Watch, which is a wearable that collects a lot of data but is easy to wear for a long time. Mobihealthnews

But what to do with this All That Data? Where this differs from a car is that the operational data goes into feedback loops that tune the engine’s performance, perform long-term monitoring, electrical system, braking, and more. (When the sensors go south or the battery’s low, watch out!) It’s not clear from the talk where this overwhelming amount of healthcare data generated goes to and how it becomes useful to a person or a doctor. This has its own feedback loop this Editor dubbed a few years ago as the Five Big Questions (FBQs): who pays, how much, who’s looking at the data, who’s actioning it, how data is integrated into patient records. That’s not answered, but presumably these technologies will incorporate machine learning and AI to Crunch That Data into bite-sized parts.

Which leads us back to Verily’s parent, Alphabet a/k/a Google. All that data into Verily devices could be monitored by Google and fed into other Google programs like their search engines and Adwords. Another privacy problem? 

Perhaps health systems are arriving at the realization that they have to crunch the data, not avoid it. For the first time, this Editor has observed that a CMIO of a small health system in Illinois and Sanford Health‘s executive director of analytics are actually welcoming patient data and research. Startups in this area such as PreventScripts labor on that “last mile” of clinical decision support, preventative medicine. EHRs are also into the act. Epic launched Share Everywhere, where patients can grant access to their data and clinicians can send updates into the patient portal (MyChart). What’s needed, CMIO Goel admits, is software that combines natural language processing and algorithms to track by disease and specialty–once again, machine learning. Healthcare IT News 

Can Google Glass’ enterprise iteration solve the patient documentation crisis?

[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2017/07/Glass-EE.jpg” thumb_width=”200″ /]”Glass is a hands-free device, for hands-on workers.” What a marketing position! Google Glass finally arrives at where it should have started–not a techie toy or a social snooper banned from bars, but a tool for specific work needs that solve specific but important problems. This is not only ‘on trend’, but also the ‘professional case’ is steak on the grill as a powerful way to lend legitimacy to a new product (the classic is Tang ‘orange drink’ going into space in the early ’60s). The recent announcement of Glass Enterprise Edition (EE) marking its emergence from stealth mode was a refreshingly low-key (for Google and parent Alphabet) surprise. Even the revamped look is sturdy and utilitarian in full glass mode (left) or in clip-on (and also serves as eye protection). 

Their on-trend position for healthcare is to reduce the amount of time that doctors spend charting and documenting patients. Augmedix, a Glass partner, built the documentation automation platform for Sutter Health and for Dignity Health that captures the information from the interaction between patient and doctor via a ‘remote scribe’. Jay Kothari, the Glass project lead, quotes data from Dignity that it reduces clinician daily documentation time from 33 percent to less than 10 percent,  The Sutter Health estimate is two hours per day. Out of the gate this is extremely valuable because it improves the clinician-patient face-to-face (and presumably virtual) visit in eye contact, reduces the break in taking notes, and reduces time pressure generated by post-visit review. Netherlands-based swyMed concentrates on facilitating virtual visits, and is testing a home visit pilot with Loyola University Health System practitioners in Maywood, Illinois. Others, like John Nosta, have been continuing to use Glass in business. Our Readers may want to check out these partners as that is how Google is making the Glass available, not directly. SF/Boston-based partner Brain Power wasn’t mentioned in Mr. Kothari’s blog, but their AI/VR applications for brain conditions such as autism and TBI, as well as other uses such as clinical trials and care for older adults. mHealthIntelligence interviewed Augmedix’s CEO Ian Shakil, who notes that Glass still needs improvements in battery life for the hard work of documenting patient visits.

Update: An interesting comment on this via Twitter. The paper is from 2015 but the regulatory and privacy questions around recording patients and information remain. Augmedix does state on its website that it is HIPAA compliant.

 
[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2017/07/Glass-Twitter.jpg” thumb_width=”250″ /]

Behave, Robot! DARPA researchers teaching them some manners.

[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2014/01/Overrun-by-Robots1-183×108.jpg” thumb_width=”150″ /]Weekend Reading While AI is hotly debated and the Drudge Report features daily the eeriest pictures of humanoid robots, the hard work on determining social norms and programming them into robots continues. DARPA-funded researchers at Brown and Tufts Universities are, in their words, working “to understand and formalize human normative systems and how they guide human behavior, so that we can set guidelines for how to design next-generation AI machines that are able to help and interact effectively with humans,” said Reza Ghanadan, DARPA program manager. ‘Normal’ people determine ‘norm violations’ quickly (they must not live in NYC), so to prevent robots from crashing into walls or behaving towards humans in an unethical manner (see Isaac Asimov’s Three Laws of Robotics), the higher levels of robots will eventually have the capacity to learn, represent, activate, and apply a large number of norms to situational behavior. Armed with Science

This directly relates to self-driving cars, which are supposed to solve all sorts of problems from road rage to traffic jams. It turns out that they cannot live up to the breathless hype of Elon Musk, Google, and their ilk, even taking the longer term. Sequencing on roadways? We don’t have the high-accuracy GPS like the Galileo system yet. Rerouting? Eminently hackable and spoofable as WAZE has been. Does it see obstacles, traffic signals, and people clearly? Can it make split-second decisions? Can it anticipate the behavior of other drivers? Can it cope with mechanical failure? No more so, and often less, at present than humans. And self-drivers will be a bonanza for trial lawyers, as added to the list will be car companies and dealers to insurers and owners. While it will give mobility to the older, vision impaired, and disabled, it could also be used to restrict freedom of movement. Why not simply incorporate many of these assistive features into cars, as some have been already? An intelligent analysis–and read the comments (click by comments at bottom to open). Problems and Pitfalls in Self-Driving Cars (American Thinker)

Wearables: it’s a journey, but is it really necessary?

[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2015/08/is-your-journey-neccessary_.jpg” thumb_width=”150″ /]Increasingly, not in the opinion of many. We’ve covered earlier [TTA 21 Dec, 6 Feb] the wearables ‘bust’ and consumer disenchantment affecting fitness-oriented wearables. While projections are still $19 bn by 2018 (Juniper Research), Jawbone is nearly out of business with one last stab at the clinical segment, with Fitbit missing its 2016 earnings targets–and planning to target the same segment. So this Washington Post article on a glam presentation at SXSW of a Google/Levi’s smart jeans jacket for those who bicycle to work (‘bike’ and ‘bikers’ connote Leather ‘n’ Harleys). It will enable wearers to take phone calls, get directions and check the time by tapping and swiping their sleeves, with audio information delivered via headphone. As with every wearable blouse, muumuu, and toque she’s seen, this Editor’s skepticism is fueled by the fact that the cyclist depicted has to raise at least one hand to tap/swipe said sleeves and to wear headphones. He is also sans helmet on a street, not even a bike path or country lane. All are safety Bad Doo-Bees. Yes, the jacket is washable as the two-day power source is removable. But while it’s supposed to hit the market by Fall, the cost estimate is missing. A significant ‘who needs it?’ factor.

Remember the Quantified Selfer’s fascination with sleep tracking and all those sleep-specific devices that went away, taking their investors’ millions with them? Fitbit and many smartwatches work with apps to give the wearer feedback on their sleep hygiene, but the devices and apps themselves can deliver faulty information. This is according to a study published in the Journal of Clinical Sleep Medicine called “Orthosomnia: Are Some Patients Taking the Quantified Self Too Far?” (abstract) by Kelly Glazer Baron, MD with researchers from the Feinberg School of Medicine at Northwestern University. “The patients’ inferred correlation between sleep tracker data and daytime fatigue may become a perfectionistic quest for the ideal sleep in order to optimize daytime function. To the patients, sleep tracker data often feels more consistent with their experience of sleep than validated techniques, such as polysomnography or actigraphy.” (more…)

Robot-assisted ‘smart homes’ and AI: the boundary between supportive and intrusive?

[grow_thumb image=”http://telecareaware.com/wp-content/uploads/2016/06/Robot-Belgique-1.png” thumb_width=”200″ /]Something that has been bothersome to Deep Thinkers (and Not Such Deep Thinkers like this Editor) is the almost-forced loss of control inherent in discussion of AI-powered technology. There is a elitist Wagging of Fingers that generally accompanies the Inevitable Questions and Qualms.

  • If you don’t think 100 percent self-driving cars are an Unalloyed Wonder, like Elon Musk and Google tells you, you’re a Luddite
  • If you have concerns about nanny tech or smart homes which can spy on you, you’re paranoid
  • If you are concerned that robots will take the ‘social’ out of ‘social care’, likely replace human carers for people, or lose your neighbor their job, you are not with the program

I have likely led with the reason why: loss of control. Control does not motivate just Control Freaks. Think about the decisions you like versus the ones you don’t. Think about how helpless you felt as a child or teenager when big decisions were made without any of your input. It goes that deep.

In the smart home, robotic/AI world then, who has the control? Someone unknown, faceless, well meaning but with their own rationale? (Yes, those metrics–quality, cost, savings) Recall ‘Uninvited Guests’, the video which demonstrated that Dad Ain’t Gonna Take Nannying and is good at sabotage.

Let’s stop and consider: what are we doing? Where are we going? What fills the need for assistance and care, yet retains that person’s human autonomy and that old term…dignity? Maybe they might even like it? For your consideration:

How a robot could be grandma’s new carer (plastic dogs to the contrary in The Guardian)

AI Is Not out to Get Us (Scientific American)

Hat tip on both to reader Malcolm Fisk, Senior Research Fellow (CCSR) at De Montfort University via LinkedIn