Things & Thinks — Issue XXXVII
For this thirty seventh issue of Things & Thinks, the research brief summarizes some interesting findings from a recent study about adoption of AI driven clinical decision support systems. The Digital Healthcare news section covers quite a bit. The monthly sections with a Longread, Tweet and Chart o the Month, with the newly introduced Mental Model of the Month. Happy to hear your thoughts and feedback…
Research Digest
Designing and deploying AI
A lot of research is being done on AI/ML driven clinical decision support systems and there is an increasing focus on thinking about implementing these systems in the real world.
Integrating ML could present a challenge in time-constrained clinical contexts, where clinicians must rapidly evaluate whether and how to act on recommendations while managing competing demands on their time and attention. Indeed, most systems typically report users responding to only 6–45% of alerts or requiring dedicated staff. It is not yet well-understood how clinicians perceive ML recommendations in these contexts, and thus it remains unclear how to most effectively deploy ML systems to maximize clinical benefit. Two important hurdles identified include-
- Experts may struggle to develop trust with ML-based systems due to a large number of inputs and the complex integration of data involved, which can make it challenging or impossible to convey the specific logic behind an alert or recommendation.
- Many view ML as being impoverished relative to human expertise and question whether it can add clinical value for highly trained expert users. By making a competing diagnosis, ML systems could also be perceived as encroaching on clinicians’ professional role, presenting a “threat to autonomy” that may make clinicians reluctant to use, rely on, and trust them.
Teams from Bayesian Health and the Johns Hopkins University worked to develop and deploy a sepsis early detection tool TREWS (Targeted Real-time Early Warning System); it was deployed at five hospitals over a 2-year period. While the team has provided details about the high accuracy of the system, what’s more interesting is their work on identifying factors influencing the adoption of the system. Some of these factors included-
- High predictive performance and low alert burden leading to better trust
- Ease of use and integration into the provider’s EHR workflow
- Inclusion of alert context to avoid ‘black box’ presentation
- Different kinds of trust factors including dispositional trust, learned trust and situational trust
Digital Healthcare News
Tech in Digital Health
Google will delete location data after people visit abortion clinics, domestic violence shelters and other sensitive locations; Additionally, the company will soon add a feature for users to delete multiple menstruation logs at once on Google Fit and Fitbit apps, seen as a direct result of overturning Roe v Wade.
New York-based health system Northwell Health is partnering with Google Cloud to use cloud technology and artificial intelligence to automate administrative workflows and identify patient risk factors for early intervention.
Amazon will buy One Medical-primary health care provider- for roughly $3.9 billion, approximately at an 80% premium of the market price.
Regulatory/Policy Brief
FDA cleared LiveMetric‘s smartwatch-like, cuffless blood pressure sensor, which uses an array of tiny sensors that can physically measure the minuscule movements of blood as it passes under the skin beat by beat.
Google’s Verily and iRhythm Technologies have received 510(k) clearance from the US FDA for a new AI-powered smartwatch system to monitor cardiac arrhythmia.
ResApp Health received the US FDA’s 510(k) clearance for SleepCheckRx, its mobile sleep apnoea screening app.
Healthy.io received 510(k) clearance from the U.S. FDA for Minuteful Kidney test, allowing people to test their kidney function from home and get clinical-grade results instantly.
Biopharma/Devices Brief
Roche and EarlySign expanded their partnership to include AI-powered lung cancer diagnosis by developing and commercializing EarlySign’s LungFlag technology.
Health2Sync and AstraZeneca to roll out digital support tool for early CKD detection in Taiwan, seeking to promote the early diagnosis of CKD and optimise everyday CKD care.
Boehringer Ingelheim will collaborate with VisionHealth on its Kata app to provide therapy support for those living with asthma and chronic obstructive pulmonary disease (COPD). The Kata app is a digital platform, certified as a Class 2a medical device, that is able to provide targeted inhalation support for those living with chronic respiratory diseases.
AstraZeneca has signed a cooperation agreement with docdok.health, a Swiss-Israeli startup company to develop a new digital solution that can monitor patients’ asthma status in real-time, identifying situations in which asthma may worsen and sending immediate alerts to the treating physician.
Anumana will partner with Novartis to develop artificial intelligence tools to detect cardiovascular diseases. Anumana is a joint venture between EHR data company nference and the Mayo Clinic
Funding, Deals, Mergers & acquisitions
SD Biosensor (a diagnostics company headquartered in South Korea) with PE firm SJL Partners agreed to acquire Meridian Bioscience in an all-cash deal valued at about $1.53 billion
Gingko Bioworks bought Zymergen for $300m.
Cardiovascular disease-focused Cleerly raised $192 million; it offers an AI-backed tool that evaluates CT angiograms for plaque build-up on the heart arteries’ walls,
Direct primary care provider Everside Health raised $164 million in growth equity funding.
Telehealth giant KRY, known as Livi in the UK and France, raised $160m.
PicnicHealth raised a $60m Series C to expand patient-centered real-world data.
Data exchange platform Particle Health raised $25m; it offers an API platform that aggregates and standardizes patient medical record data for provider organizations, pharmacies and virtual care companies.
EHR maker Canvas Medical raised $24m in a Series B funding round.
Mental health chatbot Wysa raised $20m in a Series B financing round led by HealthQuad.
Moving Analytics raised $20m to scale Cardiovascular Telehealth programs in underserved communities.
Eka Care, a healthcare technology startup in India, raised $15 m in a Series A funding round
Digital health platform MFine will merge with LifeCell‘s diagnostics arm to form a new entity called LifeWell.
3M announced that it would spin off its medtech business by the end of next year.
UnitedHealth Group has combined certain Optum Health assets with Red Ventures’ healthcare portfolio to create a new consumer health joint venture called RVO Health. These assets include Optum’s Store and Perks and Red Ventures’ Healthline Media and Healthgrades, Virtual coaching platforms Real Appeal, Wellness Coaching and QuitForLife.
PatientBond, Vizient team up for digital behavior change tools to provide Vizient member healthcare organizations with digital patient engagement and behavior change programs.
Ultrasound devicemaker Exo acquired imaging AI developer Medo for an undisclosed amount.
Other News
Digital Medicine Society (DiMe) released four toolkits to help scale wearables, remote patient monitoring, based on its Sensor Data Integrations project, which includes organizations like Amazon Web Services, Oracle, the Moffitt Cancer Center, Takeda and the U.S. Department of Veterans Affairs.
Healthcare data breach costs reached record high at $10M per attack, up 9.4% from past year, according to IBM Security’s annual Cost of a Data Breach Report.
Health analytics assets of IBM Watson Health has led to a new healthcare data company called Merative.
Nearly 75% of the 1,000 patients surveyed by Savvy Cooperative expressed concern about protecting the privacy of personal health data, and only 20% of patients said they knew the scope of companies and individuals with access to their data.
Medidata will collaborate with Aural Analytics, a speech collection and analysis software developer, to add speech biomarkers to its drug research offering.
Teladoc’s Primary360 added care coordination support and health plan in-network referrals alongside free same-day medication delivery from Capsule and in-home, on-demand phlebotomy services backed by Scarlet Health
NewYork-Presbyterian will support Cornell with $15 million over three years to fund the initiative aimed to improve heart failure treatment, as well as predict and prevent heart failure.
Philippine’s biggest conglomerate Ayala launched health superapp-KonsultaMD-by consolidating three Filipino health-tech platforms KonsultaMD, HealthNow, and AIDE.
Medable is partnering with Withings Health Solutions to use the company’s connected devices for decentralized clinical trials.
A new partnership between Macmillan Cancer Support (Macmillan) and digital therapeutics company Big Health will see newly diagnosed cancer patients across the UK get free access to Big Health’s Sleepio for insomnia and Daylight for anxiety to manage their mental health needs.
Mental Model of the Month
With a lot of upheaval in the digital health world, it feels right to talk about Survivorship bias. This blog by Farnam Street excellently summarizes the bias.
When we only pay attention to those who survive, we fail to account for base rates and end up misunderstanding how selection processes actually work.
This bias is quite common in innovation/start-up world and can lead to blind spots and ignorance.
Almost every single generic presentation for startups starts with “Ninety Five percent of all startups fail”, but very rarely do we pause for a moment and think “what does this really mean?” We nod our heads in somber acknowledgement and with great enthusiasm turn to the heroes who “made it” — Zuckerberg, Gates, etc. to absorb pearls of wisdom and find the Holy Grail of building successful companies. Learning from the successful is a much deeper problem and can reduce the probability of success more than we might imagine.
Longread of the Month
Apple published a 60 page report about their work in healthcare. Lots of interesting details, giving a glimpse of current capabilities and future ambitions…
Tweet of the Month
This makes so much sense especially for healthcare consumers.
Chart of the Month
New data from investment firm Rock Health shows venture funding fell steeply to $4.1 billion in the second quarter amid fears about a digital health bubble and a broader economic recession
