|
AI Disease Predictions, Smarter Acquires Pieces, and Everyone’s Turning 100
October 2, 2025
|
|
|
|
|
Together with
|
|
|
|
“The lesson for everyone else in healthcare is clear: don’t build a product, build infrastructure. Because when everything becomes the same, integration is the only thing that matters.”
|
|
OpenLoop CEO Dr. Jon Lensing
|
|
|
|
Clinical decision-making relies on understanding patients’ past health to improve their future health, an impossible task without first understanding how diseases progress over time.
That’s where a new study in Nature suggests AI is ready to help.
It starts with generative pretrained transformers. Researchers built a GPT, dubbed Delphi-2M, to predict the “progression and competing nature of human diseases.”
- Delphi-2M was trained on 400k UK Biobank participants (which lean healthier than the average person), and then externally validated on 1.9M Danish patients.
- The training was designed to predict a patient’s next diagnosis and the time to it, using only data readily available within the EHR: past medical history, age, sex, BMI, and alcohol/smoking status.
How’d it do? The results speak for themselves:
- Delphi-2M was able to forecast the incidence of over 1,000 diseases with comparable accuracy to existing models that are fine-tuned to predict single diseases.
- Death could be predicted with eerily impressive accuracy (AUC: 0.97), and the survival curves that it simulated lined up almost perfectly with national mortality statistics.
- Comorbidities emerged naturally from the training, and Delphi-2M was able to understand the progression from type 2 diabetes to eye disease to nerve damage.
- Delphi-2M’s ability to predict heart attack and stroke matched established scores like QRisk, and it even outperformed leading biomarker-based AI models.
Better forecasts inform better policies. If policymakers can consult the Oracle of Delphi to see how many people will develop a disease over the next decade, the authors conclude that they’ll also be able to implement better regulations to prepare.
- Not a bad theory, assuming models trained on historical data can make forecasts that hold up to evolving treatments and populations (and that politicians act in the best interest of the people:).
The Takeaway
AI is reaching the point where it can predict thousands of diseases as well as the best narrowly focused models, and that could have big implications for everything from early screening to policymaking.
|
|
Start Getting Leads From ChatGPT and Google AI
If your healthcare company isn’t getting leads from ChatGPT and Google AI Overviews, it’s probably because your content isn’t optimized for AI search. Tely AI runs keywords and questions research, generates and publishes GEO content, and captures new leads – all on autopilot. Create your first article on us to start filling your pipeline with patients and partners from today’s lead sources.
|
|
Scale RPM With BPM Pro 2
BPM Pro 2 is the next generation cellular blood pressure monitors, empowering care teams to scale remote patient monitoring and streamline operations. Discover why leading providers are choosing BPM Pro 2 to collect highly precise measurements and enrich data with Patient Insights from their daily lives.
|
|
- Smarter Technologies Acquires Pieces: Smarter Technologies is rounding up the missing pieces to its AI revenue management platform, and it hit that goal right on the nose with the acquisition of Pieces Technologies. Pieces’ ambient scribing tech will power the just-launched SmarterNotes, which embeds Smarter’s RCM intelligence directly into every note. Smarter Technologies was formed through the recent merger of SmarterDx, Thoughtful.ai, and Access Healthcare, so it has plenty of dots to connect to the actual clinical conversation.
- Apple Watch Hypertension Data: A study looking at the Apple Watch’s new hypertension alert found that it missed over half of cases among the 2,200 participants. Although that seems to suggest that Apple decided to play it safe with the feature’s sensitivity tuning, 7% of participants without hypertension also received the notification. Balancing that tradeoff is going to be the name of the game until cuffless blood pressure monitoring becomes a reality, and it looks like there’s still a way to go before we get there.
- Thyme Series D: Thyme Care just became the latest digital health startup to join the unicorn club after notching $97M of Series D funding. The value-based cancer care company works with payors, employers, and risk-bearing providers to “break through the biggest bottlenecks” across the oncology ecosystem. That’s achieved by helping patients navigate their diagnosis, receive care between visits, and supporting them with a dedicated team of providers and nurses. Thyme is already available to 8M people across all 50 states, and the fresh funding was earmarked to help scale that reach even further.
- Twice as Many People Living to 100: The number of U.S. citizens over 100 years old is up nearly 50% over the last decade. The Census Bureau counted 80,100 centenarians in 2020 (up from 53k in 2010), which added up to about 2.4 centenarians for every 10k residents. It’s great to see that treatment breakthroughs and new technologies are keeping people alive longer, but it’s also a potent reminder of how underprepared the U.S. healthcare system is for the burden that aging seniors are about to place on it.
- Mayo Clinic Platform_Orchestrate: Mayo Clinic just took the lid off Mayo Clinic Platform_Orchestrate, a new program designed to accelerate clinical development while helping medical device and biopharma companies bring therapies to patients faster. Mayo Clinic Platform_Orchestrate gives partners a single point of access to the system’s clinical expertise, de-identified clinical data, and international partner network. Sounds like a seriously valuable resource for companies looking to accelerate anything from discovery to deployment.
- Confido Closes $10M: Confido Health landed $10M of Series A funding to fuel the development of AI voice agents fine-tuned for healthcare. The agents are EHR-agnostic, go beyond simple scheduling (RCM, referrals, intake), and are “ROI-obsessed” (the scorecard is first-call resolution and staff hours saved). In the last 10 months Confido has grown 10x, and now covers 1M+ patients across various specialties.
- Claims Errors Driving Denials: Experian’s State of Claims 2025 report showed that 41% of providers have over 10% of their insurance claims denied. That’s up from “just” 30% of providers in 2022, in large part because payors are leveraging new AI-powered tools to catch errors. More than half of providers (54%) say claims errors are increasing – many of which appear to originate at registration – and 68% report that submitting clean claims is more challenging than a year ago.
- Marathon and Medbridge MSK: Marathon Health and Medbridge debuted a new virtual musculoskeletal program built directly into Marathon’s advanced primary care model. Unlike standalone solutions, Medbridge integrates its hybrid MSK pathways straight into Marathon’s in-person and virtual primary care, along with triage, care planning, and motion capture tools for improving assessment and outcomes. The program is rolling out in select markets before expanding across Marathon’s 600+ health centers serving 3M members.
- Koda Series A: Koda Health raised $7M of Series A funding to help more health systems scale goals-of-care conversations and align treatment with patient preferences. The Koda platform enables education and informed decision-making around serious illnesses that patients are often forced to face without adequate support. Nearly one in four healthcare dollars is spent in the last year of life, and a study with Houston Methodist showed that Koda reduced terminal hospitalizations by 79%, increased hospice use by 51%, and lowered costs by $9k per patient.
- FDA Mulls Approach to Real-World AI Data: The FDA issued a request for public comment on how to evaluate AI’s effectiveness for real-world clinical tasks. The move addresses concerns that positive AI results in clinical studies aren’t replicated in real-world use, due to changes in clinical practice, patient demographics, data inputs, and healthcare infrastructure. The topic came up during the first meeting of the FDA’s Digital Health Advisory Committee in November 2024, and most likely will be discussed at the next DHAC meeting on November 6.
|
|
Under the Hood of Navina’s AI
Navina’s AI engine harnesses over 600 proprietary algorithms to transform fragmented patient data into actionable clinical intelligence at the point of care. It’s shaped with the expertise of physicians to turn multiple data sources (EHR, HIE, claims, care gap files, etc.) into contextualized insights like suspected conditions or evidence for care gap closures – each linked back to the original source. Download the whitepaper to see examples of Navina’s AI in action.
|
|
Next Generation Ambient Tech and Agents
The ambient AI transformation is already sweeping across health systems, reducing administrative burdens and improving patient outcomes. So, what’s next? Tune into this on-demand session to learn how systems like Carle Health and Denver Health are leveraging Nabla to eliminate Pajama Time and build a future where agentic AI unlocks true workforce sustainability.
|
|
- 10 Strategies to Expedite Provider Credentialing: Lagging provider credentialing workflows can create a wave of unexpected care delays and financial setbacks. Sidestep these pitfalls by tuning into Medallion’s recording of 10 Key Strategies to Expedite Provider Credentialing and see how your organization can keep a well-oiled team and patient care intact through the ever-evolving changes seen in healthcare.
- Achieve Ambient AI Scale With Abridge: According to a new report from MIT, 95% of generative AI implementations fail. At Abridge, we have helped 150+ health systems scale ambient AI to tens of thousands of clinicians—many in a matter of weeks. For the first time ever, we are sharing those steep adoption curves at partners, along with impact metrics and testimonials from 19 health systems. If your ambient AI implementation isn’t scaling at speed, Abridge can help.
|
|
|
|
|
