Clinical & Diagnostics
IKS Health Debuts MyCareHub — Agentic AI Platform That Automates Patient Engagement
MyCareHub runs on an active, multi-agent behavioral algorithm that integrates natively with Epic, enabling autonomous execution of patient engagement workflows across appointment reminders, care gap closure, chronic disease monitoring, and follow-up scheduling. The platform's orchestration layer selects and sequences micro-agents in real time based on patient-specific context, eliminating the fixed rule trees of legacy automation tools.
6 Health Systems Enhancing Care Delivery with Ambient AI Scribes
A JAMA-published study across five academic medical centers found ambient scribes reduced total EHR time by 13.4 minutes and documentation time by 16.0 minutes per patient encounter. Cooper University Healthcare's Dragon Copilot deployment achieved 4.15 minutes saved per patient, compounding to 60+ minutes daily per clinician. Systems also reported improved note quality scores and reduced time-to-close for visit documentation.
Amazon Connect Health Brings Agentic AI to the Point of Care
Amazon Connect Health combines contact center AI (automated patient routing, sentiment analysis, call transcription) with clinical context layers that pull from EHR APIs via FHIR R4 connections. Agentic components handle multi-turn patient conversations, appointment scheduling, and escalation to human agents with a full interaction summary pre-loaded. Integration is available for Epic, Cerner, and HL7-compliant EHRs.
AI Is About to Overhaul the Entire U.S. Health Care Industry — But Is That a Good Thing?
The piece surveys AI integration across payer operations (UnitedHealth's 22,000 software engineers, 80%+ using AI to write code or build agents), clinical workflows (ambient scribes, diagnostic support), and patient-facing applications (chatbots, insurance navigation). Experts note that current AI systems lack robust hallucination controls in clinical contexts and that liability frameworks have not caught up with autonomous AI recommendations in care delivery.
Research & Science
Clinical AI Has Boomed. New Stanford-Harvard Report Shows What Actually Holds Up in Practice.
The report reviewed the most influential clinical AI studies published in 2025 through three lenses: real-world performance degradation versus trial results, bias and generalizability failures across patient subpopulations, and workflow integration barriers that caused clinical abandonment of technically valid tools. Key finding: the field is moving from Gartner's "Peak of Inflated Expectations" into the "Slope of Enlightenment," with evidence accumulation accelerating for validated tools while a significant percentage of peer-reviewed AI studies remain unreproducible in live deployments.
First Fully AI-Designed Drug Completes Phase IIa with Statistically Significant Efficacy
INS018_055 was conceived, designed, and optimized using Insilico's PandaOmics (target identification) and Chemistry42 (generative molecular design) platforms in 18 months at approximately $6 million in computational and discovery cost — compared to an industry average of $100M+ for a traditionally discovered compound at the same stage. The Phase IIa IPF trial measured FVC decline as the primary endpoint, achieving statistical significance against placebo. Full Phase IIb design is underway.
Merck and Mayo Clinic Announce AI-Enabled Drug Discovery and Precision Medicine Collaboration
The collaboration integrates Mayo Clinic's longitudinal clinical, imaging, and genomic patient datasets with Merck's AI-enabled virtual cell technologies and advanced analytics platforms. Research focus areas include target identification for complex diseases, patient stratification for clinical trial enrichment, and biomarker discovery using multimodal data fusion. The partnership leverages federated learning architectures to enable analysis across Mayo's multi-site data assets without centralized data exposure.
Deep Learning Framework Integrates Pathology and Radiology for AI-Assisted Diagnosis
The framework combines the Adaptive Multi-Resolution Imaging Network (AMRI-Net) with an Explainable Domain-Adaptive Learning (EDAL) strategy to handle heterogeneous imaging modalities and variable acquisition protocols. Evaluated on cancer diagnosis and progression monitoring datasets, the system achieved superior AUC versus single-modality baselines across multiple tumor types. EDAL provides saliency maps that highlight diagnostically relevant image regions, improving interpretability for clinical adoption.
Policy & Regulation
Aidoc Wins FDA Clearance for Comprehensive Foundation Model AI Covering 14 Indications
The CARE model achieved an average 97% sensitivity and 98% specificity across its 11 newly cleared indications in an FDA-reviewed pivotal study. The foundation model architecture enables a single model deployment to cover pulmonary embolism, aortic dissection, intracranial hemorrhage, vertebral fractures, and seven additional critical findings — eliminating the workflow complexity and IT overhead of managing separate AI modules per finding. This is one of the first FDA clearances explicitly granted to a foundation model architecture rather than a single-indication algorithm.
FDA Announces Sweeping Pullback on Oversight of AI-Enabled Devices and Wearables
The guidance redraws the regulatory boundary under the 21st Century Cures Act's clinical decision support (CDS) software exemption. Products that meet specific criteria — including displaying their clinical basis for a recommendation and not being intended to replace clinical judgment — can now enter the market without 510(k) clearance or De Novo review. Separately, the FDA's Quality Management System Regulation (QMSR) update aligns U.S. device quality standards with ISO 13485:2016. Most high-risk AI device obligations take effect August 2026, with full compliance required by August 2027.
TEFCA Reaches Nearly 500 Million Health Records Exchanged as HHS Leverages AI to Reduce Burden
TEFCA's growth is powered by FHIR-native architectures connecting Qualified Health Information Networks (QHINs) across payers, providers, and public health agencies. HHS is actively leveraging the network's AI capabilities to automate administrative burden reduction including prior authorization processing, claims reconciliation, and adverse event reporting. The ASTP/ONC simultaneously released draft USCDI v7 (January 29, 2026) proposing 29 new standardized data elements to expand interoperability to nutrition, quality improvement, and adverse event data streams.
Industry & Business
UnitedHealth Group Is Making a $3 Billion Bet on AI — What Does It Mean for Patients?
UHG employs 22,000 software engineers worldwide, with over 80% now using AI to write code or build agents across its Optum and UnitedHealthcare divisions. Current AI deployments span prior authorization automation, claims adjudication, clinical documentation (through Optum's provider tools), and network optimization. Hundreds of active job postings signal continued expansion in data science, AI agent development, and machine learning infrastructure. Critics argue the prior auth AI systems disproportionately deny claims for complex patients.
U.S. Digital Health Funding Surges to $4 Billion in Q1 2026, Average Deal Size Highest Since 2021
Q1 2026 saw average deal size reach $36.7M — the highest single-quarter average tracked since Q4 2021. The U.S. captured 76% of global digital health funding. AI-enabled startups received an 83% valuation premium over non-AI health tech companies, with top rounds including Abridge's $300M Series E ($5B valuation), Ambiance's $243M Series C ($1.04B), and OpenEvidence's $250M at $11.75B pre-money. Top funded categories: non-clinical workflow automation, clinical workflow tools, and data infrastructure.
Jimini Health Raises $17M to Launch AI Mental Health Chatbot Sage with Large Behavioral Health Systems
Sage is designed as a clinician-supervised AI companion rather than an autonomous therapist, operating within a hybrid care model where it conducts structured CBT and DBT skill-building conversations between patient appointments, collects PHQ-9 and GAD-7 data continuously, and generates session-preparation summaries for clinicians. Duke University's $15M NIMH grant is funding parallel research into LLM-based behavioral health monitoring using passive smartphone data — analyzing sleep, mobility, communication patterns, and home time to flag high-risk states.
Qualified Health Raises $125M Series B to Scale Generative AI Across Health Systems
Qualified Health's platform uses large language models fine-tuned on healthcare-specific datasets to automate revenue cycle tasks (prior authorization drafting, denial management, coding review), clinical documentation workflows, and patient communication. The Series B positions the company to expand into mid-market health systems following success with large academic medical centers. NEA's lead mirrors the firm's earlier bets on Abridge and other ambient AI companies that have since reached unicorn valuations.
Tucuvi Raises $20M Series A to Scale AI Care Management Platform Globally
Tucuvi's AI voice agent conducts structured clinical assessments via telephone — asking patients about symptoms, medication adherence, and functional status — and triages responses using validated clinical protocols to escalate high-risk patients to care coordinators. The system integrates with EHR platforms to push structured data from calls into the patient record and has demonstrated readmission rate reductions in post-surgical and heart failure patient populations in peer-reviewed studies.
Social Buzz
Survey: Americans May Be Losing Trust in AI for Health Care
The survey data comes amid high-profile news coverage of AI-driven prior authorization denials, chatbot errors in clinical settings, and debates over AI replacing clinical judgment. The trust gap is particularly pronounced among older Americans and those with complex chronic conditions — precisely the patient populations where AI tools are most aggressively deployed. Healthcare leaders are being warned that technical performance metrics are insufficient — patient trust must be treated as a deployable system requirement, not an afterthought.
AI in the Mental Health Workforce Is Met With Fear, Pushback — and Enthusiasm
The story highlights new Stanford HAI research finding that AI therapy chatbots may underperform human therapists on therapeutic alliance measures and could reinforce stigma through rigid response patterns. Simultaneously, large behavioral health systems report measurable documentation time savings using AI scribes tuned to psychotherapy notes (progress notes, treatment plans, safety assessments), with some practices reporting 30–40% reduction in post-session administrative time. The field lacks standardized evaluation criteria for AI tools in behavioral health contexts, with no equivalent to radiology's sensitivity/specificity benchmarks for mental health AI.
Healthcare's AI Obsession Is Missing the Point on Nursing Shortages
The piece lands against a backdrop of 250,710 RN shortage and 24% annual nurse turnover costs in 2025. While AI staffing optimization tools (predictive scheduling, acuity-based allocation, early deterioration alerts) are reducing overtime costs and improving patient-to-nurse ratios at deploying sites, workforce researchers argue that technology investment is being used as a proxy for structural workforce reform — delaying mandated staffing legislation in multiple states. The data shows burnout severe enough to consider leaving the profession in one-third of nurses, a metric that AI documentation tools have not moved.