AI in Healthcare Pulse — 2026-03-28
This week saw the FDA clear a new Philips AI system for real-time cardiac guidance, state-level AI regulation continued to fill federal gaps, and two major funding rounds totaled over $165 million for AI health infrastructure. Oncology AI deployments and a novel patient-facing GPT tool also made headlines.
AI in Healthcare Pulse — 2026-03-28
Regulatory & Policy Watch
1. FDA Clears Philips AI for Real-Time Heart Valve Repair Guidance
- What happened: The FDA cleared a Philips AI solution designed to provide real-time guidance to clinicians during complex, minimally invasive heart valve repair procedures. The system integrates with Philips' EchoNavigator imaging platform to assist physicians intraoperatively.
- Impact: The clearance expands the footprint of AI-assisted cardiovascular intervention, giving interventional cardiologists real-time decision support during high-stakes procedures that previously relied entirely on physician judgment and conventional imaging. This represents a meaningful step toward AI as an active procedural co-pilot rather than a post-hoc diagnostic aid.
2. States Take the Lead on Healthcare AI Regulation
- What happened: As federal regulatory frameworks remain in flux, states have stepped up to establish their own guidelines governing the use of AI in health settings. A new analysis published this week details the patchwork of state-level activity filling the void left by federal ambiguity.
- Impact: For AI healthcare companies operating across state lines, this fragmented landscape creates compliance complexity. Companies must now track and adapt to multiple regulatory regimes simultaneously, increasing legal and operational overhead — particularly for startups seeking to scale nationally.
3. Penn Medicine Faculty Flag "Holes" in Federal AI Healthcare Regulation
- What happened: Faculty at the University of Pennsylvania's medical school publicly voiced concerns this week about current federal AI regulatory frameworks, calling attention to specific gaps and proposing potential solutions. The commentary appeared in The Daily Pennsylvanian.
- Impact: Academic and clinical voices are increasingly joining the policy conversation, adding pressure on federal agencies to clarify oversight of AI diagnostic and decision-support tools. The commentary signals that the medical establishment views existing frameworks as inadequate for the pace of AI deployment in clinical settings.
Clinical Frontlines
Hartford HealthCare (Connecticut) — PatientGPT Launch for Lab Result Explanation
- The AI: Hartford HealthCare, a Connecticut-based health system, launched "PatientGPT," a patient-facing AI tool designed to explain lab results in plain language, identify potential medication interactions, and answer patient health questions.
- Results: The system is described as "pioneering" by the health system. Specific outcome metrics have not yet been published, as the deployment is newly announced.
- Significance: PatientGPT represents a notable move toward AI-mediated patient engagement — shifting AI's role from purely back-end clinical decision support to direct patient communication. If validated, this model could reduce call center burden and improve patient health literacy at scale.
CancerNetwork — AI Evolution in Oncology Clinical Workflows
- The AI: A detailed review published this week outlines how specialized large language models (LLMs), AI-assisted CT scanning for early cancer detection, and foundational pathology models are being integrated into oncology workflows across clinical sites.
- Results: The review highlights that AI-assisted CT is enabling earlier detection, while foundational models are described as "democratizing pathology" — potentially extending access to high-quality pathological analysis beyond major academic medical centers.
- Significance: The combination of LLMs for clinical trial matching, imaging AI for earlier detection, and democratized pathology could dramatically compress the timeline from symptom onset to treatment initiation for cancer patients, particularly in under-resourced settings.
Medscape / Clinical Research — ML Model for Liver Cancer Risk Prediction
- The AI: A machine learning model using routine clinical data was reported this week to predict liver cancer risk more accurately than existing screening tools. The model gauges risk from data already collected in standard care, requiring no additional testing.
- Results: The model outperformed existing risk-stratification tools and showed particular promise in identifying high-risk patients who would have been missed by current screening criteria.
- Significance: Liver cancer is frequently diagnosed at late stages when outcomes are poor. A model that can flag at-risk individuals earlier — using only existing clinical data — has the potential to dramatically improve survival rates without adding cost or workflow burden.
Funding & Deals
Qualified Health — $125M (undisclosed round)
- What they do: Qualified Health is a startup that partners with health systems to evaluate, build, and manage artificial intelligence tools — essentially serving as an AI infrastructure and governance layer for hospitals.
- Investors: Specific lead investors were not disclosed in available reporting.
- Why it matters: The $125M raise is one of the largest health AI infrastructure rounds in recent months. It signals strong investor conviction that health systems need dedicated support to navigate AI adoption — not just the AI models themselves, but the operational scaffolding around them. As more hospitals move from AI pilots to production deployments, companies like Qualified Health occupy a critical intermediary role.
Doctronic — $40M Series B
- What they do: Doctronic is a New York-based AI doctor startup known for a provocative experiment in Utah in which a chatbot handled drug prescription renewals. The company is building AI systems intended to handle primary care tasks autonomously.
- Investors: The Series B was co-led by Abstract and Lightspeed Venture Partners.
- Why it matters: Doctronic's fundraise reflects continued high-risk, high-conviction bets on autonomous clinical AI. The Utah prescription chatbot experiment put Doctronic in the national spotlight and drew both enthusiasm and regulatory scrutiny. The $40M raise suggests investors believe the company can navigate those headwinds and scale its model.
Research Spotlight
"Rethinking Health Agents: From Siloed AI to Collaborative Decision Mediators"
- Published in: arXiv (preprint, submitted March 2026)
- Key finding: The paper argues that AI in healthcare should be repositioned not as an autonomous decision-maker but as a "mediator" — a system that surfaces contextual information, reconciles different mental models held by clinicians and patients, and scaffolds longitudinal decision-making over time. The authors propose a conceptual framework for designing AI systems that are ethically grounded and collaborative rather than directive.
- Clinical relevance: As AI systems move deeper into clinical workflows, the question of how AI and clinicians share authority becomes critical. This framework offers a practical design philosophy for developers building AI that can be trusted in high-stakes settings — and may inform how health systems govern AI tool deployment.
"The Role of Agentic Artificial Intelligence in Healthcare: A Scoping Review" — npj Digital Medicine
- Published in: npj Digital Medicine (Nature Portfolio)
- Key finding: This scoping review systematically examines the emerging literature on "agentic AI" — systems capable of operating autonomously to achieve defined clinical goals. The review finds significant conceptual confusion in the field distinguishing AI agents from agentic AI, and notes that few studies have rigorously evaluated these systems in clinical contexts.
- Clinical relevance: Agentic AI is poised to be the next frontier in clinical deployment, with systems that could autonomously order tests, triage patients, or coordinate care. This review provides a critical baseline assessment of where the science actually stands — and flags the evaluation gaps that must be addressed before widespread clinical adoption is responsible.
What to Watch Next Week
- State AI regulation momentum: With multiple states now actively developing healthcare AI guidelines, watch for additional state-level announcements or model legislation proposals. A coordinated multi-state framework could emerge as an alternative path to federal standards.
- Doctronic regulatory interactions: Following the $40M raise and the ongoing scrutiny of its Utah prescription chatbot model, watch for any FDA communications or state-level regulatory responses to Doctronic's approach to autonomous prescribing.
- Qualified Health expansion: With $125M in fresh capital, Qualified Health is likely to announce new health system partnerships. Watch for signals about which regions or hospital networks are moving fastest on AI infrastructure investment.
- Agentic AI in clinical settings: The npj Digital Medicine scoping review and the arXiv preprint both point to agentic AI as a rapidly developing frontier. Watch for early-stage trials or pilot announcements from health systems beginning to test autonomous AI agents in bounded clinical tasks.
Reader Action Items
-
For healthcare AI developers and vendors: The Penn Medicine faculty commentary and state-level regulatory activity underscore that federal inaction does not mean regulatory inaction. Begin mapping your product's exposure to state-level AI healthcare laws now — particularly if you operate in states with active legislative calendars. Compliance complexity will only grow.
-
For health system administrators and CIOs: The Qualified Health raise signals a maturing market for AI governance infrastructure. Before deploying additional AI tools, evaluate whether your organization has the operational scaffolding — evaluation frameworks, monitoring protocols, and vendor management capabilities — to manage AI tools responsibly at scale.
-
For clinical researchers and practitioners: The liver cancer ML model and the oncology AI review both highlight a pattern: AI tools built on routine clinical data are outperforming disease-specific screening criteria. Review your specialty's current screening and risk-stratification protocols with fresh eyes — the next high-impact AI application in your field may already have the data it needs.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Create your own signal
Describe what you want to know, and AI will curate it for you automatically.
Create SignalPowered by
