Dr. Chuck Tuchinda, EVP, COO and deputy group head at Hearst Health, and chair of FDB (First Databank)
Photo: FDB
As artificial intelligence applications proliferate across healthcare, the model context protocol is an emerging industry standard that defines how AI systems, large language models and agent-based applications connect with trusted knowledge sources.
By adopting MCP, provider organizations can deploy agentic AI tools faster, and with more safety, consistency and scalability, says Dr. Charles Tuchinda, EVP, COO and deputy group head at Hearst Health, and executive chairman of FDB (First Databank), which offers drug knowledge that helps healthcare professionals make precise medication decisions.
"The model context protocol is the missing connective tissue that makes AI in healthcare safe, useful and scalable," Tuchinda explained.
"The next wave of clinical AI will not come from bigger models but from smarter connections – connecting those models to trusted, governed sources of truth and clinically proven workflows such as prescribing, medication reconciliation and verification. AI systems are only as good as the context they can securely access and apply.
"Today, every integration between an AI system and a data source – whether a drug database, formulary or decision-support tool – is custom-built," he continued. "Each introduces variation, delay and risk. MCP changes that by defining a standard language for discovery, permissions and interaction."
Automatic understanding
With MCP, an AI agent can automatically understand what a resource offers, how to use it safely and when it's appropriate to call it – all through a secure, auditable and governed framework, Tuchinda explained.
"In simple terms, MCP is like a 'FHIR-for-AI,'" he said. "Just as FHIR standardized how healthcare systems exchange data, MCP standardizes how AI interacts with those systems. It creates an AI-native layer of context awareness and governance that lets organizations control what an AI can see and do, ensuring every recommendation is explainable, traceable and aligned with clinical standards.
"This matters because healthcare requires verifiable, accountable intelligence," he added. "MCP gives organizations the ability to govern which knowledge sources an AI can access, enforces policies on agent behavior and maintains full transparency. It ensures that generative AI does not make things up but operates only within the boundaries of trusted, clinically validated content."
Hospitals and health systems can deploy agentic AI systems faster, and with more safety, consistency and scalability, by using MCP, Tuchinda contended.
"Every health system faces the same challenge: Integrating AI into clinical workflows requires months of custom development, redundant validation and manual review," he explained. "Each implementation introduces variation and risk.
"MCP solves this by providing a standardized framework that defines how AI systems securely connect with trusted, validated knowledge sources, tools and workflow," he added.
What does a source provide?
With MCP, any compliant AI agent can automatically discover what a source provides. For example, for FDB, it would be the company's drug knowledge, decision-support tools or prescribing workflows. MCP then uses those resources safely within governed parameters. This, Tuchinda explained, results in:
- Speed. Deployments that once required months of custom integration can now be completed in weeks using repeatable, standards-based patterns.
- Safety and consistency. Using governed clinical content ensures that every AI-driven decision is grounded in validated data, with auditable actions and clear traceability already relied upon across thousands of health systems.
- Scalability. Once integrated through MCP, drug knowledge and services, for example, can scale seamlessly across new systems, departments and vendor applications without re-engineering.
"MCP enables healthcare organizations to move from pilot projects to enterprise-scale AI adoption." Imagine a busy clinician finishing a patient visit and saying, "Let's start you on amlodipine 5 milligrams once a day for your blood pressure," said Tuchinda by way of example.
"In most clinics today, that single sentence triggers a cascade of manual steps: Someone retypes the prescription into the EHR, runs a separate interaction check, confirms pharmacy routing and updates the patient instructions," he noted. "Each step adds time, friction and risk."
AI performs in seconds
"Now picture the same encounter in an MCP-enabled world," he said. "As the physician speaks, an ambient AI assistant connected through the MCP instantly recognizes the prescribing intent. It securely queries trusted drug knowledge and automatically drafts a structured prescription for review."
In just seconds, the AI has:
- Verified dose appropriateness for the patient's age and kidney function.
- Checked for interactions and contraindications.
- Suggested cost-effective or clinically appropriate alternatives when needed.
- Routed the order through an e-prescribing network, sending it to a patient app or pharmacy.
"The clinician simply reviews and approves with a single click," said Tuchinda. "The patient leaves with clear instructions, fewer delays and virtually no risk of transcription error. Behind the scenes, every AI action is permissioned, governed and auditable because MCP allows the agent to connect to trusted drug knowledge and apply the correct clinical criteria for safe decision making.
"That is the value of MCP," he explained. "It transforms fragmented, manual steps into a single, intelligent, governed workflow that pairs deep clinical expertise with the agility of modern AI."
Follow Bill's health IT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
The HIMSS AI & Cybersecurity Virtual Forum is free to attend on Nov. 18. Learn more and register.
WATCH NOW: Managing government policy and regulation in healthcare AI adoption
