On this page
Published: May 30, 2024 · Last reviewed: May 1, 2026
Key Takeaways
- Every interface to another vendor's system extends trust beyond your control boundary.
- Input validation matters more when the sender is not your product.
- Shared protocols like HL7 and DICOM were designed for trusted networks.
- Availability failures propagate across connected systems faster than confidentiality ones.
- Environmental assumptions must be stated in labeling, not assumed silently.
- Shared responsibility boundaries belong in your architecture views and your customer documentation.
Interoperability creates cybersecurity risk because each connection extends trust to a system you do not control and cannot patch. Every interface your device offers to another vendor's product is an attack surface, and every assumption you make about that product's behavior is a claim you may be asked to substantiate. Managing this means authenticating peers, validating inputs, and stating your assumptions in labeling rather than relying on them silently.
Reviewed September 17, 2026
Interoperability is a clinical requirement. A monitor that cannot send vitals to the record, a pump that cannot receive orders, and an imaging system that cannot share studies all create manual work and transcription errors. Hospitals buy connectivity for good reasons.
The security consequence is that your device's risk profile now depends partly on products built by companies you have never met, running software versions you cannot see, on a network administered by someone else. Your submission has to show you have thought that through.
Why Interoperability Changes Your Risk Position
The February 3, 2026 premarket cybersecurity guidance expects manufacturers to consider the environment their device operates in, including the systems it connects to. That expectation exists because connected devices fail together. A misbehaving peer can flood your device with traffic, send malformed messages that crash a parser, or supply data your device treats as authoritative.
The IEC 80001 family of standards was developed for exactly this reason: risk management for medical devices on IT networks is a shared activity between the manufacturer and the healthcare delivery organization, because neither controls the whole picture. The hospital owns the network; you own the device's behavior on it. Problems arise where each side assumes the other has handled something.
The same shared-responsibility logic drives the FDA's interest in labeling. If your device is safe only when it sits on a segmented network with controlled peers, that condition is part of the device's safe use, and it belongs where the customer will see it.
Where Trust Breaks Down
| Interface | What you extend trust to | What can go wrong |
|---|---|---|
| HL7 messaging | The sending system's message content | Malformed or malicious content parsed as clinical data |
| DICOM associations | Any peer that can reach the port | Unauthorized query, storage, or bulk retrieval |
| Middleware and gateways | An intermediary that reframes your data | Loss of end-to-end integrity and attribution |
| Cloud integration APIs | A third party's authentication and scoping | Over-broad access tokens; data readable beyond intent |
| Shared hospital network | The hospital's segmentation and hygiene | Lateral movement from an unrelated compromised host |
| Third-party AI or analytics | Another vendor's handling of your data | Data retained, moved, or altered outside your control |
Input validation is the control that pays for itself across all of these. A parser that assumes well-formed input because the sender is another medical device is trusting a claim it cannot verify, and message parsers have historically been a reliable source of crashes and memory safety bugs. Validate structure, bounds, and plausibility on everything that crosses the boundary, regardless of who sent it.
[KEY REQUIREMENT] Authenticate the peer, not the network. "The traffic came from inside the hospital VLAN" is a statement about network topology, not about identity, and it is not a control you administer.
Availability Is the Property That Propagates
Interoperability failures are usually availability failures, and availability failures spread. When an integration engine goes down, every device feeding it queues or drops data. When a device floods a network segment, everything on that segment suffers. When a clock source drifts, timestamp-dependent workflows break across multiple systems at once.
That matters for risk rating. Under ISO 14971, the harm is what reaches the patient, and delayed data, missing alarms, and unavailable priors are clinical events rather than IT inconveniences. A threat model that considers only confidentiality has missed the failure mode most likely to occur.
Designing for it means defining what your device does when a peer is unavailable or misbehaving. Does it buffer, and for how long? Does it alarm locally? Does it degrade to a safe standalone mode? Those behaviors are design decisions worth documenting, because a reviewer asking about network dependency wants to see that you chose them deliberately.
Documenting Shared Responsibility
| Item | Who owns it | Where it must appear |
|---|---|---|
| Device authentication of peers | Manufacturer | Architecture views, threat model |
| Input validation on all interfaces | Manufacturer | Design documentation, testing evidence |
| Network segmentation | Healthcare delivery organization | Labeling as a stated assumption |
| Peer system patching | Healthcare delivery organization or other vendor | Labeling and customer security documentation |
| Configuration of your security features | Usually the customer | Installation and administration documentation |
| Vulnerability notification for your product | Manufacturer | Vulnerability disclosure and management plan |
Labeling is doing real work in that table. An assumption that appears only in an engineer's head cannot be verified by anyone, and it cannot be relied on in a risk argument. Writing it down converts it into something the customer can act on and something the reviewer can evaluate.
Be specific rather than generic. "Deploy on a secure network" tells a hospital nothing. "This device must be deployed on a network segment that restricts inbound connections to the following ports from the following systems" tells them exactly what to configure and gives you a defensible basis for the assumption.
Practical Controls That Hold Up
See also: NeuroTech Cybersecurity Risks, QNX Vulnerabilities in Medical Devices, and Cellular Connectivity Device Security.
The controls that survive contact with a real hospital network are the ones that do not depend on the network behaving well.
Authenticate every peer connection, with credentials unique to the deployment rather than shared across your product line. Encrypt every interface, including internal ones, because segmentation you do not administer is not a control you can count on. Validate every input against structure and bounds before using it. Apply rate limits so a flooding peer degrades throughput rather than availability. Log security-relevant events in a form the hospital's tooling can ingest. Define and test the device's behavior when its peers are unreachable.
None of these are exotic. What makes them difficult is that each one is a place where security and interoperability pull against each other, since the most compatible configuration is usually the most permissive. Documenting why you chose your position on that tradeoff is what a submission needs, more than a claim that no tradeoff was made.
How Blue Goat Cyber Approaches This
We test interoperable devices at the interfaces, which means sending the messages a misbehaving peer would send rather than the ones your integration partner sends. That covers protocol handling, authentication of peers, input validation, and behavior under peer failure, with findings written so the shared responsibility boundary is explicit.
Our threat modeling work maps each external connection to what it trusts and what crosses it, and our medical device penetration testing exercises those interfaces directly rather than through the application layer above them.
Frequently Asked Questions
Why is interoperability a cybersecurity risk?
Because each connection extends trust to a system you do not control, cannot patch, and cannot observe. Your device's security then depends on another vendor's implementation, on the hospital's network administration, and on assumptions about both. Every interface is also an attack surface reachable by anything else on the same network.
Are HL7 and DICOM secure protocols?
Both support security mechanisms and neither requires them. HL7 version 2 messaging has no built-in authentication or encryption and relies on the transport and network beneath it. DICOM supports TLS and identity negotiation but permits open associations. In both cases, the security of a deployment depends on configuration choices rather than on the protocol defaults.
Who is responsible when two connected devices create a risk together?
Responsibility is shared, which is the premise of the IEC 80001 family of standards. The manufacturer is responsible for the device's own behavior, its authentication and validation, and for disclosing the conditions under which it operates safely. The healthcare delivery organization is responsible for the network and for the overall risk of the connected system it assembles.
What should our labeling say about network requirements?
Exactly what a hospital would need to configure: required segmentation, ports and protocols used, which inbound connections must be permitted and from where, what the device does when peers are unreachable, and which security features the customer must enable. Generic advice to use a secure network is not actionable and does not support a risk argument.
How should we handle a vulnerability in a peer system we do not control?
Through coordinated disclosure to that vendor, plus an assessment of whether your device's exposure changes because of it. If your device becomes more exposed when a peer is compromised, that path belongs in your threat model, and your customer communication should say what compensating action a hospital can take.
Does interoperability testing differ from normal penetration testing?
The emphasis differs. Interoperability testing sends malformed, unexpected, and unauthorized messages across each integration interface, tests peer authentication, and observes behavior when peers fail or flood the connection. Standard application testing typically exercises the intended client behaving correctly, which does not reproduce the conditions a real network produces.
Connected to Systems You Do Not Control?
Every integration is a trust decision. We can show you which of yours are unstated assumptions and what a reviewer will ask about them. Book a strategy session.
Blue Goat Cyber specializes in medical device cybersecurity, from threat modeling and penetration testing through premarket submission support. Our team works exclusively with device manufacturers preparing FDA submissions. Learn more about Christian Espinosa, our founder and CEO.
About the author
Christian Espinosa, MBA, CISSP · Founder & CEO, Blue Goat Cyber
U.S. Air Force Academy graduate and veteran with 30+ years in cybersecurity. Founded Alpine Security in 2014 (acquired 2020), then Blue Goat Cyber in 2022. Has supported 275+ FDA medical device submissions; no client has failed to clear due to cybersecurity. Author of three books including The Smartest Person in the Room. Ironman triathlete and mountaineer.