Three posts into this series, the problem should be clear.
Responsibility dissolves in clinical handovers not because people fail, but because systems lack the structural conditions to carry it. Administrative controls ask humans to remember. Engineering controls make the unsafe state unreachable. And at constitutional boundaries, responsibility doesn't just blur. It forks into different legal frameworks with different accountability structures.
This post defines what we've been building toward: the minimum structural conditions that must hold before clinical responsibility can safely move between any two parties. We call this Minimum Viable Responsibility Transfer.
The definition
Minimum Viable Responsibility Transfer (MVRT) is the smallest set of structural conditions that must be satisfied before clinical responsibility can safely move between teams, organisations, or constitutional authorities.
If any condition is missing, the transfer has not occurred. The system must not proceed as though it has.
MVRT is not a protocol. It's not a checklist. It's a set of invariants. Conditions that must hold regardless of care setting, pathway design, or system vendor. They define the structural minimum. Any transfer that satisfies fewer conditions is unsafe by definition.
Five conditions. No exceptions.
The five conditions
Condition 1: Declared sender
A named individual, team, or service must explicitly declare that they are releasing responsibility. This declaration must be active, not inferred from workflow progression.
Completing a discharge summary is not declaring you are releasing responsibility. Sending a referral is not declaring it. Filing a note is not declaring it. These are information actions. Declaration is a governance action: I am handing over responsibility for this patient's care in this specific domain.
Without a declared sender, no one has released responsibility. It may have evaporated, been forgotten, or silently expired. But it hasn't been transferred.
Condition 2: Declared receiver
A named individual, team, or service must be identified as the intended recipient of responsibility. This must be specific enough to be actionable.
"The GP" is not a declared receiver. "Community services" is not a declared receiver. "The next team" is not a declared receiver. These are categories, not actors. A declared receiver is specific: Dr Chen at Elm Road Surgery, or the Eastborough Community Heart Failure Team, led by Claire Ward.
Without a declared receiver, responsibility has been released but not directed. It exists in transit. No one has been asked to hold it.
Condition 3: Defined scope
What exactly is transferring must be bounded and enumerable. Responsibility is not a blob. It has edges.
Is this a transfer of medication management? Monitoring responsibility? Clinical decision-making for a specific condition? Follow-up scheduling? Full care responsibility? The scope determines what the receiver is accountable for and, equally important, what they are not accountable for.
Without defined scope, the receiver cannot know what they've accepted. Accountability becomes total or zero, depending on who's asking. Both states are dangerous.
Condition 4: Verified acceptance
The declared receiver must actively acknowledge that they accept responsibility for the defined scope. Silence is not acceptance. Workflow progression is not acceptance. Information arriving is not acceptance.
This is the condition healthcare most consistently violates. A discharge summary lands in a GP inbox. A referral enters a queue. A care plan is shared with a community team. The system proceeds as though responsibility has transferred because information has moved. But information moving and responsibility transferring are different events.
Without verified acceptance, the sender may have released responsibility but no one has picked it up. The patient exists in a gap where everyone assumes someone else is watching.
Condition 5: Recorded state
The transfer must produce an auditable record showing who held responsibility, for what scope, under which constitutional authority, and the exact moment it changed. This record must be reconstructable after the fact.
When something goes wrong, the first question in any investigation is: who was responsible at that point? If the answer requires retrospective inference, assumption, or negotiation between organisations, the transfer was structurally incomplete.
Without recorded state, accountability becomes a matter of argument rather than evidence. Investigations reconstruct what probably happened rather than what the system showed at the time.
How aviation does this
The image above this article shows two aircraft paths crossing. In aviation, this happens thousands of times a day. Aircraft move between air traffic control sectors, crossing boundaries where one controller's jurisdiction ends and another's begins.
Every sector crossing follows MVRT, even though aviation doesn't use that term. The protocol is structural:
The transferring controller (declared sender) initiates a handoff by identifying the aircraft to the receiving sector. They specify the aircraft's callsign, current altitude, assigned heading, speed, and any restrictions that apply. This is the defined scope: here is exactly what you're taking on, and here are the constraints.
The receiving controller (declared receiver) reviews the incoming aircraft. If they can accept it safely, they send an acceptance. If they can't, they can refuse or negotiate: request a different altitude, a speed restriction, a delay. The transfer doesn't proceed until acceptance is confirmed.
Verified acceptance happens electronically. The system records the exact moment jurisdiction transfers. Until that moment, the original controller retains full responsibility, even if the aircraft has physically entered the new sector's airspace.
Recorded state is continuous. At any point, the system can show exactly which controller was responsible for which aircraft. There is no ambiguity, no assumption, no gap. The record is the source of truth, not people's recollections.
When two aircraft paths cross without incident, it's not because the controllers were skilled or attentive, though they are both. It's because the infrastructure made the unsafe state structurally difficult to reach. You cannot have an unowned aircraft. The system won't allow it.
Healthcare has no equivalent. You can have an unowned patient. The system not only allows it, it hides it.
What healthcare gets wrong
Healthcare violates MVRT conditions so routinely that the violations have become invisible.
No declared sender. Discharge happens as a process, not a declaration. The system generates a summary. The bed is freed. The patient leaves. At no point does a clinician explicitly declare: I am releasing responsibility for this patient's ongoing care.
No declared receiver. The GP is assumed to pick up responsibility by default. But default is not declaration. The GP may not have seen the discharge summary. May not have capacity to act on it. May not agree they're the appropriate recipient for this particular aspect of care.
No defined scope. A discharge summary contains information about what happened. It rarely specifies what exactly the GP is now responsible for. Medication review? Monitoring for deterioration? Follow-up investigations? Wound care? Mental health review? Everything?
No verified acceptance. The summary arrives. The system records delivery. Everyone assumes the handover happened because the information moved. But the GP hasn't accepted anything. They've received an email.
No recorded state. When something goes wrong, the investigation tries to reconstruct who was responsible at the point of harm. The hospital says the patient was discharged. The GP says they hadn't reviewed the notes yet. The community team says they were waiting for a referral. Nobody's wrong. The system simply never established who held responsibility.
This is not a description of a broken system. This is a description of the standard system, working exactly as designed.
MVRT across constitutional boundaries
The five conditions become harder, not simpler, when responsibility crosses from NHS to Care Act authority.
When care moves between constitutional frameworks, two additional attributes must be visible within the defined scope:
Which constitutional authority governs each element of the transfer. Medication management transfers under NHS Act clinical governance. Daily living support transfers under Care Act eligible needs. These may transfer at different moments, to different receivers, with different accountability structures. The scope must distinguish them.
Where authorities intersect on the same action. Falls prevention involves both clinical factors (NHS authority) and environmental factors (Care Act authority). The scope must make this intersection visible rather than hiding it inside a generic "care plan."
Without constitutional binding, integrated care creates exactly the risk surface described in the previous post: boundaries dissolve, legal frameworks merge invisibly, and accountability becomes impossible to reconstruct after harm.
The minimum, not the ideal
MVRT is deliberately minimal. It does not require:
- A specific technology or platform
- Consensus between sender and receiver on the care plan
- A complete patient record
- Integration between systems
- Organisational alignment or cultural change
It requires five things. That's all. If those five conditions hold, the transfer can be shown to have occurred. If any condition is missing, it cannot.
This is a structural argument, not an aspirational one. MVRT doesn't describe what good care looks like. It describes the minimum conditions under which responsibility can be said to have moved. Everything else, the care plan, the clinical communication, the shared decision-making, the collaborative working, builds on top.
You can have excellent clinical communication and still fail MVRT if no one explicitly accepted responsibility. You can have poor communication and satisfy MVRT if the structural conditions hold. These are orthogonal concerns.
The confusion between communication quality and transfer governance is precisely what allows responsibility to dissolve. We keep improving the communication. We keep leaving the governance implicit.
Where this leads
MVRT defines conditions. It doesn't yet describe how to implement them.
The next posts in this series will examine what infrastructure would need to exist for these conditions to hold at scale. Not as a custom protocol bolted onto existing systems, but as a governance substrate beneath them.
The question isn't whether clinical professionals are capable of safe handovers. They are, when the conditions allow it. The question is whether the infrastructure makes it structurally possible for responsibility to move safely, or whether it makes dissolution the default outcome.
Right now, dissolution is the default. Not because anyone chose it. Because nobody built the alternative.
Safe Responsibility Transfers Series
- The Architecture of Trust
- When Safety Becomes Machine Authority
- The Handover That Never Happened
- Minimum Viable Responsibility Transfer (current)
Next in this series: What infrastructure would need to exist for the five MVRT conditions to hold at scale across NHS and integrated care boundaries. View the full series
