When the Francis Scott Key Bridge in Baltimore collapsed in March 2024, headlines focused on the enormous scale of the disaster: six lives lost, a major port paralysed and a reconstruction bill estimated at $5 billion. Behind the tragedy is a story for healthcare.
Image: NTSB https://www.ntsb.gov/news/press-releases/Pages/NR20251118.aspx
Investigators have traced the initial failure on the 984-foot containership Dali to a single wire. According to the NTSB (National Transportation Safety Board) a misplaced label on the wire prevented it wire from being fully inserted into its gate, leaving an inadequate connection (see figures below). Over more than a decade the wire gradually worked loose. When it finally disconnected completely, the breaker unexpectedly opened, triggering the first blackout. The ship lost propulsion and steering. Four minutes later, after a second blackout and failed attempts to restore power, the Dali calamitously struck the 2.37-mile-long bridge.
Images: NTSB – https://www.ntsb.gov/news/press-releases/Pages/NR20251118.aspx
At the NTSB’s public meeting, the story that was told sounded like something out of a patient safety textbook. Layers of defence were present, but they were thin. Human responses were rapid and competent but constrained by the design of the system. Signals reached some people but not others. Organisational actors defended their role and blamed others. In other words: a classic Swiss cheese model event.
Image: Wikipedia
https://en.wikipedia.org/wiki/Swiss_cheese_model
The Swiss cheese model of accident occurrence says many defense layers typically lie between hazards and accidents. But, there are flaws in each layer that, if aligned, allow the accident to occur. In this diagram, three “hazard vectors” are stopped by the defenses, but one passes through where the “holes” line up.
There’s a lot for the shipping industry to learn from this catastrophe. It also offers a mirror for clinicians, hospital leaders, safety officers and anyone interested in why preventable harm persists in healthcare.
Catastrophe rarely results from one thing, even when it starts with one thing
A wrongly positioned label caused the wire to loosen, but the disaster required several more holes in the cheese:
- Two blackouts occurred, not just one. After the first, the ship began heading to toward Pier 17. The pilots and bridge team attempted to correct the vessel’s direction, but propulsion was lost too close to the structure for human intervention to be effective.
- The breaker that tripped could have been configured to auto-reset within 10 seconds. Instead, the crew had to manually reset it, costing almost a minute.
- Backup generators came online but stalled because they relied on a pump never designed for that purpose. The pump had stopped in the initial blackout and needed a manual restart.
- Communications worked, but only partially. The pilot called a dispatcher, who alerted police, who successfully shut the bridge to traffic, a move that almost certainly prevented many deaths. But no one called the road crew working on the bridge. They had a direct contact number and were only metres from safety.
In healthcare we see the same pattern in medication errors, delayed diagnoses, wrong-site surgery, maternal deaths and airway catastrophes. Something small begins the sequence: a missing lab result, a look-alike medication, an incomplete handover. But serious harm depends on multiple thin or absent defences lining up: a design flaw, a workflow workaround, a communication missed, an assumption that someone else has checked.
The lesson is familiar but still sometimes hard to accept: the last act is the visible one, but the real causes are upstream and layered.
“Human error” is the least helpful explanation
The NTSB report emphasised that the ship’s bridge team and pilots had acted promptly and appropriately. A vessel 300 metres long, drifting without propulsion toward a fixed object, can’t be turned or stopped by human action in seconds. Investigators confirmed that quick action by the pilots, dispatchers and the Maryland Transportation Authority did prevent greater loss of life.
This is a message we still struggle with in healthcare. When something goes wrong, the immediate impulse from leaders, regulators, the public and sometimes clinicians themselves is to focus on the individual: Who made the error? Who failed to act? Who didn’t follow protocol?
But on the Dali, humans were not the root cause. The complex interaction of design decisions, maintenance practices, ill-fitting backup systems, incomplete communication chains and ambiguous responsibilities determined what was possible in those four minutes.
In healthcare the ongoing migration toward systems thinking, including Safety-II, high reliability organisations and human-factors design, supports better understanding and prevention of disaster.
Maintenance debt accumulates silently, until it doesn’t
The loose wire took a decade to inch its way out of place. Investigators also found earlier electrical failures, jury-rigged workarounds and unclear ownership of maintenance responsibilities among the ship’s company, operator and builder. They compared finding this one loose wire among thousands of connections to “hunting for a loose rivet on the Eiffel Tower.”
Image: OpenAI
Hospitals know maintenance debt all too well. Old monitors, outdated infusion pumps, legacy IT systems, low-priority repairs and temporary fixes (“just keep it running until Monday”) slowly accumulate. None of them is catastrophic on its own. But they stack up, interact and lie dormant until the critical moment.
One sobering parallel is the number of adverse events linked to alarms disabled “just for now”, outdated drug libraries in smart pumps or unserviced backup equipment. As with the Dali, these technical debts don’t announce themselves. They erode safety quietly, until the day they break through and cause a tragedy.
Communication failures look small on paper and enormous in hindsight
Police acted quickly and saved many lives by closing the bridge to traffic. But the one call that wasn’t made – to the road crew – is the one now being intensely scrutinised. The crew had a direct contact number but the officer intended to drive to them instead. Too late.
Healthcare has similar moments:
- A critical potassium level phoned to a ward but not to the covering registrar.
- An anaesthetist “sure” someone else called the ICU before an unstable patient is transferred.
- A radiologist flags an unexpected mass but sends the message to the wrong recipient after hours.
These examples don’t amount to negligence. They are symptoms of communication systems that rely on human memory, unclear escalation pathways and assumptions about who is responsible. As with the road crew, people often do their part, but the system doesn’t ensure the right person hears the right thing at the right time.
After the event, the finger-pointing looks the same
In the disaster’s aftermath, the ship’s operator blamed the manufacturer. The manufacturer insisted maintenance is the owner’s responsibility. State officials blamed the ship. The ship’s representatives blamed the state for failing to conduct a vulnerability assessment. Lawyers filed suits in all directions.
The NTSB also identified a critical factor outside the ship itself: the Key Bridge lacked protective countermeasures to reduce vulnerability to vessel impact. When a much smaller ship struck the bridge in 1980, the damage was minor. Today’s ships are ten times larger but the infrastructure has not been updated to reflect the new risk. Nationwide, the NTSB has found many bridge owners to be unaware of their structures’ vulnerability despite longstanding guidance.
In healthcare, the same pattern. After a serious adverse event, different groups often defend their corner: nursing blames pharmacy, junior doctors blame IT, IT blames procurement, management blames clinicians. Even when well-intentioned, the defensiveness makes organisational learning harder.
The Swiss cheese model is not just about holes in defences; it’s about how organisations react, sometimes repeating the same patterns that allowed the accident to occur.
Making thicker cheese
NTSB has issued a broad set of recommendations to US federal agencies, international standards bodies, the ship manufacturer, the ship operator, the electrical component maker, and multiple bridge owners. This was a systemic event requiring systemic action.
The NTSB recommendations align with healthcare safety practice:
- Better design including automatic resets, more robust components, redundancy that actually works.
- Clearer responsibility – who maintains what, who owns each step in the safety chain.
- Resilient communication systems with no single point of failure and no reliance on memory or good luck.
- Routine vulnerability assessments not just after an incident but as a normal part of operations.
- A culture of learning rather than assigning fault, even when lawyers are circling.
Ultimately, the bridge collapsed because multiple thin defence layers lined up at the wrong moment. In healthcare we face the same challenges daily. Every safe day in a hospital depends on preventing small errors from becoming large disasters.