144 Deaths, One Database Nobody Trusts

What happened

Between January 2000 and December 2013, the US Food and Drug Administration collected reports tying surgical robots to 144 patient deaths and more than a thousand injuries. Researchers pulled these figures straight from the agency's own adverse-event archive, working through fourteen years of submissions filed by hospitals, patients, and device makers for a study published in 2015. The authors were careful to note their count sits well below the true total, since it only reflects incidents someone bothered to report.

The number that should worry people isn't the death toll. It's how little the filings actually explain. More than a thousand of the recorded incidents involved broken or burned instrument parts falling into a patient's body during an operation, a category with an obvious mechanical signature that accounted for over a hundred injuries and at least one death. For the deaths as a whole, though, investigators found the paperwork thin: most filings gave no real account of what caused the harm.

The UK's Royal College of Surgeons challenged the study on solid ground. The researchers never benchmarked robotic outcomes against comparable procedures done without a robot, so there's no way to know from this data alone whether the machines carry more risk than a surgeon's own hands. The paper also skipped peer review. Both objections hold up. Neither one touches the deeper problem: the source records were too sparse to answer the question in the first place, regardless of who reviewed the math.

That's the actual failure worth sitting with. The FDA's system leans on the hospitals and manufacturers involved to report their own equipment's failures, completely and honestly, and most of them didn't fill in the part that mattered. Once the cause field is blank, a regulator can't tell a software bug from a training gap from an unrelated complication during surgery. A raw count of deaths with no attached explanation doesn't prove the robots were dangerous, and it doesn't clear them either. It just shows that the reporting pipeline was never built to settle the argument, only to log that something happened.

Reported impact

Affected parties

Not publicly disclosed

Harm type

Not publicly disclosed

Scale

Not publicly disclosed

Financial impact

Not publicly disclosed

Regulatory action

Not publicly disclosed

Classification

Organization

Not publicly disclosed

AI system

Not publicly disclosed

Industry

Not publicly disclosed

Country

Not publicly disclosed

Provider

Not publicly disclosed

Incident type

Not publicly disclosed

Relevant governance controls

Governance control mapping is not available for this record.

• No controls mappedNot publicly disclosed

Control mapping is analytical. It does not state that any control would have prevented the incident.

Sources and evidence

This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.