In-flight medical diversion rates differ four-fold and nobody can reconcile them

Published diversion rates for in-flight medical emergencies range from 1.7% to 7.3% of events, with a pooled estimate of 11.1 diversions per 100,000 flights. Those figures are not interchangeable — they come from different populations, different denominators, and different reporting streams.

No binding global rule requires airlines to report medical events at all, so neither number is a universal diversion rate. What both datasets do agree on is that the overwhelming majority of in-flight medical events end without an unscheduled landing.

Two studies, two diversion rates that don’t match

The larger dataset comes from JAMA Network Open, which published a cohort study in September 2025. It tracked 77,790 in-flight medical events across 84 airlines between January 1, 2022 and December 31, 2023. Among those events, a diversion followed 1.7% of the time, according to the study indexed by PubMed Central.

The second dataset is older and narrower. Peterson et al. reviewed 11,920 medical-emergency calls logged by five domestic and international airlines between 2008 and 2010, and found that 7.3% of cases ended in a diversion. Same phenomenon, roughly four times the rate.

A third figure complicates the picture further. A 2021 systematic review and meta-analysis — 18 studies covering roughly 1.5 billion passengers — pooled en-route diversion at 11.1 per 100,000 flights.

Why 1.7% and 7.3% can’t be averaged

Start with the denominator. The 1.7% figure divides diversions by reported medical events. The 7.3% figure does the same on its own event set.

The pooled 11.1-per-100,000 figure divides diversions by flights. Converting one into another requires data neither study provides.

Then there is the reporting gateway. Only events that made it to a ground-based medical support center entered the 2022–2023 cohort. If a crew resolved a case on its own, with no consultation and no case file, it left no trace in the dataset.

Airline-to-airline variation in event frequency was wide, the authors note, and they link part of it to how readily each carrier’s crews escalate to a ground consultation. That is a selection effect, and it can push a reported rate in either direction.

A second confound comes from the years between the two studies. The newer rate may be lower because telemedicine improved and airlines revised their medical protocols — or because the threshold for calling ground support has simply dropped. The authors concede as much.

On this evidence, no trend over time can be drawn.

The meta-analysis compounds the problem. It reports an I² of 97% — heterogeneity so high the authors graded the underlying evidence low-certainty. A pooled number built from studies that disagree that much is a rough guide, not a benchmark.

The reporting gap that makes both numbers provisional

No binding worldwide rule requires an airline to report every in-flight medical event. What exists is guidance — recommended forms, anonymous submissions, voluntary adoption — and that gap is the structural reason two credible studies can land so far apart without either being wrong.

The consequence for a reader is simple. Neither the 1.7% nor the 7.3% figure describes a universal diversion rate, because neither study saw the whole picture. Each counted the events its own reporting gateway happened to catch.

Who actually decides to divert

The decision is never automatic, and it is never made by the medical side.

When a passenger becomes seriously unwell, the crew typically contacts a ground-based medical support service for clinical guidance. That consultation shapes what happens next. It does not decide it. Sole authority sits with the pilot in command, who weighs the medical picture against fuel remaining, weather at candidate airports, and what local medical infrastructure is actually available on the ground.

One documented case shows the chain in motion. In September 2026, United Airlines Flight 142 was bound from Tokyo Narita to Denver when the crew reported a passenger medical emergency over the Pacific. The Boeing 787 set down at Seattle-Tacoma International Airport instead, as media reported citing the Federal Aviation Administration; the passenger later died, People reported. The case confirms a diversion can happen on a trans-Pacific sector. It says nothing about how the cockpit weighed the call.

Two factors shift the odds. Neurological and cardiac conditions are the categories most consistently linked to unscheduled landings. And when a physician volunteer happens to be among the passengers, a diversion becomes statistically more likely — an association the data does not explain, and not a cause.

What this means for your next long-haul

For a Western traveler booked on a long Asia-Pacific sector, the useful conclusion is reassurance built on mechanics rather than marketing. If a medical emergency happens on your flight, the base rate says it almost certainly will not end in an unscheduled landing, whatever the true diversion rate turns out to be.

It matters most on trans-Pacific and trans-Asian routes, where diversion options are sparse. The factors in play — fuel load, weather at candidate airports, ground medical capacity — are exactly the ones that get tight on a nine-hour ocean crossing. A diversion out there is a deliberate trade, not a reflex.

Cost explains part of the caution. A 2021 systematic review put average diversion costs at $15,000 to $893,000 per unplanned emergency landing, drawing on the studies that examined cost. Other published estimates run narrower, from about $3,000 to more than $500,000, and the two ranges belong to different source sets — they should not be merged into a single current figure.

What they share is the reason airlines think hard before turning a widebody around. The bill lands in the six figures often enough to matter, which is worth remembering the next time a flight continues to destination instead of landing short.

Key terms

Diversion

A diversion is an unscheduled landing at an airport other than the flight’s ticketed destination. In commercial aviation it is a deliberate operational decision, taken when continuing to the planned destination is judged riskier than landing early. For a traveler, a published diversion rate is best read as a property of the dataset behind it rather than a fixed odds of landing short.

Ground-based medical support center

A ground-based medical support center is a service that flight crews contact by radio or satellite link for clinical advice during an in-flight medical event. Staffed by physicians and nurses, it gives real-time guidance on whether a case can be managed in the cabin or warrants a diversion. Because such centers only see the events crews choose to escalate, a rate built on their records captures consulted cases rather than every onboard event.

Pilot in command

The pilot in command is the captain legally responsible for the aircraft and everyone aboard during a flight. Under international aviation practice that authority is final, and no ground service, dispatcher, or medical consultant can override the captain’s decision. In a medical emergency, this is why the diversion call rests with the cockpit even though a ground-based medical service supplies the clinical advice.

Enplanements

Enplanements are the number of passengers boarding flights, counted each time a traveler boards rather than once per person. Regulators and researchers use enplanements as a denominator to express event rates per unit of passenger traffic. Because enplanements count boardings rather than flights, a rate expressed per million enplanements is not directly comparable to a rate expressed per flight.

Meta-analysis

A meta-analysis is a statistical method that combines the results of multiple separate studies into a single pooled estimate. It is used when individual studies are too small or too varied to settle a question on their own. When the studies being combined disagree as much as those in the 2021 review cited here, the pooled figure is a rough guide rather than a benchmark.

Questions? Answers.