Slashdot Mirror
Flaw In Emergency Response System May Have Killed Hundreds
Hugh Pickens writes "The Telegraph reports that a flaw in the way emergency response software was set up to handle Category A responses in Great Britain may have cost hundreds of lives over the past ten years. Most ambulance services use an international computerized system designed in America and in the US version, a fall of more than 6 feet receives the maximum priority response. However, the government committee which governs its use in Great Britain decided that such cases should be deemed less urgent, and excluded from an eight minute category A target response time. If a call involved a fall of more than 6 feet it was designated a lower priority 'category B response' despite the presence of life-threatening conditions which were supposed to receive the most urgent category A response. The flaw came to light after Bonnie Mason, 58, fell 12 feet down the stairs and died from a head injury after emergency controllers in Suffolk failed to identify her situation as 'life-threatening.'"
The summary sounds like "we underestimated how dangerous a medium distance fall can be, so we didn't have the correct priorities and more people died than could have". That isn't really a flaw in the algorithm, it's just a flaw in one specific parameter in the algorithm.
The system itself wasn't flawed, but rather whoever set it up decided that they should be category B. The system did exactly what it was told, it just was told to do something different than in the US, and something that was later deemed to be suboptimal.
How is this a flaw in the Emergency Response System if the change initiated by a government committee is how the incidents were classified wrongly?
Support NYCountryLawyer RIAA vs People
Caller: Please hurry!! He's fallen down a 30ft well! Can't you get here any faster!?
A&E Drone: *clackety clackety* ...... Computer say Nooooo....
May the Maths Be with you!
I think that the comments I have read above me are missing the point, or maybe I am.
The software was changed so that falls of more than 6 feet no longer caused a case to be considered "category A", the problem is that (through a mistake when rewriting that bit of code I assume), mention of a fall was causing a case to be considered "category B" even if other things were present that would normally have made it "category A".
You're right: it isn't a flaw in the software per se, and I would not assign any blame to those who developed it (as opposed to those who implemented it).
However, it is a predictable of administration, and the use of information technology is often integrated into systems in just this way. The idea that risk can be rationalized and reduced to a number (class A, class B, and so on) is itself potentially dangerous. Though it is not necessarily dangerous in any particular situation, it is nevertheless predictable that administrative or technical rationality would make this kind of outcome more common.
You see, the problem was not simply that the response categories were incorrect. The problem was that the system (including its operators, operating procedures, and so on) was too rigid, too rationalized, and therefore unable to respond to unexpected events:
This kind of event was clearly unexpected by the systems implementors. But even if they had assessed the danger of falls differently, there is likely some other event that would fall outside the systems parameters. (Most falls probably should be category B events, not category A.) That's why you want to have human judgement and human overrides.
Treating a system in terms of independent technical components has a number of benefits, including efficiency. That's what happened here. The process was rationally divided into tasks for the humans and tasks for the computer. Nice, neat, clean: and likely to produce outcomes like this.
At least the Brits can be credited for the genius of their new number and catchy jingle. Oh one one eight, nine nine nine--eight eight one nine nine, nine one one nine seven two five! .... three.
The actual event in question happened a year ago. Given the recent news in the USA - something to do with some sort of bill about healthcare, and the imminent UK general election, I find the timing from a right wing newspaper here in the UK to be highly sensational - especially since the issue has been corrected in the new version of the software, that was released and rolled out last year.
Because there were two things which went wrong:
1) falls from over six foot were deemed to not be that serious but also
2) a computer error (in program or configuration) caused that to override anything else.
That is, under the decisions made by the bureaucracy the rules should have said:
1) if someone has fallen from a great height it is a type A (highest class) emergency
2) if someone has fallen from a moderate height it is a class B emergency unless there is something else which elevates it to a class A
3) if someone has fallen from ground level (slipped or tripped), it is a class B emergency unless there is something else which elevates it to class A
4) if someone is having trouble breathing, it is a class A emergency
5) is someone is bleeding profusely, it is a class A emergency
(etc).
The rules as implemented by the system actually effectively said:
1) if someone has fallen from a great height it is a type A (highest class) emergency
2) if someone has fallen from a moderate height it is a class B emergency
3) if someone has fallen from ground level (slipped or tripped), it is a class B emergency unless there is something else which elevates it to class A
4) if someone is having trouble breathing, it is a class A emergency
5) is someone is bleeding profusely, it is a class A emergency
(etc).
Note the shortened version of rule 2 and why this is relevant: moderate falls were being categorized as class B even when there were other risk factors. In the example case, we had an ambulance which was on its way to treat a woman who had fallen a moderate height and was bleeding profusely. It was diverted to instead help a woman who had slipped and fallen and was having trouble breathing because the first was classified as urgency level B and the first as urgency level A. This is clearly an error in either the software code, configuration, or the use of the program (most likely the configuration, I would guess, but that's just speculation on my part).
The committee's decision was that moderate falls shouldn't automatically elevate to class A. In practice, the system was assuming that moderate falls always were class B. So something was going wrong with the program, its configuration, or its use.