Our resident statistician reveals which occupations and countries have the highest risk of being killed on the job.
Health and Safety. I know, I know, I can already hear your groans. But bear with me please. We all know what dismal reputations those “killjoys” and “jobsworths” have in health and safety. In the UK, the Health and Safety Executive (HSE) is often forced to dispel “myths” about rigid rules it is said to have enforced, such as banning cast members throwing sweets into the audience during pantomimes.
But if we conduct a safety analysis of the workplace using the statisticians’ trusty tool of the micromort, defined as a one-in-a-million chance of sudden death, we can begin to appreciate what various initiatives have actually achieved.
For instance, these initiatives appear to get little popular credit for the reduction in industrial accidents over the decades. In 1974, when the UK’s HSE was formed, there were 651 employees killed at work, an average risk of 29 micromorts per year for an employee. By 2010 this figure had fallen to 120, which is 5 micromorts per year, an impressive 82% decrease in risk. (So you know, 5 micromorts is roughly equivalent to the risk of travelling 30 miles (48 kilometres) on a motorbike, or a single scuba-dive.) People who are self-employed carry more risk, and 51 were killed in 2010, which bumps up the overall UK average to 6 micromorts per year.
Britain comes out well in comparison with its European Union counterparts. The statistical office of the EU, Eurostat, reports that in 2007 British workers were on average exposed to 10 micromorts per year (excluding road transport deaths), compared with 17 in France, 19 in Germany, 26 in Spain, 35 in Poland and 84 in Romania. Of course the low UK figures could be because nobody in modern Britain does anything but stand in shops or sit behind computers.
On the other side of the Atlantic, the US Bureau for Labour Statistics provides some extraordinary statistics on the fate of 130 million workers in 2010. A total of 4,547 workers were killed, a rate of 35 micromorts per worker per year. The most common cause was highway accidents, which are excluded from the European figures: without these the rate falls to around 28 micromorts per year, around that of Spain.
But, believe it or not, the second most common cause of death, larger than falls or being hit by things, is “assault and violent acts”. This comprises 18% of all work-place fatalities, and includes 506 homicides (this was down from 860 homicides in 1997). So this means that each year US workers have on average around 4 micromorts risk of being murdered at work – not much less than the total risk to UK workers.
Comparing these figures with the wider world is tricky, as reliable statistics are hard to come by. For example, the International Labour Organisation (ILO) said that India reported 222 fatal accidents at work in 2005, while the ILO reckoned the true number was nearer 40,000.
Out of 2 billion workers worldwide in 2008, the ILO estimated that there were 317 million injuries requiring more than four days absence, and 320,000 people killed while at work, although they only heard about 22,000 of these deaths through official channels and so had to guesstimate the rest. This makes it an average of 160 micromorts per year per worker.
But it is important to bear in mind that all these figures are averages. This includes vast armies of workers toiling over computers, which although causing stress, repetitive strain and back injury, rarely actually kill you on the spot.
Some occupations, though, are a different matter.
Apart from niche jobs such as providing security in unstable regions of the world, the highest-risk occupation in the UK today is, believe it or not, commercial fishing. A recent study found 160 deaths in commercial fishing in the UK between 1996 to 2005, which works out as 1,020 micromorts per year per fisherman. This is staggeringly high – about the same as the risk British coal-miners faced in 1911. Commercial fishing is also the most dangerous job in the US, with a risk of 1,160 micromorts per worker in 2010. Perhaps surprisingly, being a police officer was only the 10th most risky job in the US, at 180 micromorts a year.
We have looked at how dangerous some occupations can be, but how risky does something have to be before an organisation such as the HSE decides that something should be done? The HSE’s philosophy is based on what is known as the Tolerability of Risk framework, and can be rather nicely related to micromorts.
Basically, potential hazards are thought of as being on a spectrum of increasing risk, divided into three fairly loosely defined regions. At the top are “Unacceptable” risks – where something must be done to protect either the workers, or the public, or both, whatever the benefits of the activity may be. At the bottom are “broadly acceptable” risks, which are not exactly zero but are considered insignificant – the sort of thing we would regard as normal in our daily lives.
In between these extremes lie the “tolerable” risks – those that we may be prepared to put up with if there is sufficient benefit, such as providing valuable employment, personal convenience, or keeping the infrastructure of society going. After all, somebody has to do the dirty work.
Acceptable or not?
But how does anyone decide what is unacceptable, broadly acceptable, or tolerable? This cannot be necessarily put into numbers, but again we can turn to the HSE because they have come up with some rough rules-of-thumb.
First, they state that an occupational risk might be considered “unacceptable” if the chance of a worker being killed were greater than one in 1,000 per year. This equates to 1,000 micromorts per year, around the current level for commercial fishing. “Exceptional groups” are excluded from this: presumably serving in a war-zone counts as exceptional, for example, when the average risk faced by 10,000 servicemen and women in Afghanistan reached 33 micromorts a day, or around 10,000 a year.
For members of the public (rather than employed workers), the HSE considers a one in 10,000 a year risk – which is 100 micromorts a year – as being generally unacceptable. At the other extreme, risks are considered broadly acceptable if they are less than one in a million per year – or 1 micromort. This, by the way, is the current estimated risk of being killed by an asteroid.
But even such a minimal risk, if applied to the whole population of the UK, would mean around 50 deaths a year. So what would happen if these occurred all at once, say disasters with multiple fatalities, hazards that affect vulnerable groups such as young children, or risks imposed upon people just because of where they live?
In these circumstances, “societal concerns” can trump these cold micromort calculations. And again, the HSE has come up with some rough guidance that reflect these concerns: for any potential hazard, the risk of an accident involving 50 deaths should be less than 1 in 5,000 per annum. So let us say there is a dam, which, if it burst, would be likely to kill 50 people. In this case it needs to be designed so that this could be expected to happen only once every 5,000 years. This is a fairly stringent criterion. If 10,000 people live below the dam, that is less than 1 micromort a year each.
From an individual perspective this might be considered as being “acceptable”. But because we do not like disasters (although people clearly love reading about them in newspapers) then huge amounts of money are spent to make tiny risks even tinier.
So, perhaps it is time to unfurrow those brows and ungrit those teeth, and say hurray for health and safety!