Flying razor blades

It is frequently claimed that corrugated galvanised iron (CGI) sheets are ‘flying razorblades’ in storms, and that they shouldn’t be used in reconstruction. This has caused their complete prohibition, for example after the 1991 Cyclone in Bangladesh. However, they are durable, relatively cheap, and usually very popular with affected people. So what is the truth, and should they be used in recovery and reconstruction projects?


Having looked for evidence, I have found absolutely no evidence that CGI is any more dangerous than any other flying debris in windstorms. In fact the leading causes of death in windstorms are drowning (due to storm surges and wind-induced flooding), falling trees and falling debris (walls falling over, tiles falling off roofs, etc). Where data is available there are cases of injury and death (more injuries than deaths) from windborne debris, but not one that cites windborne CGI as the cause.

There is no reason to avoid the use of CGI in post-disaster reconstruction. Rather than banning CGI, reconstruction efforts should focus on effective evacuation of flood-prone areas and on stronger, safer buildings in general.

Details of the evidence I found are below:

Wind-borne debris is extremely hazardous for buildings

It is well established that wind-borne debris is a serious hazard for buildings. Objects carried by the wind can puncture the walls and windows of buildings, allowing the wind to get inside, potentially destroying the building. However, none of the literature or evidence states that CGI is any more dangerous than anything else.


“Wind-borne debris generated from loose-laid ground materials and/or from dislodged building components can fly an unbelievable distance, up to several hundred feet, and impact neighboring buildings. Almost all building envelope materials, particularly windows, doors and skylights, as well as most wall sidings and roof coverings used for homes, are at risk for impact from such high-energy projectiles. Breaches to the building envelope by such projectiles will allow rainwater intrusion and subsequently damage the building interior. It may also pressurize the interior of the building, increasing the uplift on the roof and the outward forces on the walls, making them more likely to fail. In high category hurricanes (e.g. category 3 or higher), wind-borne debris damage occurs essentially in parallel with the damage due to differential pressures.”

1992andrew6 Gabriel_Fernandez_del_Pino_CARE (66)
Flying debris from hurricane Andrew (left) and Typhoon Phailin (right). Clearly if these hit you, it will cause injury or death.

Damage to buildings from windborne debris (Tamura 1991)

Wind Debris TableStructural design codes have special provision to ensure that fixtures, fittings and cladding of buildings is well enough attached so it does not become detached and form flying debris (see right, for example, from IBC 2009).

Is is well established that in built-up areas windborne debris from one building can hit the next, and so on, causing what is know as the damage chain.

Capture1Damage to roofs, and the damage chain (Tamura 1991)

See also:

But does wind-borne debris kill lots of people?

The description of the effects of different category hurricanes mentions wind-borne debris injuring or killing people, and common sense says that any solid object flying at 60+mph that hits people will severely damage them. But what’s the actual evidence?

  • Category One Hurricane: Loose outdoor items will become projectiles, causing additional damage. Persons struck by windborne debris risk injury and possible death.
  • Category Two Hurricane: Loose outdoor items will become projectiles, causing additional damage. Persons struck by windborne debris risk injury and possible death.
  • Category Three Hurricane: Persons struck by windborne debris risk injury and possible death.
  • Category Four Hurricane: Windborne debris will cause extensive damage and persons struck by the wind-blown debris will be injured or killed.
  • Category Five Hurricane: Severe injury or death is likely for persons struck by wind-blown debris.

Most of the studies I can find focus on disasters in rich countries, as it is only richer countries that have systematic collection of mortality and morbidity information.

  • One paper, notes that 23% of the 63 fatalities from Cyclone Mireille, which struck Japan in 1991, were due to flying debris (Tamura 1991). The paper also gives a good overview of the damage chain. The paper does not mention CGI.
  • A systematic review into the health impacts of windstorms (Goldman 2013) in the UK concludes that becoming airborne, being struck by flying debris or falling trees and road traffic accidents are the main dangers, with a number of secondary effects like electrocution also being significant. The review also finds that flying debris causes far more injuries (mainly fractures and blunt trauma) than it does fatalities. Most of the fatalities from flying debris are due to head injuries. The review does not mention CGI, but largely only references research from the UK and USA.
  • A 1984 paper reviewed the multiple casualties caused by a gale in the UK: “Four patients were injured by falling slates and twelve by other falling or flying objects, including window glass, a tree, a milk-crate and a dust-bin.” This was out of a total 116 injuries treated. (Illingworth 1984)
  • Another similar paper found 36% of injuries were due to flying debris and falling masonry (Cugnoni 1992)
  • Medical examiner reports in the US report that 9 deaths were caused by blunt or penetrating trauma in Hurricane Andrew (1992) and 2 of 18 deaths were caused by flying debris in Hurricane Charley (2004).
  • Hurricane Katrina killed over 1,200 people in the USA in 2005. One study, looking at 971 deaths in Louisiana, found that drowning (40%), injury and trauma (25%), and heart conditions (11%) were the major causes of death. The mechanism of injury was unspecified for 226 trauma or injury deaths; specified injury-related causes of death included heat exposure (n 6), unintentional firearms death (n 4), homicide (n 2), suicide (n 4), gas poisoning (n 3), and electrocution (n 1), but numbers due to flying debris are unfortunately not mentioned. (Brunkard 2008). Interestingly the paper strongly links likelihood of death to age, poverty and social exclusion.
  • Hurricane Sandy in 2012 killed around 160 people. The CDC examined the causes of 117 deaths captured by American Red Cross Mortality Tracking. Of the 117 deaths, 67 (57.3%) were classified as directly related deaths, and 38 (32.5%) were indirectly related to the storm. Of the directly related deaths, the most common mechanism was drowning (40 [59.7%]), followed by trauma from being crushed, cut, or struck (19 [28.4%]). Poisoning was the most common indirectly related cause of death; of the 10 poisonings, nine were caused by carbon monoxide. Most directly related deaths occurred during the first few days of the storm, whereas indirectly related deaths continued from the day before the storm into the middle of November. More information is provided in an Earthsky article, based on data from the New York Times. Flying debris was not one of the top 7 causes of death: Capture

So what happens when you look at the most deadly windstorms, which don’t tend to happen in rich countries, but for which the data is much harder to find:

  • The 1970 Bhola Cyclone in Bangladesh killed up to 500,000 people. Sommer and Mosley undertook two surveys after the cyclone looking mainly at mortality, but also at morbidity. The primary causes of death were drowning, although they did note a ‘cyclone syndrome’ where trauma was caused by hanging on to trees and being buffeted by the floodwaters. (Sommer, 1972). In fact six of the top ten deadliest windstorms of all time have hit Bangladesh, with the lead cause of death drowning. (Washington Post, 2013)
  • Typhoon Haiyan/Yolanda hit the Philippines in 2013 and killed 6,300 people. The National Disaster Risk Reduction and Management Agency final report states simply that most deaths were caused by drowning and trauma. It includes an extract, but not the entirety, of the deaths reporting, containing details of 163 deaths. The cause of over half of these is unknown and the leading known cause of death is drowning. Falling debris and blunt or penetrating trauma represent only 6 of the known deaths:Yolanda cause of death


Windborne debris is dangerous, but it is not the leading cause of death or injury in windstorms. Drowning, falling trees and collapsing walls/roofs are more significant causes of death.

CGI, rather than general windborne debris, is not cited as the cause of injury in a single one of the papers or reviews of storm mortality and morbidity I have found. There is no evidence that CGI is any more dangerous than any other windborne debris. The only way to limit wind-borne debris, especially important in built-up areas to avoid subsequent damage to adjacent buildings, is to ensure all parts of a building are tied down and robustly attached.

There is no reason to limit use of CGI in post-disaster reconstruction, but there is every reason to ensure it is used correctly and buildings are well-built.


KBrunkard J; Namulanda G; Ratard R, Hurrican Katrina Deaths, Louisiana, 2005, American Medical Association (2008)

Cugnoni & Whitworth, Injuries related to wind speed, Annals of the Royal College of Surgeons of England (1992)

Goldman A. et al, The health impacts of windstorms, a systematic literature review, Public Health (2013)

Illingworth R N; Illingworth K A, Multiple casualties caused by a gale, Archives of Emergency Medicine (1984)

Sommer A; Mosley W H, East Bengal Cyclone of 1970 Epidemiological Approach to Disaster Assessment, The Lancet (1972)

Tamura Y, Wind-induced damage to buildings and disaster risk reduction

Thanks to the ‘community of shelter geeks’ for the discussion that started this off!

Does ‘building back better’ mean what you think it means?

This was first published in a slightly different form on CARE International’s Insights blog. 

After natural disasters the phrase ‘Build Back Better’ is a constant refrain from politicians, donors, aid agencies and the media. This short, alliterative phrase has captured the imagination, and seems at first glance to be a simple, powerful and necessary principle.

Sitting on the panel of a recent event at the Overseas Development Institute looking at the response to Typhoon Haiyan/Yolanda in the Philippines, I said that CARE doesn’t use ‘Build Back Better’, and for want of a better phrase it prefers ‘Build Back Safer’. Why is this, and what could possibly be wrong with it?

For the comprehensive answer, the best thing you can do is read The Meaning of ‘Build Back Better’: Evidence From Post-Tsunami Aceh and Sri Lanka. It’s behind a paywall though, so here’s a shorter explanation:

Evidence from a succession of disaster responses has shown that the simple rhetoric of building back better is extremely difficult to apply in highly complex situations. While better sounds good, it means very different things to different people.

The media tend to represent it as building houses that will never again fall down in a storm, or earthquake, or flood. Earthquake-proof, wind-proof and other guarantees are frequently mentioned. Unfortunately there’s no such thing (as evidenced by the consequences of Hurricanes Katrina & Sandy in the US and the 2012 tsunami in Japan – even the best-prepared, richest nations cannot build strong enough to resist all hazards).

Humanitarian aid workers tend to think it means reducing risk and increasing resilience: building stronger homes in safer places, with access to sustainable livelihoods, health and social services, and involving and ensuring the protection of men, women, girls and boys from any dangers they may face.

But many disaster-affected people, when told someone is going to (help them) build a better house, might think of modern materials like concrete or masonry instead of timber, or mud, or wattle & daub. They might think of a second storey, a bathroom and an indoor kitchen, with running water and electricity. They might think of artistry and decoration on their houses that will show them to be more successful and wealthy. They might think of internet access, or somewhere to entertain guests, or glazed windows. And who is to say they are wrong? We, as aid workers, here today and gone tomorrow, do not get to decide what is better.

There’s a strong ethical obligation on those who come to give help after disaster to be honest and transparent about what they can and cannot do, and what they will and will not do. The word better gets in the way of that. It raises very high expectations and leaves us unable to deliver what people think we have offered.

The shelter sector now prefers to say that we (humanitarian shelter actors) will seek to ‘build back safer’. Still not a perfect phrase, it is less ambiguous, it is easier to explain, and it more accurately reflects what we can actually achieve with the funding and time typically available to us. Building back safer involves building houses which have improved capacity to survive hazards, and working with the community to improve their ability to respond to and recover from shocks themselves. It involves making achievable, steady, incremental improvements in the safety and resilience of communities and their infrastructure. It involves making people less vulnerable and putting them on a firmer footing from which they can continue to build what they know is better for themselves.

Is aidwork really a profession?

J (@TalesFromthHood) recently published a post on Aidspeak explaining why using unqualified volunteers in overseas aid and development work is always a bad idea. I won’t re-state his points here, you can read the post.

I agree with everything J says. I’ve seen the effects of unqualified people (both paid and unpaid) on projects and the people involved in them. I also strongly agree that aid work should be done by professional, qualified people.

Some of the comments in the post noted that paid aid-workers can themselves be incompetent, harmful or a waste of time. Others accused aid-workers claiming that only professionals can do aid-work of being arrogant. Of course some aid-workers are incompetent, harmful or a waste of time. The same can be said for some teachers, or some doctors, or some engineers. Teachers, doctors or engineers would never be accused of being arrogant for denying well-meaning volunteers the opportunity to do their work. Both these claims are may or may not be true when applied to individuals, but are clearly not valid arguments for using volunteers.

Nonetheless, I have serious concerns about describing aid-work as a profession. This upsets me, because it really does need to become more professional. Aid-work is, at best, a profession in its infancy.

The dictionary definition of a profession is:

“A paid occupation, especially one that involves prolonged training and a formal qualification”

Or, alternatively:

“a type of job that requires special education, training, or skill”

Most professions are regulated in some way by industry bodies which register members of the profession, ensure standards and accredit training & qualifications. To become a chartered structural engineer, I had to spend four years at university and a further four years doing professional training, followed by day-long long professional exam with a typical pass-rate of less than 30%. I have to do continuing professional development, which I have to record and is regularly audited. If I don’t meet the stringent standards of my profession I can be fined or struck off. I’m liable under civil and criminal for the quality of my work.

To become an aid-worker someone can get an internship with the HR department of a humanitarian NGO straight after university, transfer to the programme department and be working ‘in the field’ within a couple of years, regardless of qualifications or even suitability for the work. This is a real example, and I wish it was an aberration.

Of course, the majority of aid workers have very relevant qualifications. Many are highly-qualified specialists. Many have done master’s degrees. Many have years of professional experience more valuable than any formal qualification. Aid work is a job that requires special education, training, and skill. It cannot be done by unqualified people, and shouldn’t be, regardless whether they are paid or not.

Most aid-worker are professional, but aid-work, to my mind, does not (yet) qualify as a profession, because it doesn’t have consistent or rigorous ways to ensure the professional standards of its members. It doesn’t even have a way to register its members.

Aid-work needs to move away from its current state of being a highly skilled community of practice. The humanitarian community has developed professional standards, it has organisations which capture and share knowledge and learning. It has all the trappings of being a profession. It is, perhaps, a profession in its infancy. I wish it would go one step further and establish a professional body to represent & regulate the professionals and ensure professional standards are maintained. Perhaps we could then finally stop re-hashing the volunteering/voluntourism arguments over and over again. How does the Institution of Chartered Aid-Workers sound?

Does health and safety matter in other countries 2?

This is the second of my two blogs about health and safety. The first one is here.

Part 2: Construction & Sport

The Guardian newspaper has recently been reporting on the conditions construction workers have to bear working in the Middle East; with a strong focus on Qatar and the 2022 world cup. Many of the workers are effectively slaves, with passports and pay withheld and no rights, under a system called Kafala. Accident rates are hard to find, but the Guardian has established that 382 Nepali workers died in Qatar in 2012 and 2013 (with the real figure likely to be higher). Nepali workers make up only one sixth of Qatar’s entirely immigrant construction workforce. That suggests to me that possibly thousands have died. This is a scandal, or it should be. Sadly the conditions construction workers are subjected to in the Middle East have been well known the construction industry for as long as I’ve worked in it[1]. It’s similar in many countries around the world, especially those that rely heavily on migrant labour.

The fact is that it’s only on highly visible projects like major sports events where anybody takes any notice. That however means that these events are an opportunity to raise standards and awareness; and indeed there is evidence that such improvements occur at many games. The European Institute for Construction Labour Research has investigatedpractices at several Olympic Games and notes that often the opportunities for workers are good and the pay and conditions better than normal; but it also notes some pretty shocking practices. Even in Beijing it is noted that improvements were made on normal practices. Athens is an exception, where construction practices and safety took a big step backwards, while London set new standards and achieved record low accident rates and zero fatalities, for the first time in the Modern Games’ history.

The table below shows the best data I’ve been able to find about fatality rates at various games. The background fatality rate in the wider construction industry is given where known.

Games Deaths Background rate: deaths per 100,000 per year, from ILOStat Reliability
Barcelona ‘92 2 26 (1999)10.6 (2008) 12 deaths have been claimed by theEuropean Institute for Construction Labour Research, but it notes that the accident & fatality rate was lower than for normal construction in Spain.Sources: 1, 2
Atlanta ‘96 1 10 (2008) Can’t find any official statistics – at least one person was killed. Sources:1, 2
Sydney ‘00 1 4.4 (2008) Sources: 1, 2, 3
Athens ‘04 14 Not known Real number likely higher, as documentation of deaths is not good. Higher numbers are widely claimed. Sources: 1, 2, 3
Beijing ‘08 6 Not known Confirmed by China, but a figure of 10 or more is claimed. Sources: 1, 2
Delhi Commonwealth Games ‘10 43 Not known Real number may be higher. Sources: 1
Vancouver ‘10 1 8.7 (2008) In a road blast accident. Source: 1
South Africa World Cup ‘10 2 Not known Source: 1
London ‘12 0 4.5 (2006) One worker died on site of a heart attack unrelated to his work. Sources: 1, 2, 3
Brazil World Cup ‘14 9 26.4 (2000) Source: 12, 3
Sochi ‘14 25+ Not known Sources: 1, 2
Brazil ’16 0 Not known
Russia World Cup ‘18 3+ Not known Sources: 1
Qatar ‘22 100s Not known Sources: 1, 2, 3
Know any better sources? Please add a comment below – I’ll update the table.

Against a background of continual improvement, and sports events generally improving on the status quo, Sochi and Qatar seem to be very regressive steps. FIFA has been forced to admit that the conditions in Qatar are “absolutely unacceptable” – a tacit admission also that it must do better in choosing hosts in the future. Let’s hope it lives up to this. Qatar has been forced to improve conditions for workers, but will only do so for those working directly on the stadia – a pathetic fraction of the workforce involved in constructing the World Cup infrastructure. There are no plans to make lasting changes to the Kafala system.

There has been quite a lot of research into what made the London 2012 construction so safe, and it’s freely available. The major international sporting bodies and their sponsors must take an active role in promoting these lessons and the large body of knowledge about safety in construction, and in promoting good working practices. And not just promoting, but requiring. It can’t just be down to human rights organisations to campaign for such improvements, and major events are such a big opportunity to showcase how well things can be done.

Sport is meant to bring us together and celebrate human achievement. It surely can’t do that when the whole event is built on the suffering and exploitation of the builders.

Does health and safety matter in other countries 1?

This is the first of a two-part blog I’m writing about health and safety. It’s forty years since the UK’s Health and Safety at Work Act 1974 came into force. It, and subsequent legislation, has transformed all UK construction places, but has had a particular effect on construction sites.

Part 1: Construction safety & NGOs

We’ve all seen them: the proud photos in fundraising literature of people in disaster-stricken countries rebuilding with money from relief agencies. In bare feet or sandals, without gloves, balanced on the roof, young men are helping provide new houses, or a school, or a clinic. This is evidence of people being helped. It’s definitely a good thing.


But is it good enough?

Construction work is consistently amongst the most dangerous professions in the world. From the ILO[1]:

“In the construction industry, at least 60,000 people are fatally injured on building sites every year. Many hundreds of thousands more suffer serious injuries and ill health. In fact, these estimates are conservative. In many countries, less than 20 per cent of construction injuries are reported, and little account is taken of the longer-term impact of occupational diseases. The main causes of fatalities in the sector include falls, fatal crush injuries and the impact of falling objects, and electrocution. Major health problems range from deafness to vibration syndromes, back injuries, other musculoskeletal disorders, and exposure to hazardous substances (solvents, isocyanates, pesticides in timbers, chemical treatments for damp courses, fire retardants, welding fumes) and to dust and fibres (cement dust, silica, wood dust, fibreboard and, worst of all, asbestos). Stress is a frequent problem brought on by the other factors, notably the fear of falling. Construction workers tend to live away from home in substandard accommodation, especially in the developing countries. Tuberculosis, cholera, dengue, malaria and HIV/AIDS can therefore pose particular risks.”

The construction industry is dangerous everywhere. In the UK, which has world-leading health and safety practices, only agriculture and fisheries have a higher fatality rate, and the number of injuries and people suffering long-term ill effects is large[2]. The UK has strong legislation to protect construction workers. The Health and Safety Executive will prosecute breaches of legislation and workers are well insured and will be compensated if injured at work.

In most countries around the world this is not the case. Most construction workers are uninsured and cannot expect support if hurt because of doing their jobs. NGOs build quite a lot; they employ trades people and labourers all the time. Some NGOs build a huge amount, it’s pretty much what they do. In most cases (not all), the safety provisions they put in place for local labour are minimal. Rarely is decent footwear provided. Scaffolding, to allow safe working at height, is usually considered too expensive (and remember that falls are a leading killer). Training in safe working and good site management, which are totally free – are not often seen.

None of this would be acceptable in the rich countries in the north, but somehow it’s deemed too much trouble and expensive in poor countries. This while in countries with no social security systems, with no safety nets, the impact of an injury at work can be destitution for whole families.

If NGOs are serious about disaster risk reduction, perhaps they should start with reducing the very real risk of disaster to individuals working in construction by training them in safe construction methods and by providing them with safe working environments. I’m not suggesting we can achieve the very low accident rates of the UK; after all it’s taken the UK over 40 years of concerted effort to get there. But surely we can do much better, and we shouldn’t accept dangerous working practices just because the workers are poor?

The ILO has a code of practice for Safety & Health in Construction[3][4], which should ideally be followed, but there are some simple things NGOs should ensure are included in the budgets and project plans for all projects involving construction:

  • Train workers in safe working practices.
  • Provide, on loan, personal protective equipment (hard hats, gloves, high-visibility jackets, steel toe-capped boots, safety glasses).
  • Ensure scaffolding is provided for all working above ground, or at least minimise working at height.
  • Maintain a tidy site, to prevent trips and falls.
  • Have access control for the site to keep children out, ensure nobody is below people working above (and to keep track of your stuff).

None of these are difficult, none are expensive. It would be out of the question not to provide them in the west. There’s no excuse for not providing them anywhere else.


PS. It’s been said to me before that decent footwear is too expensive, and surely it’s not that important? This is what happens if your skin is exposed cement – I warn you, it’s not pleasant: