'How can we stop the carnage on the sites?' screamed a headline in Construction News last year after statistics showed that it was likely - as proved the case - that there would be more deaths on construction sites in 2000 than in the previous year. The horrific deaths of four workers, whose gantry collapsed as they were working on the Avonmouth Bridge, drew further attention to the issue.
In considering a response, comparison with the rail sector is instructive. Two months after the Avonmouth Bridge accident (which merited little news coverage), four people were killed in the Hatfield rail crash. Notwithstanding the great improvements in speed and safety that had taken place on the railways in recent years, the aftermath of this rare tragedy led to a heavy-handed precautionary approach to mainstream working practices and a ritualistic requirement for strict compliance with rules and regulations.
The concentration on zero-risk operational practices has decimated the rail service. Timetables have evaporated and costs are due to increase, which we are told we should be thankful for in the name of greater safety.
Gerald Corbett, the ill-fated chief executive of Railtrack, said: 'Performance must take a back seat to safety.' But this motto, whether applied to railways, the construction industry, or life itself, is in danger of clouding real improvements, which should make us a more relaxed and rational in our approach to health and safety - though not complacent.
Aside from a few blips, the safety trend on construction sites has shown a remarkable improvement over a long period (see graph below).
Even when the devastating recessionary shake-out in the industry in the early '90s is taken into account, the number of deaths on site have, in general, decreased.
In the past five years, there have been, on average, 83 deaths on building sites a year, compared with an average of 104 in the previous five years. A 1989/90 peak of 165 deaths was down to 68 deaths in 1998/99.
And although there is no full and proper explanation for the rise to 85 deaths in 1999/00, it was in tandem with the continued increase in employment totals within the sector.
Notwithstanding last year's figures, long-term trends in the fatality rate per 100,000 employees suggest that building sites are becoming safer places to work. The rate is on a par with fatalities per 100,000 workers in the combined 'agriculture, hunting, forestry and fishing' sector.
It is significant to note that fatalities rose slightly in 1996/97, when the Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 1995 (RIDDOR) came into force. The figures for major injuries rocketed to 3,227 in the same year and to 3,857 the following year - double the average of the mid '90s - and this is cited by some commentators as evidence of the continuing dangerous nature of the sector. But these increases can be placed in context by looking at the significant variations in recording requirements of RIDDOR during this period.
The latest version of RIDDOR has dropped the previous stipulation that 'immediate admission' to hospital be required to classify an accident as 'major'. Less speedy transfers are now logged.
More importantly, a new range of injuries has become reportable, including those resulting from acts of physical violence at work. And the statistics now include figures from the Chemical and Hazardous Installations Division and the Nuclear Safety Directorate, which were not there before.
Even taking all this into account, the rate of major accidents has decreased to 120.1 per 1,000 employees, the lowest rate since the introduction of RIDDOR, and the trend is downwards.
It is true that a worker is six times more likely to be killed in the construction industry than in manufacturing, but, given the nature of the business of construction, is that really that surprising? While employers who flaunt safety standards should be prosecuted fully, and employees who pose a threat to themselves and others by their actions reprimanded, it is worth remembering that the fatality rate in Britain is one third of that on construction sites in mainland Europe. Falls are the most common cause of death in the construction industry (averaging 48 deaths per year) - proportionately this is only twice the fatality rate from falls in the home (1,500 people per year).
The confusing introduction of the CDM regulations has had no real effect on the general downward trend of deaths on site and employment levels. Maybe this should call into question the effectiveness of legislative frameworks on work practices and the disproportionate amount of time, money and effort they entail.
Currently, health and safety risk assessments are completed more to shift responsibility than to save lives.
The onus is on the employee to read the health and safety plan, which is kept on site for reference. Employee failure to read the employer's risk assessments, it could be argued, brings danger - and liability - on themselves.
Another effect of the precautionary approach to risk is for shrewd contractors to postpone straightforward tasks on site, because their own risk assessment determines that it is 'too hazardous', which can lead to 'legitimate' time extensions with all the consequent damages and inconvenience.
While there is no cause for complacency - 80 deaths a year in any sector is unsatisfactory - attention should be shifted away from paper shuffling. This might enable safety to become a natural corollary of work practices, rather than their driver.
For example, to address the construction industry fatalities caused by a fall from a height the solution is not just to train people in the use of ladders, or to formulate a standard risk assessment.A reduction in fatalities is more likely to come from easing up on time pressures, profits and subcontractors' margins; improving site supervision (from architects, clerks of works and contractors); and, yes, harnesses. This would have greater effect than any exhortations of a planning supervisor.
The recent invention of the health and safety professional has bolstered some architects' incomes in the short term, but led to an overinflated view of risks in the sector. Simple, straightforward precautions are, of course, sensible, but overemphasis - or overinflation of the problem - could lead to a paralysis of common sense.