David Learmount/CAPE TOWN

Learning only from serious accidents and incidents is a flawed way of advancing flight safety. It took until the 1990s to create a system which is more effective and workable, and until now to persuade most of the world's regions to consider adopting it. The system is data based safety decision making.

At the Flight Safety Foundation (FSF) international air safety seminar (IASS) in Cape Town, South Africa, last month it was revealed, however, that the only two remaining world regions yet to set up a local flight safety organisation - without which safety standards usually remain static - are among those which need it most: Africa and the South-East Asia/Asia- Pacific region. The logic - born out by practice - is that without a forum there is no communication, therefore no information exchange, therefore no basis for decisionmaking, therefore no action.

Meanwhile, in the USA, Federal Aviation Administration Administrator Jane Garvey has just announced a long-awaited programme mandating the use of flight data recorder (FDR) information to advance what the FAA calls flight operations quality assurance (FOQA).

Downloading FDR data may be an effective analytical tool, but is not the only way of gathering useful, useable safety information, however. The first principle of today's precision approach to improving flight safety is that policy decisions should be based on any intelligently gathered and analysed data which takes account of far more than just major accidents and their primary causes.

Once lessons have been extracted from the information, the other basic principle is applied. This consists of setting priorities for corrective action. The priorities simply acknowledge the reality of limited resources, and aim first at those accident categories which statistically pose the greatest danger to human life and cause the most aircraft damage.

Targeting safety priorities is not new. It came before the systems to enable data driven safety programmes had been widely put in place. For example, the FSF started its campaign to reduce controlled flight into terrain (CFIT) in 1992, because it did not take quantum physics to work out that CFIT was - and still remains - the biggest killer category. In fact, information presented by Boeing at the IASS, reviewing jet hull loss accidents in 1998, revealed that an annual downward CFIT trend has been reversed even before the year is over. It also reveals that approach and landing accidents are proving stubbornly resistant to the FSF campaign currently being waged by its Approach and Landing Accident Reduction task force (see bar chart).


Many individual nations have long kept massive libraries of accident investigation information, but until the arrival of desktop data processing, this sort of information not been widely turned into the usable data now available. Since the mid-1980s, by airline choice rather than national policy, a system pioneered during the 1970s by the UK Civil Aviation Authority has brought computer analysis of safety data into the realms where it has the greatest potential to do good: data regularly downloaded from digital flight data recorders (FDRs) and quick access recorders (QARs) to gain aircraft health monitoring and operational information. The data on its own is useless, however and needs to be analysed using specialist software to determine trends.

This final data processing step allows the "accident waiting to happen" to be seen before it does. It is also a step which much of the world appears ready to take, because the methods and technology are sufficiently mature, and the will to implement it more widely is gathering momentum.

So when the FSF, together with the International Air Transport Association (IATA) and the International Federation of Airworthiness (IFA) conceived the theme for their Cape Town safety conference, they chose "Making a safe system safer", and the speakers were analytical in their approach. Kicking off, the FAA presented the rationale for its Safety Analysis and Strategic Intervention (SASI) programmme, which deputy director aircraft certification service Elizabeth Erickson described as "a structured data-driven process to continually reduce accident rates". The SASI is a product of the "Safer Skies" initiative by the Commercial Aviation Safety Strategy Team (CASST), an FAA/industry partnership. Perhaps the greatest shortcoming of sophisticated flight safety monitoring systems is that they have the potential to intimidate with an excess of acronyms.

Following Erickson at the seminar came the European Joint Aviation Authorities' (JAAs) Koos van der Spek, certification director, to describe the JAA's new Safety Strategy Initiative (JSSI). Van der Spek says that the JSSI's task is to set priorities for the JAA's regulatory activity, on the simple premise that it cannot do everything at once.


Both these programmes, representing the latest policy thinking in the world's two most influential aviation authorities, have in common the principles of decisionmaking based on hard data, followed by setting priorities for action from a list of safety objectives. As van der Spek puts it: "Interventions will be based on analyses of existing databases. These analyses will define more specific issues or 'hot spots' [and] will respect human factor principles. The main rationales...are that the list should be acceptable by the European authorities [governments/ European Commission] and needs as far as possible to be harmonised with Safer Skies."

Each agency independently drew up a priorities list for study and action. CFIT action headed them both, but the complete lists were similar, perhaps reflecting van der Spek's hope that the two programmes would be "harmonised". The fact that there are differences reinforces the argument that safety decisionmaking should be regionally based and orientated.

The JAA says that its first priority is with large turbine aircraft (JAR25 operators), since they carry the majority of the JAA airlines' passengers. Leaving aside CFIT, the JAA's action list includes:

Approach and landing accidents; loss of control accidents; design related accidents; weather related accidents; occupant safety and survivability.

This list, the JSSI notes, covers 89% of all European accidents. The FAA has given SASI a set of guidelines, of which the principal ones are to identify the "critical few interventions" needed to achieve an 80% reduction in accidents. In other words, 80% of the US accidents studied came into the following categories which makes up the SASI "intervention" list:

Loss of control; uncontained engine failure; runway incursions; approach and landing; weather.

Finally, Erickson says: "The focus must be on prevention". That is the aim of Garvey's proposed FOQA programme based on the FDR data.

It could be argued that only the nations or regions where safety performance is already good need a sophisticated analytical tool to decide how to improve safety standards further, because their safety oversight systems are already in place and the operating infrastructure is in good condition, so achieving improvements requires well-aimed action.

In the areas of the world which perform poorly, however, simply getting the oversight and infrastructure in place would be a good start. In continents such as Africa, which remains bottom of the world safety league despite early signs of progress, it tends to be perfectly clear what needs to be done. ICAO and IATA agree that, in Africa, with a few national exceptions, almost everything needs to be improved.

That, however, is no argument for failing to prioritise action. Indeed, as IATA's director for infrastructure in Africa Trevor Fox makes clear, when resources are particularly sparse, setting priorities is even more important. His priorities for the continent's infrastructure could be summed up in two items:

Improved communications including links between air traffic control centres, ground to air, and the provision of a good aeronautical fixed telecommunications network; depoliticisation of air traffic services, making them autonomous cost centres to which user charges flow direct.


Fox does not attempt to prescribe for safety oversight or to talk about flight operations quality assurance.

With the infrastructure in place however, safety oversight provision would come next. Compared with a world hull loss/fatal accident rate of 1.8 per million departures during the past 10 years, according to Boeing figures presented at the Cape Town conference, most African states exceed that by a massive amount. The worst, Sudan, shows 45 hull loss/fatal accidents per million departures. Boeing also presented a projection for the number and causes of African hull loss and fatal accidents to large commercial jets for the next 10 years (see pie chart). More than half of them will be landing accidents, if Boeing's projection is correct.

It may be easy to see, without sophisticated safety analysis tools, that landing accidents are Africa's main flight safety scourge. It is true that any regionally developed safety strategy needs to work on the basis that a system must be able to "walk before it can run". But if there were a system in place to analyse FDR data, Africa would know why landing is so dangerous there and be able to change the way things are.

Source: Flight International