‘Incident prevention’ regarded as key to ‘accident prevention’

For safety, the year 2007 can be described as the good, the bad and the ugly, according to U.K.-based consultancy Ascend. The group tracks accidents worldwide and produces a year-end review of accidents in the form of a “Special Bulletin.”

Paul Hayes

“2007 was ‘good’ in that, overall, it was a relatively safe year and continued the trend of recent years to fewer accidents,” said Ascend’s Paul Hayes. “However, from the point of view of insurers, things could now have turned ‘ugly’ with the incurred losses during the year exceeding premiums.”

The Ascend data focuses on transport category aircraft, both Western and Eastern built, including jets and turboprops.

The accident history for 2007 is good, in that the fatal accident rate, compared to that of the early 1990s, has been halved. For passengers, the fatality rate for 2007 has turned out to be the second best of all time, being better only in 2004 (for Ascend’s report of 2006, see Air Safety Digest, February 2007, p. 2, in the archives section). According to Hayes:

“During 2007 there were 25 fatal accidents world-wide where passengers and/or crew were killed. This was two more [accidents] than in 2006 but still better than the average for the current decade (27) and markedly better than the experience of the 1990s, which saw an annual average of 37 accidents. …[See Figure A]

“There were 15 accidents during 2007 where a passenger was killed on a revenue passenger flight. This is two more than in 2006 but slightly better than the annual average for the decade (16) and very much better than the 1990s average of 24. …

“During the year, 631 passengers were killed in revenue passenger flights. [See Figure B] This is 159 less than in 2006, when 790 passengers died, and 89 less than the average for the decade (719). 2007’s performance and the average for the decade are both very much better that the 1990s average of 954.

Figure B

Passenger fatalities on revenue passenger flights (aircraft of 15 or more passenger seats; excludes acts of violence. Note that in only 5 of the 42 years shown have 500 or less passengers been killed in a year. Source: Ascend

“Aviation as measured by passenger fatalities is now very much safer than in the 1990s. The passenger fatality rate in 2007 was 0.24 per million carried, while the rate for the period since the start of 2000 is 0.32. This compares with a rate of 0.56 passengers killed per million carried during the 1990s. [See Figure C]

Figure C

Fatal accidents per million flights (all airline jets and turboprops). Note the tendency of accidents to peak about every seven years, and the indication from the results of 2007 that the trend may continue. Source: Ascend

“The worst accident in 2007 was the TAM Airbus A320 crash in July at Sao Paulo, Brazil, which killed all 181 passengers and 6 crew on board plus a further 12 people on the ground when it overran on landing and struck an office building. Apart from this accident, there were two others where more than 100 people were killed (Kenya Airways Boeing 737NG – 105 passengers and 9 crew killed – and Adam Air 737CFMI – 96 passengers and 6 crew killed). There were also two other accidents with high loss of life (One-Two-Go MD80 – 85 passengers and 5 crew killed – and Atlasjet MD80 – 50 passengers and 7 crew killed). These five accidents together resulted in 517 passenger fatalities, about 80% of the total passenger fatalities for the year.”

Despite the relatively good record in 2007, insurance premiums were not sufficient to pay the losses, prompting at least one insurer to stop writing airline insurance. Other insurance companies may pull out of the sector, also, Hayes surmised. The money-losing situation may be self-inflicted. Premiums in 2001 were three times the amount for 2000, reflecting the huge losses following the 9/11 terrorist attacks. But since that high-water mark, premium income has fallen to less than half that of 2001. (See Figure D)

Figure D

Airline hull and liability claims and premiums ($m). Note that aviation insurance was a losing business in 2007, as it has been twice before in the past two decades. Source: Ascend

Airline insurance premiums are “probably about $1 billion below the point insurers consider adequate for a healthy business,” Hayes surmised. The price of accidents is high. According to Hayes, hull claims for Western-built jet airliners in 2007 cost a total of $655 million, about $100 million more than in 2006 and the highest since the terrorist attacks of 2001.

He anticipates a gradual hardening of insurance rates to the point where it is profitable to underwrite airliners.

Western built jets, which carry more than 90% of the world’s traffic, experienced just seven fatal accidents in 2007, the same number as occurred in 2006, but with a slightly increased loss of life. Overall,

Western-built jets racked up a fatal accident rate of just 0.26 per million flights.

By comparison, the fatal accident rate for Eastern-built jets this decade is some 13 times higher.

The really significant problem is for Eastern-built turboprops flying in Africa. Hayes summed up the situation thusly:

“Eastern-built turboprops had a poor year with 17 known losses. There were just 9 total losses in 2006 and the annual average for the decade so far is 14. The average for the 1990s was 11. However, as in a number of previous years, many of the losses suffered in 2007 were in Africa, 10, with six of these being in the Congo (Zaire). There were 11 fatal accidents during the year resulting in 122 passenger and crew deaths. … Eastern-built turboprops also produced the most third party ground fatalities when an AN26 crashed shortly after take-off from Kinshasa, DRC, killing at least 28 people on the ground and injuring some 30 others.”

Ascend’s year-end report focuses on accidents, where damage to aircraft is at least “substantial” (cost of repairs exceed $1.0 million or 10% of the aircraft’s value) or where people are injured or killed.
Mark Goodrich, with Aviation Consulting Enterprises, Inc., of Reno, Nev., believes accidents present only part of the safety picture. Indeed, he believes using accident rates as an indicator of airline safety can be misleading. He maintains that a focus on incidents is key to accident prevention. Writing in a recent issue of the British magazine, Aircraft Technology Engineering & Maintenance, he argued for a more diligent examination of incidents:

“It is well understood that accidents usually follow a chain of incidents – that is, a series of individual and system errors, many relatively innocuous when viewed alone, combine and lead to accidents. Indeed, it is said that breaking the chain is crucial to accident prevention. What often remains unsaid is that breaking the chain amounts to ‘incident prevention.’

“Thus, a focus on incidents – rather than accidents – is the best way to foster a culture of safety in the aviation workplace. …

“How often have we read reports of accidents like the recent 737-800 fire in Japan, or the A320 runway overrun in Brazil, only to learn in the following months that a number of incidents bearing directly on the causes had gone either unreported or ignored – sometimes for years. It is often only an accident that suddenly brings the preceding history of incidents into the harsh light of reality. …

“It has been said that safety is not the destination but the journey itself. Incidents are the events we see in daily operation, and creating a culture of safety requires a focus on incident prevention. That only works if the reporting of incidents is encouraged – even rewarded – at every level. Managers must lead by example – one cannot hide evidence of incidents in the carpeted hallway, and expect each incident on the shop floor to be reported faithfully. Incidents not reported are also not available as an educational tool to create increased awareness, and the breaking of a future accident chain. …

“The truth behind accidents is that the incidents leading up to them are the real story.”

Goodrich’s company is involved in evaluating repair stations, conducting pilot training, test flying aircraft and investigating accidents. Based on that work, Goodrich estimates that incidents have increased in the United States a whopping 450% since 2000. The Federal Aviation Administration (FAA) and the National Transportation Safety Board (NTSB) maintain very incomplete incident databases. Thus, while the NTSB database on airline operations reflects a 39% decline in accidents, Goodrich estimates an explosive increase in the number of incidents:

“Our work reflects that incidents are up substantially from the line, to MRO (Maintenance, Repair & Overhaul) to flight operations. In no small measure, we’ve observed this to be related to: (a) the increased outsourcing of functions (both domestically and overseas), with reduced oversight and quality control; (b) the increased willingness of regulators to allow the extension of training intervals, and reduction in curriculum content (usually in response to airline requests for such action in order to reduce training budgets); and (c) increasing allowances for airlines and MROs to operate without diagnostic equipment, without training and without documentation for the very airplane, engines, equipment and functions as to which they hold certified repair station operations specifications credentials or approved maintenance program certifications.

“The recent NASA study [see elsewhere in this edition of Aviation Safety & Security Digest] – suppressed as it was by both the FAA and NTSB (‘not invented here’) – supports our observations that incidents substantially exceed all ‘official’ data. The existence of the NASA study and the attempts by ‘official’ Washington to discredit it lends credence to our quite unofficial observations as to the actual levels of incidents in service.

“The increasing trend in ‘turn backs’ and ‘in flight deviations’ – both resulting in unscheduled landings for Part 121 and Part 135 flight operations – likewise lend credence to our observations.”

It has been said that an incident is an accident that was lucky, and that the difference between an accident and an incident is a few feet or a few seconds. If this is the case, and if incidents are truly on the upswing, it is only a matter of time before the accident trends tracked by Ascend tick upwards.