Accident Investigation

James Reason [1, 2] developed a model of understanding aviation and other organizational accidents by showing layers within an organization that all contribute to the cause of the accident. Reason’s model became known as the Swiss Cheese Model and showed that “holes” in each layer create the conditions for an accident and that those at the end of the chain, for instance the pilots, are rarely solely responsible. Organizational factors, environmental conditions, and team actions all play a part. The latent conditions within the organization turn into active failures when they are triggered by a certain event, such as deteriorating weather in aviation, that in combination may lead to an accident.

Reason distinguishes slips and lapses that are unintended actions as well as mistakes and violations that are intended actions. Although this part of Reason’s work is based on concepts developed by Rasmussen [3, 4] and other parts were influenced by the information processing model by Norman [5], his approach resulted in a widespread improvement of understanding organizational accidents. Today’s safety investigation analysis methods for aviation are largely based on his ideas.

The Human Factors Analysis and Classification System (HFACS) by Shappell and Wiegmann [6, 7] became a popular tool to investigate accidents, classify the underlying factors, and distinguish unintended and intended actions that may have caused the accident. It was to become particularly successful for aviation [6, 7]. A number of other models followed, for instance Threat and Error Management [8], that were more schematic but specifically designed for aviation. Criticism on such approaches by Dekker [9] and others found that these approaches were too limited and studied the process only after the accident had already happened.

The approach of HFACS became dominant but was used to investigate accidents rather than practices that could possibly lead to an accident. Accidents lead to serious injury, fatal injury, or a substantially damaged or destroyed aircraft; reporting aircraft accidents is mandatory and aviation organizations, such as the National Transportation Safety Board in the USA, collect, analyze, and publish statistics on accidents. Incidents cause only minor damage or injury and, in case of the USA, are mostly collected by the Federal Aviation Administration. All such reports are historical and record the problems only after they have resulted in damage and injury.

Efforts are made to collect reports on problems that only potentially lead to an incident or accident. The aviation organizations have endeavored to collect voluntary reports on such problems. In addition the International Civil Aviation Organization (ICAO) supports a program known as LOSA [8], a line operational safety audit, in which, for instance, a person joins the crew in the cockpit and monitors their actions. This approach records practices rather than accidents and is quick to identify areas in which procedures, training, or policies can be improved.

Part of the criticism on accident investigation analysis was the focus on relatively rare events rather than current practices [overview in 10]. Although, this is often true for commercial aviation, general aviation shows a different picture.

Military and Civil aviation have traditionally been studied separately although some of the flight characteristics resemble one another, particularly that of helicopters. Within Civil or Civilian aviation there is a distinction between commercial and non-commercial flights. The first concerns the major airlines and Reason’s work has transformed the investigation practices of airline accidents. Airlines and commercial flights, such as air taxis, transport goods or people to generate revenue. General aviation, on the other hand, commonly uses smaller aircraft and features a wide range of operations. They include fire-fighting, logging, crop-dusting, emergency medical services (EMS), and other operations often specific to helicopters as well as a range of flights with balloons, gliders, ultralights, gyroplanes as well as small planes and helicopters that are used for personal or business use, for flight training or just for recreation.

In 2007 [11], there were 228,000 active private pilots and 220,000 registered General Aviation aircraft in the USA. In the years prior, 91% of all aviation crashes and 94% of all aviation fatalities were in GA. The crash rate was reported as 1.31 fatal crashes per 100,000 flight hours, which is 82 times the rate for major airlines. This striking difference between general aviation and major airline operations has been in place for many years and the accident investigation models were not designed to address this difference.

HFACS has been applied to general aviation and has classified the types of mistakes leading to accidents. The crucial element of the investigation models is the layers beyond the pilot’s mistake and in GA these layers are often not accessible or not investigated. But since accidents are not rare in general aviation, there is still much to gain from accident analysis without the danger of focusing on rare and unusual events only.

Apart from relatively high numbers of accidents, the general aviation sector is characterized by highly diversified operations and aircraft. Next to overviews of general aviation accidents, each type of aircraft and each specific operation has required and has been given attention. The characteristics of these smaller groups of accidents create a better insight and more specific recommendations to improve the safety record of this industry.

Epidemiological Studies

The main source of the literature for epidemiological studies in aviation is the journal of the Aerospace Medical Association: Aviation, Space, and Environmental Medicine, which is a continuation of its two predecessors Aerospace Medicine and the even earlier Journal of Aviation Medicine. Together they published almost 75 years of aviation-related epidemiological studies including countries other than the USA although rarely outside the UK, Australia, and New Zealand [12]. An important part of the early epidemiological studies concern the age and occupation of the pilots, topics that are less commonly studied today [13]. They did result in the 60-year rule that is still used as an upper limit for airline pilots but is highly debated [Overview in 14]. A review of this journal from 1994 [15] states that “in some cases, the deficiencies in study design and data analysis have resulted in controversial findings.” Li mentions the age-60 rule as a particular example and adds: “Many of the contributions were made by uncontrolled studies based on crash analysis in the form of case reports and case series studies.” Case–control and cohort studies are still rare, mostly because the case reports are easily available and still need much study while control studies require more effort.

One case–control study of risk factors for fatal and nonfatal injury in New Zealand’s civil aviation [16] showed that environmental and operational factors were key determinants of the injury outcome rather than pilot or aircraft characteristics. This confirms the shift in the literature from pilot characteristics to environmental and operational factors. The type of aircraft is still mentioned since O’Hare et al. [16] conclude that flying a twin-engine aircraft was a risk factor for fatal injury, while piloting a microlight aircraft was a risk factor for all serious injury. The most significant risk factors for fatal and serious injury were aerobatic flight, post-crash fire, not having a certificate of airworthiness and off-airport location.

Publications of the last 10 years add much insight to the particularities of certain operations and aircraft. They identify the relevant environmental factors that are reported by the accident investigators. As such, the shift of research attention that was initiated by Reason has resulted in the inclusion of the environment and the operational factors but rarely the organizational factors when it comes to GA. GA is not part of one large organization but is governed by national laws, international aviation regulations, many different company policies as well as individual preferences, which in the case of, for instance, flight instruction [17] creates a complex picture for which improvements are possible but for which policies are not always related to the high number of accidents.

Types of Accidents

Apart from aircraft types and, sometimes closely linked, aircraft operations, a number of studies have focused on a type of accident. For instance, studies on the geographical location of the aviation accident, the type of accident according to the gender of the pilot, accidents relating to disorientation including wrong airport landings and also mid-air collisions.

Mid-air collisions are not limited to a type of aircraft nor to a particular operation but are considered the most dangerous type of accident. Apart from statistics, studies on mid-air collisions have focused on human factors, in particular the inability of pilots to detect converging aircraft in time to prevent a collision [18, 19]. Morris [19] analyzed mid-air collisions in GA to illustrate the limitations of the see-and-avoid principle and a subsequent study [20] showed that radio communication also has little effect to prevent such mishaps.

Extensive literature exists on spatial disorientation in both clinical and aviation studies [21]. Spatial disorientation is associated with aerobatic maneuvers, flying into clouds and vertigo in which pilots are no longer able to understand the position of the aircraft in relation to the ground. In aviation, way-finding problems have been termed geographical disorientation [22] and are often seen as part of a situation awareness problem. In such cases, the position of the aircraft in relation to a (mental) map is no longer understood. Landings at the wrong airport are a curious result of geographical disorientation [22, 23]. The development of improved displays and the introduction of GPS navigation systems in GA can be seen as the main solutions for the prevention of such accidents in the future.

Aviation in terms of geography is partly made insightful by studies from different countries. In the USA Grabowski et al. [24] also identified localized problems that point at climate, regulatory and flight activity differences within one country. In the USA, the state of Alaska has for that reason received increasing attention from researchers and aviation authorities because of its relatively high accident rate.

Studies by Baker et al. [25] on gender show pilot characteristics that still play a role in GA accidents since female pilots make different types of errors compared to male pilots. Although pilot characteristics are no longer central in aviation accident studies, they are not to be ignored altogether.

Aircraft-Specific Studies

Specific operations have received attention well before Reason and his followers. In GA it has been recognized that crop-spraying, fire-fighting, and logging operations by helicopters require specific attention and have little to gain from helicopter air taxi accident statistics even though air taxi accidents are more numerous. Although helicopters seem to have the most diverse types of operation, small airplanes are involved in crop-spraying and fire-fighting as well, allowing comparisons of the two machines. This is less common for the other types of aircraft that make up only a minor part of GA flights but which require separate recommendations (Table 8.1).

Of these other aircraft, hot-air balloons have received most of the attention. Historically, the first aviation accident occurred with a hot-air balloon and the medical profession has benefitted from the studies on altitude sickness conducted almost exclusively in these aircraft up until the space age. Next to hot-air balloons there are less frequent gas balloon accidents that are sometimes included in these studies as well as zeppelins that are usually ignored. The number of zeppelin-related accidents at any given time has been too small even in countries such as Australia, the UK, and the USA to allow any statistical analysis.

Fatal accidents in hot-air balloon crashes are few compared to those for small airplanes or helicopters but still warrant investigation [26, 27]. A study from the UK [28] highlighted victims with severe burns as one of the dangers in ballooning. De Voogt and van Doorn [29] showed that the severity of the injuries and the damage to the balloon were not significantly related. Crashes occur in which the balloon is destroyed but the passengers remain unharmed. The envelope of the balloon may catch fire when it strikes an electrical wire in which case the passengers can often leave the basket unharmed. In contrast, hard landings cause most of the injuries in ballooning but the basket and the envelope commonly remain in one piece. When the basket drags across the ground injuries may become more severe or even fatal.

Fatal glider accidents share characteristics with homebuilt airplanes. Van Doorn and de Voogt [30] found that fatal accidents were predicted by pilot error, flight phase, and homebuilt aircraft. Gliders and ultralights [31] are often homebuilt so that maintenance and construction errors are more common. Homebuilt aircraft make up 3% of all hours flown in GA but 10% of all crashes and also have a higher fatality rate [32].

Research on gyroplanes has shown that pilot experience is a significant predictor of accident fatality [33]. Pilots with less than 40 flight hours were four times as likely to be involved in a fatal accident as their more experienced colleagues. This result is similar to that found for ultralights in an earlier study [31]. Gyroplanes, also known as autogiros, were regarded as a relatively safe and stable type of GA aircraft. They were developed in the 1920s as a safe alternative to airplanes but were replaced by helicopters in the 1940s. Helicopters do not have a reputation as safe aircraft, partly because helicopter operations can be considered more dangerous.

Helicopters

A survey of US helicopter accidents between 1994 and 2005 showed that accident rates were between 11.6 per 100,000 h in 1994 and 6.3 in 2005 [30]. A rare study of helicopter accidents in Russia [34] states that “the Western helicopter relative statistical data conform to the Russian helicopters ones” in particular the accident rate that is higher than for airplanes. For the period 1982 up to 2006, 15.2% of all US helicopter accidents were fatal [30]. Adverse weather and night conditions showed a significantly higher proportion of fatal accidents. At night the accidents were 3.0 times and in instrument-flying conditions 6.3 times more likely to be fatal. These results are similar to those found in specific helicopter operations. For instance, fatalities after helicopter EMS crashes are associated especially with post-crash fire and with crashes that occur in darkness or bad weather, according to Baker et al. [35]. Helicopter EMS operations from 1997 to 2001 showed a lower crash rate than General Aviation but fatal crashes showed a rate higher than all other categories of aviation: 1.7 per 100,000 h compared to 1.3 for general aviation.

Table 8.1 Accident severity and aircraft specificity for 2,141 accidents occurring between 1982 and 2007 in the USA, extracted from the ntsb online database
Full size table

Taneja and Wiegmann [36] reviewed 84 autopsies of helicopter pilots from 1993 to 1999 in the USA. Blunt trauma was cited as the primary cause of death in 88.1% of these cases with ribs (73.8%), skull (51.2%), and facial bones (47.6%) as the most common bony injuries. There was no relation of the injury pattern with the age of the pilot or the phase of flight. They argue for crashworthy aircraft in order to reduce the number of fatalities.

Hayden et al. [37] recommend crash-resistant fuel systems for Civil helicopter operations, which have shown close to 100% effectiveness in Army helicopters and which already prove to reduce post-crash fires in helicopter crashes. A study that includes both GA helicopters and airplanes in the USA [38] also related fire or explosion to pilot death and found that pilots who failed to use both lap belt and shoulder harness were more likely to die as well as those who used only the lap belt compared to both restraints.

Helicopters are uniquely equipped for specialized operations such as forestry and fire-fighting, often categorized as external load operations, as well as operations at low altitude in confined areas, such as crop-spraying also known as aerial application. The environment in which these operations take place allow little room for error and the machine’s performance is limited by the specialized loads.

In low-altitude operations, helicopters have an increased risk of colliding with trees, transmission wires, power lines, or other high structures. One of few published studies on external-load helicopter operations, also known as sling-load operations, was presented by Manwaring et al. [39] who described the causes and circumstances of accidents between 1980 and 1995. They linked their outcomes to studies of Alaskan wood-logging helicopter operations in which progress on safety had been made. Research on aerial application flights has concentrated on the danger of the applied chemicals rather than the danger of the flight operation but a combined study of sling-load and aerial application accidents [40] shows that safety can be improved with additional crew and better fuel management. Additional crew may assuage the number of external load and aerial application accidents since the number of tree and wire strikes make up 18 and 24% of all such accidents, respectively.

Pathology

The differences between aircraft is also reflected in pathology studies. Wolf and Harding [41] in their study of sport aircraft-related deaths included nine deaths in ultralight and experimental aircraft. The patterns of injuries included trauma predominantly to the chest, abdomen, or head in addition to blunt force injuries in chest and abdomen or head and torso. Only two extremity factures were found while injuries to the symphysis pubis appeared in six cases. They conclude that “these cases illustrate the varied pathologies associated with deaths due to crashes of sports aircraft and reveal the lack of uniformity associated with the investigations of such deaths.”

A rare electrocution is reported for ballooning where collisions with powerlines are a frequent occurrence in accidents [42]. Deaths in ballooning are usually a result of blunt trauma from falls. Although burn victims are reported, burn injuries in balloon and small airplane crashes are usually survivable if the patient arrives at a burn center alive [43].

In gliders and light airplanes decelerative injury is a common pathological feature. A study by Byard and Tsokos [44] of an extreme case highlights the avulsion of the distal tibial shaft. The death of the two occupants of the light airplane was caused by severe injuries involving craniocerebral, skeletal, soft tissue, and organ trauma. The legs were shortened and fragments of distal tibial shaft had been forced through the soles of the occupants’ shoes, which indicated a fall from considerable height and a direction along the axis of the legs. They conclude that the observation of extrusion of bone fragments downwards through the shoes and the fracturing of the lower limb bones can add to the understanding of the crash, particularly the position of the occupants prior to impact.

Wiegmann and Taneja [45] report on the basis of 559 autopsy reports in the USA that in GA blunt trauma is the primary cause of death (86%) with fractured ribs, skull, facial bones, tibia, and pelvis as the most common. They also mention a fractured larynx in 14.7% of the cases; a common injury not reported before. Common organ injuries include laceration of the liver (48.1%), lung (37.6%), heart (35.6%), and lung (32.9%). They add that individuals who sustained brain hemorrhage were also more likely to have fractures of the facial bones more so than skull fractures.

A study of fatal light aircraft accidents in Canada [46] showed that out of 68 victims that were studies about half of them suffered multiple trauma, 29% drowned, 16% died of head or neck injuries and 3% due to a coronary disease. Neck trauma occurred mostly with pilots and also occurred most often with drowned occupants. Passengers suffered relatively more craniofacial fractures and abdominal or retroperitoneal trauma. Ethanol intoxication was implicated in two cases but other drugs did not appear to be a definite factor in accident causation.

Although drugs and alcohol are sometimes suspected, particularly with general aviation, Kuhlman et al. [47] already found that there is no consistent pattern of drug involvement in civil aviation fatalities. Not counting nicotine and ethanol they found 12.6% of 377 US cases positive for one or more drugs, with acetaminophen and salicylate as the most frequent ones. Ethanol at greater than 10 mg/dL was found in 14.8% of the cases but only 4.5% were concluded to be due to ingestion of ethanol.

Accidents and Fatality

The least and the most dangerous operations in General Aviation, arguably flight instruction and aerobatics maneuvers, show the importance of detailing the study of GA accidents.

Flight instruction is not limited to any type of aircraft and received special attention in the literature. Studies reporting flight accidents by student pilots [e.g., 48, 49] indicate that this is a substantial subcategory of accidents. Olson and Austin [50] found that the highest rate of errors in the landing phase was found during the flare and follow through phases of the landing. Benbassat and Abramson [51] state that pilots consider the flare to be more difficult than a number of other flight maneuvers. A follow-up study by Uitdewilligen and de Voogt [52] showed that solo flights and first-time solo flights in the USA show an even higher proportion of improper flares. As with most maneuvers taking place close to the ground, the number of fatalities is small and in the case of first-time solo flights zero.

In contrast to flight instruction, aerobatic flight maneuvers are the most significant risk factor for fatal and serious injury, at least in the USA [16]. Nearly 50% of airport transport pilot-induced accidents in GA occur during aerobatics [53].

Instructional flights are the first flights of every airline pilot and aerobatics is mostly limited to the most experienced. Both show that it is difficult to separate General Aviation from Commercial Aviation since at least the pilots travel across these boundaries.

The strategies to prevent fatalities in General Aviation differ widely and although shoulder restraints and crash-resistant fuel tanks are general recommendations, the operations and aircraft are too diverse to address them all at once. The relatively high accident and fatality rate for General Aviation flights emphasize the importance of both new approaches to safety and continuing epidemiological research in aviation.