Factors associated with pilot fatalities in work-related aircraft crashes — Alaska, 1990-1999

Morbidity and Mortality Weekly Report, April 26, 2002

Despite its large geographic area, Alaska has only 12,200 miles of public roads, and 90% of the state's communities are not connected to a highway system (1). Commuter and air-taxi flights are essential for transportation of passengers and delivery of goods, services, and mail to outlying communities (Figure 1). Because of the substantial progress in decreasing fatalities in the fishing and logging industries (2), aviation crashes are the leading cause of occupational death in Alaska. During 1990-1999, aircraft crashes in Alaska caused 107 deaths among workers classified as civilian pilots. This is equivalent to 410 fatalities per 100,000 pilots each year, approximately five times the death rate for all U.S. pilots (3) and approximately 100 times the death rate for all U.S. workers (4). As part of a collaborative aviation safety initiative that CDC's National Institute for Occupational Safety and Health (NIOSH) is implementing with the Federal Aviation Administration (FAA), the National Transportation Safety Board (NTSB), and the National Weather Service, CDC analyzed data from NTSB crash (*) reports to determine factors associated with pilot fatalities in work-related aviation crashes in Alaska. This report summarizes the result of this analysis, which found that the following factors were associated with pilot fatalities: crashes involving a post-crash fire, flights in darkness or weather conditions requiring instrument use, crashes occurring away from an airport, and crashes in which the pilot was not using a shoulder restraint. Additional pilot training, improved fuel systems that are less likely to ignite in crashes, and company policies that discourage flying in poor weather conditions might help decrease pilot fatalities. More detailed analyses of crash data, collaborations with aircraft operators to improve safety, and evaluation of new technologies are needed.

Aircraft crash reports are compiled by NTSB and entered into a database maintained by FAA's National Aviation Safety Data Analysis Center. Crashes in which pilots in command died were compared retrospectively with those in which they survived. All variables, except age, were dichotomized. Wald Chi-squared analyses were then completed. Factors that were evaluated included age, shoulder-restraint use, weather conditions (used as a marker for poor visibility), light conditions (light or dark), aircraft type (plane or helicopter), occurrence of post-crash fire, location (on or off airport), flight experience (median: 4,350 hours, range: 78-20,000 hours), and whether the pilot was an Alaska resident (a surrogate for familiarity with geography and flight conditions in Alaska). The Statistical Analysis System (SAS) software was used to generate odds ratios.

The study identified 675 work-related crashes; in 567 (84%), the pilot survived, and in 108 (16%), the pilot died. The estimated likelihood of pilot death was 14 times higher when a fire occurred than when one did not, seven times higher for flights that crashed in instrument meteorological conditions than for crashes in conditions of greater visibility, and approximately two times higher for crashes that occurred away from an airport or in darkness; the estimated likelihood of a pilot dying was significantly lower when the pilot used a shoulder restraint (Table 1).

Reported by: Conway G, Moran K, Alaska Field Station, Div of Safety Research, National Institute for Occupational Safety and Health; Bensyl D, FIS Officer, CDC.

Editorial Note: The results of this study indicate that crashes involving a post-crash fire, flights in darkness or weather conditions requiring instrument use, and crashes occurring away from an airport were significantly more likely to result in a pilot fatality. Conversely, crashes in which the pilot was using a shoulder restraint were less likely to result in a pilot fatality. These findings appear consistent with other studies identifying conditions associated with pilot fatality (5,6)

These findings suggest several possible approaches to reducing pilot death rates in Alaska. Companies should direct pilots to return to base if they encounter weather requiring instrument use and to avoid flying if they are likely to encounter such weather. Additional training in procedures to follow if weather conditions requiring instrument use are encountered unexpectedly should be provided. Use of improved fuel systems that are less likely to ignite following a crash could improve post-crash survivability.

Many aircraft manufactured before July 1978 are not equipped with shoulder harnesses (2). Although installing shoulder harnesses in small aircraft manufactured before July 1978 is voluntary by the owner/operator, doing so is often relatively simple and inexpensive (depending on the amount of structural reinforcement needed for each aircraft). FAA requires shoulder harnesses to be worn only for takeoff and landing, but not during flight (8). In this study, some crashes were catastrophic events for which no restraint system would provide protective effects; in other crashes, the pilot might have been incapacitated temporarily, preventing escape before fire consumed the aircraft. For crashes in which the initial impact is survivable, using a fastened shoulder harness might decrease temporary incapacitation from crash-related injuries. Recommendations to pilots and FAA that shoulder harnesses be used throughout a flight might reduce fatalities.