detection tools proposed, tend to rely heavily on predictive fatigue models and subjective crew assessments on fatigue symptoms and sleep. Objective data sources such as physiological measures and SMS data, inclusive of roster quality indices and Flight Data Monitoring, are not considered. The model offers no guidance on the detection, investigation, risk assessment, and management process for fatigue events within the Safety Management System (the model defaults to management through an airline Safety Management System (SMS). The model also assumes there is an investigation process that is integrated into the SMS that can detect fatigue influence, causal and/or contributory, to event occurrence. Current fatigue risk investigative capability within literature only extends to the application of directed scientific methodology and application of predictive fatigue models (models that predict sleep opportunity and alertness that don’t consider individual differences and
S. Stewart, F. Koornneef, R. Akselsson, and C. Turner  - HILAS 2009
work environment). Neither process has the sufficient diagnostic capability to support the investigation of fatigue-related incidents. The question is what level of fatigue is contributory and linked to unacceptable performance relative to operational standards. Dawson & McCulloch acknowledge the lack of field performance studies from which to draw performance data to better define operational fatigue controls. The model proposes proactive and reactive controls for fatigue, but suffers from the limitations of the Reason trajectory model of causation. Crew may not manifest fatigue-related symptoms and/or behaviours, yet commit errors (easyJet HFMP study presented to NASA Ames -Intelligent Systems Division Seminar, 2007). Further, fatigue symptoms, behaviours, and

1.12.3 Safety Risk Management and fatigue-related risk
The relational models of Sutton et al (2003), ASLEF (2003) and Baker & Fergusson (2004) describe the relationship between fatigue causes and consequences. Dawson and McCulloch (2005) evolve this concept by describing ‘fatigue as a hazard’ linked to Reason’s accident causation model (Reason, 1997). The Dawson & McCulloch model describes a fatigue ‘error’ trajectory and outlines controls appropriate to the four levels of
S. Stewart, F. Koornneef, R. Akselsson, and C. Turner  - HILAS 2009

the model (sleep opportunity to fatigue related incident/accident). The level one hazard is stated as ‘sleep opportunity’ as the starting point of the trajectory however sleep opportunity in itself does not constitute a hazard, lack of sleep opportunity however does. However, Sutton et al group the antecedents to fatigue (hazard sources) as work factors, shift system and individual factors. Caldwell (2004) cites the principle antecedents as schedule factors, sleep deprivation, circadian factors and individual differences. The model further states that the first three levels of the trajectory are then addressed within the SMS as ‘error analysis’ representing a control mechanism. This would imply that some sort of risk assessment and management process is enabled from this analysis. The following quote from the Australian New Zealand Risk Standard 4360 refers to the process as:
　
中國航空網 www.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：Developing a Safety Management System for Fatigue Related Risks in easyJet(14)