Efficacy of proxy definitions for identification of fatigue/sleep-related crashes: An Australian evaluation

Sleep-related crash characteristics: Implications for applying a fatigue definition to crash reports

“…In both the police-reported and proxy analysis 14 colliding with a stationary object was significantly more prevalent in SR than not-SR crashes and had 15 one of the largest effect sizes. Collision with something in clear sight has previously been reported as 16 a feature of SR crashes (Horne and Reyner 1995a) and is one of the most common outcomes of self-17 reported SR incidents (Armstrong et al 2013). These findings provide some evidence for potential 18 benefit of the inclusion of collision with a stationary object in proxy definitions, a feature which has 19 also been reported by others (Crummy et al 2008).…”

“…While this proportion is closer to that obtained from 18 high-quality case control investigation (Connor et al 2002) it is still an estimate of SR crashes . 19 Approximately half of the police-reported SR crashes did not meet the proxy definition, suggesting 20 that proxy definitions are only identifying a subset of SR crashes (Armstrong et al 2013). The 21 accuracy of proxy definitions is particularly uncertain as they do not closely reflect drivers subjective 22 reports of SR crashes (Armstrong et al 2013) and have been criticised for being too narrow (Crummy 23 et al 2008).…”

“…single vehicle run 9 off the road crashes are most likely to be SR (Horne and Reyner 1995b, Pack et al 2006)). However, 10 there has been some criticism that proxy definitions are too narrow (Crummy et al 2008) and that 11 they only capture a specific subset of SR crashes (Armstrong et al 2013). 12…”

Sleep-related vehicle crashes on low speed roads

“…There may be several reasons for 1 this, first, police officers lack an accurate objective measure of sleepiness and self-reported difficulty 2 to identify sleepiness as a crash causal factor (Radun et al 2013). Second, in a previous study we 3 have found that of drivers who self-report having had a SR incident in the previous 5 years, only 45% 4 report that the police were involved (Armstrong et al 2013). 5…”

“…Philip Hallvig et al 2013). However, self-reported experience by drivers of SR 8 driving incidents suggest that approximately 25% of SR incidents occur on roads with a speed limit of 9 ≤ 50km/h (31mph), with a further 30% occurring on roads with speed limits between 50 and 80km/h 10 (31-50mph) (Armstrong et al 2013). To date, little attempt has been made to specifically investigate 11 these low speed SR crashes therefore it is unknown how these differ from either not-SR crashes of a 12 similar speed or high speed SR crashes.…”