**4. Discussion**

Work-related fatal injury contributes substantially to the total burden of unintentional fatal injury in New Zealand; 15% of the total burden of fatal unintentional injury in children in New Zealand during the period 1999–2014 was attributable to a work exposure. Both the former Health and Safety in Employment Amendment Act 2002 and the new Health and Safety at Work Act 2015 require employers and the responsible Government Agencies to protect all people who come into contact with workplaces, including children [19,20]. As such, there is a clear legislative mandate to prevent WRFI in children and investigate the circumstances of any that do occur. This study's findings indicate the highest risks for child WRFI are in males, those under five years of age and those of Maori or European ethnicity. Child WRFI most ¯ commonly involved vehicle crashes or being hit by moving objects in incidents occurring on farms or public road. The main industries involved included the agriculture and transport sector.

This is a largely hidden burden as New Zealand's o fficial WRFI data excludes bystander deaths, thereby excluding most WRFI deaths in children. This study's novel data therefore lead to the identification of important missed opportunities to reduce the broader public health impact of work-related fatalities in New Zealand. Updated accurate data on the current patterns of work-related fatalities inform where workplace interventions to prevent fatalities to children are needed and allows for the monitoring in trends in child WRFI over time. Combined with previously published child WRFI data, this study has demonstrated a declining trend in child WRFI between periods. This study identified 1.5 (95% CI 1.3, 1.7) work-related fatal injuries per 100,000 children for the period 1999 to 2014, a considerable decrease compared to the rate of 2.1 (95% CI 1.8, 2.4) in the study that covered from 1985–1998 [6]. However, despite the decline, similar patterns of child WRFI to those described earlier remain: bystander deaths dominate and the agricultural industry is the most common industry in which child WRFI occurs [8]. While there was an overall decline in rates of WRFI in adult workers in New Zealand over a similar period, these trends

have been variable [21]. For example, adult workers in the primary production sector of agriculture, forestry and fisheries experienced an increase in fatal injury risk over this time [21].

The majority of children who sustained WRFI were bystanders, consistent with patterns of work involvement observed in children in Australia [7,9] and, as would be expected, given the low proportion of children engaged in formal or informal work, especially at a young age. The preponderance of bystander deaths in children reflects inadequate control over the hazards generated by work activities largely under the direct control of an adult worker, thus providing insight into the failures in the control of hazards in these work situations. This study found it was rare for children less than 15 years to be fatally injured as a worker. The circumstances of work involvement change with increasing age. As children ge<sup>t</sup> older and their physical and cognitive abilities develop, there is increasing movement into the formal workforce which changes the exposure of the child to workplace hazards, and the pattern of fatalities becomes more like that of adults [21].

The age-related patterns of WRFI observed in this study are consistent with those observed elsewhere. The exposure of children to work hazards, even at a very young age, are partially explained by the norms and attitudes of parents. For example, children residing on New Zealand farms often accompany working parents, or are engaged in farm work, where there are distinctive patterns of exposure to farm tasks by age [22]. Families residing on farms can perceive that children raised on farms are more aware of hazards and are more capable of handling risky agricultural tasks even at a young age [23]. North American Guidelines informing adults of agricultural tasks appropriate to the di fferent child physical and cognitive developmental stages and corresponding abilities of children have been shown to be e ffective at reducing the burden of injury in children on farms [24].

Males are commonly over-represented in child WRFI, particularly on farms [9]. Gendered patterns of exposure to work hazards occur on New Zealand farms, with male children more involved in farm tasks involving machinery and vehicles resulting in increased risk of WRFI for males [22]. Maori children are ¯ over-represented in unintentional fatal injury, accounting for half of all child unintentional injury deaths in New Zealand [25]. This study's findings sugges<sup>t</sup> injury prevention e fforts to reduce child WRFI for Maori ¯ need to focus on reducing work-tra ffic crashes. Whilst child WRFI rates varied by level of deprivation, with the highest rate of child WRFI observed for the most deprived children, there were few other clear patterns observed. This is in contrast to the common trend of a gradient of risk of fatal injury with increasing levels of deprivation in child unintentional injury mortality overall in New Zealand [26]. Few other studies have considered this as a potential determinant of child WRFI.

Rural environments are an important determinant of child WRFI. The agricultural sector is consistently identified as being one of the highest risk groups for fatal injury in adult workers in New Zealand [27–29]. Others have previously identified children living on farms or rurally as carrying higher injury risks compared to other non-farming or urban counterparts [7,9,30,31]. In agriculture, unlike other industries, children provide informal labor for family farm operations at an early age and farms typically have dual purposes, being both a place of work and of residence. This poses higher risk of work-related injury for children residing on, or visiting farms, because they are exposed to common and high-risk hazards, such as farm machinery and moving vehicles [22], to which other children are not exposed.

Very young children are particularly vulnerable to WRFI on farms where high levels of active supervision are needed to keep them safe in a high hazard work environment. While a 2007/08 survey of NZ farms found it was uncommon for young children to accompany working adults on farms [22], when it does occur, it poses a particularly high risk of WRFI for young children. A lack of childcare options for rural farming families, and rural attitudes of including children in informal family farm work from a young age, have also been identified as a barrier to protecting farm children from the hazards of work [32]. Many farming parents in New Zealand feel that, in order to prevent injuries on farms, a child's access to active farming areas, as well to farm machinery or vehicles, should be restricted, or at a minimum properly supervised by an adult [22]. Other opportunities for prevention include the use of the most suitable vehicle to transport children on farms and the use of child restraints, such as seat belts and car seats in work vehicles to ensure the safety of children as occupants in vehicles used on- or o ff-road [33].

Other work contexts were also identified as being of concern. This study identified that work exposures make a substantive contribution to 28% of the total burden of transport-related fatalities involving children aged under 15 in New Zealand. Using public roads is a necessary part of everyday life for children; therefore, it is unsurprising that motor vehicle tra ffic crashes (MVTC) are the overall leading cause of death in New Zealand children aged less than 15, accounting for over a quarter of all child deaths for the period 2006–2010 [34]. While our finding highlights the potential for reducing child fatalities on public roads through the influence of workplace safety policy, few of these incidents were triggered by a work vehicle, which limits the possibilities for prevention through work safety actions. One area of direct influence of work, however, that needs to improve is heavy vehicle driver awareness of cyclists and pedestrians. To help with these sensory technologies, alerting drivers of the presence of cyclists or pedestrians in blind spots should be adopted on working vehicles. The implementation of other evidence-based tra ffic safety interventions, such as proper child car restraints, use of seatbelts, alcohol controls, and improved road infrastructure, such as physically divided roads on routes carrying high volumes of heavy working vehicles, provide other opportunities to reduce the burden of child WRFI [35]. However, it is really only the last of these that is primarily a work-related prevention initiative. Fatalities occurring in industrial or construction settings that are more common amongs<sup>t</sup> working adults [21] were rare in children.

It is important that children's needs for preventive actions in the workplace are assessed alongside those of the working adult population because the patterns of child WRFI di ffer from those of adults. Simply applying injury interventions intended for adults may not adequately protect children. Children are considered vulnerable to WRFI due to their lack of power in the work organizational hierarchy, and their lack of physical and cognitive development, leading to children and adults underestimating risk [32]. Effective or promising strategies for preventing child injury broadly includes (listed from most to least effective) legislation, modification of the environment or a product, the use of safety devices, educational home visits, community based interventions and education and skills development [33,34].

New Zealand performs poorly in the prevention of child and adolescent fatal injury, with rates of child injury mortality ranking the worst out of out of 24 OECD countries [4]. A wide-ranging national strategy for child safety more broadly is currently lacking, with previous national strategic e fforts, most notably the New Zealand Injury Prevention Strategy (operational from 2003 to 2013), falling victim to changes in political focus [36]. The Well Child/Tamariki Ora programme, which takes an integrated child health and development approach to improve child health and well-being, includes child safety. However, it predominantly targets home and transport-based injury risks [37]. Children are a notable omission from New Zealand's Health and Safety at Work Strategy despite the strategy's vision of "work [that] is healthy and safe for everyone in New Zealand" [38]. This situation serves to illustrate the lack of implementation of important and e ffective policy and legislative actions to improve safety for children [5]. Our study highlights the need for child safety strategies to address the substantial role of work exposures in addressing the burden of fatal injury for children in New Zealand. National strategies for farm safety have been developed in Australia; for example, farm focused interventions and education programmes have been implemented to address child safety issues on farms [39].

Overall, these findings identify where improvements in prevention e fforts are needed to address child WRFI, serving to highlight the importance of managing work-related risks for children, particularly in the Agricultural sector. Most child WRFI are highly preventable, largely by adult intervention and enforcement of current workplace health and safety legislation. Furthermore, the recurrent common narrative scenarios point to the highly repetitive circumstances surrounding these incidents, implying that it is common to fail to learn from previous fatal incidents involving children. Interventions to address WRFI in the

younges<sup>t</sup> children (0–4 years) should focus on improved adult supervision of children during play on farms, while interventions for children aged 5–14 years should focus on reducing vehicle crashes on public roads, alongside farm settings. To identify and address the hidden burden of child WRFI o fficial data capturing work-related injuries should be expanded to capture fatal and non-fatal injuries sustained by children. Regular surveillance of the burden and patterns of child WRFI will directly inform interventions to change work practices harmful to children and allow for these incidents to be included in health and safety enforcement practice.

The availability of detail-rich Coronial records allowed for the comprehensive and accurate determination of work-relatedness and the collection of new information on the causes and circumstances of child WRFI unavailable from other data sources. The mandated requirement for all sudden and unexpected deaths to be notified to a Coroner for investigation means that virtually all child deaths due to external causes receive an inquest to determine the cause and circumstance of the fatal injury providing a comprehensive population basis for informing directions for intervention. The inclusion of common narratives capturing the recurring circumstances of injury is novel for studies describing the burden of child fatalities. This study is limited to 2014 as the most recent year available due to the lengthy time it takes for a Coronial inquest to be "closed" and become available for research purposes, limiting the currency of the data. It is a strength that this study expands the range of childhood work-related deaths to consider all industry groups with many previous studies focusing only on fatalities occurring on farms. This study is limited to fatal injury in children and the mortality rates for children fatally injured while working used the total child population due to the lack of working population denominators for children. The inclusion of tra ffic fatalities where the working vehicle was not "at fault" limits the generalisability to other countries with "at-fault" systems restricted to cases where there was liability on the part of a working vehicle. However, inclusion of cases regardless of fault in this study is consistent with New Zealand's no-fault universal accident compensation system.
