**1. Introduction**

Unintentional injury is a leading cause of premature mortality in New Zealand children aged 1–14 years, accounting for two in five deaths in this age group [1–3]. Child injury mortality is inequitably distributed in New Zealand [3]. There are clear disparities for New Zealand's indigenous Maori children ¯ who experience a rate of fatal injury 3.5 times greater than that for non-Maori children [ ¯ 1]. New Zealand's child injury mortality rate is amongs<sup>t</sup> the poorest amongs<sup>t</sup> comparative Organisation for Economic Co-operation and Development (OECD) nations, being more than twice the rate of Sweden, the United Kingdom, Netherlands and Italy [4]. A 2009 report card scoring the adoption and implementation of evidence-based child injury prevention policies found New Zealand implemented only half of these interventional policies, again comparing poorly to comparative European OECD nations [5].

Children, although not traditionally thought of as part of the formal workforce in high income countries, do participate in work under less formal arrangements such as casual work during school holidays, part-time work after school, and work in family businesses. Additionally, they may be exposed to workplace hazards when they visit or live on worksites, such as farms. The only previous New Zealand

study of child work-related fatal injury (WRFI) examined incidents from 1985 through 1998 [6]. This study reported child WRFI commonly occurred when children were bystanders to another person's work process or activity, with the agriculture sector and farms in particular the dominant setting for these injuries [6]. Studies on child WRFI in other countries are also limited, although available evidence from Australia and the United States show a similar pattern as New Zealand, with the agricultural sector being the most common industry involved in child WRFI [7–9].

Child fatal injuries are therefore rarely studied in a work-related context, despite work contributing significantly to the burden of unintentional injury for this age group and such child-related deaths being included in workplace safety legislative protections. Child WRFI data are di fficult to obtain from o fficial workplace injury notifications or injury compensation claims, as these databases typically only capture those aged 15 years and older. Furthermore, injuries sustained in a work setting are not readily identifiable in external cause codes as defined by the International Classification of Diseases (ICD) framework [10] and while some could be identified using corresponding ICD place of occurrence and activity codes higher levels of use of "other specified" or "unspecified" categories mean these are less readily utilized in research [11]. In contrast, in New Zealand, coronial records provide a complete and comprehensive source of child WRFI because all deaths that are 'sudden and unexpected' are referred to Coroners to determine the cause and circumstances of death. This study, utilizing Coronial records, provides the most up-to-date and comprehensive information available on child WRFI in New Zealand.

The research aims to address the current deficit in knowledge about child WRFI nationally and internationally, capturing all child fatalities where a work exposure directly or indirectly contributed to the causes and circumstances of the fatal injury incident. All children who were fatally injured on a worksite, in a public place or on a public road as a result of employment (paid, unpaid or in-kind for family business) or due to another person's work were included. This research will provide directions for preventive actions by using coronial case file data to establish a complete and comprehensive cohort of all WRFI in children from 1999 to 2014 in New Zealand.

### **2. Materials and Methods**

Unintentional work-related fatal injuries of children aged less than 15 years were examined as part of a larger study of work-related injury fatalities in New Zealand [12,13]. In brief, potential WRFI cases with a date of death registration on or between 1 January 1999 and 31 December 2014 were identified using New Zealand's Mortality Collection, the most complete data source for all New Zealand deaths, including work-related injury fatalities. Injury deaths were identified in the Mortality Collection as those with an underlying cause of death coded to an external cause in the International Classification of Disease (ICD-10-AM) range V01 to X59, X85 to Y34, Y85 to 86, Y87.2, Y87.2 and Y89.9 [10]. Linkage to coronial records held by the National Coronial Information System was undertaken with all corresponding coronial records reviewed for work-relatedness. Coronial files were found for 98% of all 1335 relevant external cause deaths for children.

A broad definition of work-relatedness, compared with o fficial data definitions, was used to capture all child fatalities to which a workplace exposure contributed. The work-relatedness of a fatal injury event was decided based on whether the decedent, at the time of the fatal incident, was: working for pay, profit or paymen<sup>t</sup> in kind; assisting with work in an unpaid capacity; was engaged in other work-related activities even when on a break or away from the workplace; or was a bystander (as defined below) to another person's work activity. All fatal injury cases determined to be work-related were broadly classified as one of the following.

Worker deaths: the decedent was fatally injured in the course of work duties in a workplace (referred to as workplace WRFI), or on a public road (referred to as work-tra ffic WRFI).

Bystander deaths: the decedent was not working but died as a result of someone else's work activity regardless of fault (referred to as bystander WRFI). These deaths could be further classified as bystander deaths occurring on a public road (work-traffic bystander), at a work place (workplace bystander) or to students of primary school age or older where the incident occurred during school time or while they were performing a task directly connected with their course (students).

Rural deaths: the decedent was fatally injured on a rural workplace (farm) where the circumstances did not satisfy the worker or bystander definitions above. This group includes farm deaths in children where it was difficult to ascertain the relative contribution of work and non-work exposures.

Socio-demographic characteristics including age, sex and ethnicity were obtained from the Mortality Collection. Prioritised ethnicity was determined by categorising individuals with multiple ethnicity responses in the order of Maori first, then Pacific, Asian and finally European ¯ /Other to provide a single response as per Ministry of Health ethnicity protocols [14]. Small area geographical meshblocks were coded from the physical address where the injury incident occurred. Small area-level deprivation was then derived using the 2013 New Zealand Deprivation Index (NZDep), with deciles categorised into quintiles, with '1–2' representing those living in the least and '9–10' the most deprived areas [15]. Mechanism of injury, location of injury incident, and agency and industry of incident were obtained from coronial records. Standard coding frameworks including the Type of Occurrence Classification System (TOOCS) and the Australian New Zealand Standard Industry Classification (ANZSIC) were used [16,17].

To describe the burden and patterns of child WRFI, frequencies and percentages were calculated. The risk of child WRFI was calculated by age group, sex, ethnicity and deprivation using incidence rates per 100,000 person years with 95% confidence intervals (95% CI). Population estimates for children aged less than 15 from the 2000, 2006 and 2013 Census were obtained from Statistics NZ with denominators for non-census years estimated by linear interpolation and extrapolation. Data were analysed using Stata V13.1 SE [18].

To illustrate the most common circumstances of child WRFI, a series of narrative "profiles" were created using a combination of quantitative data (analysis described above) and qualitative analyses. Qualitative analyses used a thematic analytical approach examining what the decedent was doing prior to the injury incident, what went wrong to cause the injury and the cause of death.

Ethical approval for this study was granted by the University of Otago Human Ethics Committee (Ref 15/065), the National Coronial Information System (Ref NZ007), and Health and Disability Ethics Committee (Ref OTA/99/02/008/AM05).
