Accidental pediatric death: a ten year retrospective cohort

Abstract

Background

Accidental death is the leading manner of pediatric death worldwide. In Canada, this phenomenon has been explored in-depth in select provinces however published assessments of accidental pediatric death in the province of Newfoundland & Labrador (NL) remain limited.

Methods

A retrospective cohort study of cases of accidental death in children and youth aged 0- to 21- years that occurred between January 1, 2004 and December 31, 2013 and referred to the Office of the Chief Medical Examiner (OCME) of NL was carried out. Comparison was made between accidental death in NL and the provinces of New Brunswick (NB) and Manitoba (MB) matching ages and years of occurrence.

Results

A total of 139 cases occurred with a mean age of 15.8 (+/-4.9) years. Males accounted for more cases than females (p < 0.001). Leading cause of death was mechanical trauma due to a motor vehicle collision (MVC) (52%). Among all causes of death statistical significance was present (p < 0.001). Within NL regional differences were present with the highest incidence of all accidental death occurring in the Labrador Grenfell Health region (p = 0.002). When comparing NL to MB, rates of accidents were comparable. More differences were seen when comparing to NB regarding specific causes of death, and sexes.

Conclusion

Accident incidence differed significantly between NL provincial regions with rural areas manifesting more accidents per capita than urban areas. NL data was similar to MB and lower than NB in overall accidental death, cause specific death and among sexes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12887-025-06197-0.

Background

Among global literature, accidents have been cited as the leading manner of death among the pediatric population [1–4]. Common causes of such death in a pediatric population are known to include drowning [1–3, 5, 6], motor vehicle collisions (MVC) [1, 3, 5, 6], fire [2, 3, 5, 6], smoke inhalation [2, 3, 6], asphyxia [2, 3], substance misuse [1, 5], and other blunt injuries (i.e. falls) [2, 3]. While variation exists in the literature, MVC are often reported as the most common cause of accidental pediatric death, followed by drowning [1–3, 5, 7, 8].

More pediatric males die due to accidents than pediatric females [1–4, 9, 10]. In Canada male deaths represented 64.4 and 71.4% of accidental deaths in British Columbia (BC) and Quebec (QC) [3, 9]. The frequency of accidental death in a pediatric population varies based on age and cause. Children under one year of age have the lowest rate of accidental death, likely because their deaths would be attributable to the actions another person. Otherwise phrased, children under one are less likely to become a victim of an accident due to their own devices [10].

Causes of unexpected accidental pediatric death in the Canadian province of Newfoundland & Labrador (NL) have not been studied. The primary objective of this study is to describe causes of accidental death in those from birth to 21 years of age referred to the NL Office of the Chief Medical Examiner (OCME) between January 1, 2004 to December 31, 2013.

Methods

This is a retrospective cohort study analyzing cases of accidental pediatric death that occurred in the Canadian province of NL between January 1 st, 2004 and December 31 st, 2013. Cases were included if they were determined to be an accidental manner of death, occurred in those aged 0 to 21 years of age, and were referred to the NL OCME. The OCME of NL reports, investigates, and records reportable deaths in the province. Deaths considered reportable include death if a person has died under suspicious, unexpected, or non-natural circumstances [11]. This guideline ensures that accidental deaths specifically are captured and investigated by the OCME, therefore ensuring full ascertainment of death due to an accidental manner of death through the OCME in NL. These dates were selected due to data availability. Ethics approval was granted by the Health Ethics Research Board of Newfoundland and Labrador (HREB #2012.199).

In North America, the upper age definition for a pediatric patient is variable and ranges from 18y to 21y [12–14]. This study adopted the definition of a pediatric patient described by the American Academy of Pediatrics as birth to 21 years [14].

Database description

The dataset used contained fatalities, deemed by the OCME, to be due to an accidental manner of death in those aged 0- to 21-years. All data used in this study contained no personal information about those affected, therefore data was completely anonymized. In the province of NL, the OCME oversees all death investigation in the province. Regarding accidental death specifically, as dictated by the Fatalities Investigation Act (2014), the OCME is to be notified of all death occurring in the community or facility of any kind that is believed to be the results of an accident [11]. This means that full ascertainment of accidental manners of death would be reflected in the database utilized in this study reducing selection biases in this study. Variables collected in the dataset included age, sex, manner of death, medical cause of death, date of death, medications before death, non-prescription or recreational drugs, circumstances/location of death, activity level at symptom onset, and underlying pathology.

The database describes location based on the regional health authority (RHA) where the deceased lived, not necessarily the region where the death occurred. RHAs included Eastern Health (EH), Central Health (CH), Western Health (WH), and Labrador-Grenfell Health (LGH) (Fig. 1). Since 2023 health regions were amalgamated to form one health care region [15].

Fig. 1.

Newfoundland & Labrador regional health authorities. From (Statistics Canada, 2017)

Literature for comparison

The NL dataset was compared to similar cases of accidental death that occurred in Manitoba (MB) and New Brunswick (NB) in publications by Herath et al. (2014) and The Office of the Chief Coroner of NB, respectively. Inclusion and exclusion criteria of the original literature source comparing provinces was reviewed. The selected referenced databases included death time period between 2004 and 2013 and included pediatric cases.

• This MB study assessed autopsied cases of non-natural death in those 0- to 18-years in MB from 1989 to 2010 [6]. Cases of mechanical trauma due to MVC, mechanical trauma not due to MVC, drowning, smoke inhalation and drug toxicity were used as comparison between MB and NL.

• The data from NB originated from The Office of the Chief Coroner and included accidental manner of death from 2006 to 2014 in those 0 to 19 years [16–23]. Overall cases of accidental death, drowning, fire-related death, substance-related death, accidental asphyxia, and motor vehicle accidents were compared between NB and NL.

Statistical analysis

Statistical analyses were completed using SPSS version 27 [24]. Chi-square tests were used to compare frequencies and incidence rates. Fisher’s exact test was used if the sample size was small, or greater than 20% of cells had a value less than 5. A p-value of < 0.05 was considered statistically significant.

Results

Between 2004 and 2013, 139 cases of accidental death in those 0 to 21 years, in NL were referred to the OCME between the ages of 0 and 21 years. Males accounted for 74.8% (104/139) of cases and females 25.2% (35/139) (Table 1). No cases of accidental death occurred in children under 1 year. Cases had a notable rise at 15 years and saw the highest number of cases at age 19 years. Accidental death in those 15 to 21 years of age accounted for the majority of reviewed cases. In total accidental death in those aged 15 to 21 years represented 74.8% of all accidental death in the study population.

Table 1.

Descriptive statistics of all accidental death in Newfoundland and Labrador

	Cases	Incidence per 100,000	Significance (p-value)
Age
Mean (SD)	15.8 (+/- 4.9) years
Median	17.0 years
Sex
Male	104	16.4	p < 0.001 *
Female	35	0.6
Age Category
0–4 years	7	2.93	p < 0.001 *
5–9 years	11	4.30
10–14 years	17	5.03
15–19 years	75	23.17
20–21 years	29	21.66
Cause of Death
Mechanical Trauma due to MVC	72	5.8	p < 0.001 *
Drowning	27	2.2
Smoke Inhalation/Fire	14	1.1
Drug Toxicity	10	0.8
Asphyxia	7	0.6
Mechanical Trauma not due to MVC	6	0.5
Hypothermia	3	0.2
Region
Eastern Health	64	8.7	p = 0.002 *
Central Health	20	9.4
Western Health	27	14.8
Labrador Grenfell	21	20.0
Unknown/Other	7	N/A
Season
Winter	25	2.0	p = 0.011 *
Spring	33	2.7
Summer	51	4.1
Fall	30	2.4

*Statistical significance, p value < 0.05

Of all cases, 34.5% (48/139) were found to have a non-prescribed substance present. These substances were counted if detected regardless of level present. Of such cases, 66.7%(32/48) had one class of substance present while 33.3% (16/48) had multiple classes present. Substances detected included alcohol, opioids, cocaine & derivatives, tetrahydrocannabinol (THC) & derivatives, benzodiazepines, amphetamines, antihistamines, psychedelics, and inhalants. Alcohol was the most common substance, in 75% (36/48) of cases. Regarding MVCs, 31.9% (23/71) had non-prescription substances detected. Of these MVC toxicology positive cases, 82.6% (19/23) had one classification of substance present and 17.4% (4/23) had multiple classifications of substances present. As with the larger toxicology positive cases, in MVCs alcohol was the most common substance detected.

Cases of accidental manner of death were categorized and included; mechanical trauma due to an MVC, drowning, smoke inhalation/fire, drug toxicity, asphyxia, mechanical trauma not due to MVC, and hypothermia (Table 2). Statistical significance is present when comparing incidences of all causes of death, suggesting some causes occur at a rate significantly different than others.

Table 2.

Descriptive statistics by cause of death in Newfoundland and Labrador, 2004-2013

	MVC		Drowning		Smoke/Fire		Drug Toxicity		Asphyxia		Non-MVC		Hypothermia
		p-value		p-value		p-value		p-value		p-value		p-value		p-value
Age
Mean (SD)	16.4 (+/- 4.2) years		16.3 (+/- 5.3) years		11.9(+/- 6.9) years		17.9(+/- 2.1) years		17.1 (+/- 3.8) years		10.3 (+/- 5.4) years		17.0 (+/−2.6) years
Median	17.0 years		19.0 years		14.0 years		17.5 years		19 years		10.5 years		18.0 years
Sex
Male	48 (7.6)	p = 0.009 *	25 (3.9)	p < 0.001 *	7 (1.1)	p = 0.925	9 (1.4)	p = 0.022 ⍭*	7 (1.1)	◆	5 (0.8)	p = 0.220 ⍭	3(0.5)	◆
Female	24 (4.0)		2 (0.3)		7 (1.2)		1 (0.2)		0 (0.0)		1 (0.2)		0(.)
Age Category
0–4 years	2 (0.8)	p < 0.001 *	2 (0.8)	p < 0.001 *	2 (0.8)	◆	0 (0.0)	◆	0 (0.0)	◆	1 (0.4)	◆	0(0.0)	◆
5–9 years	5 (2.0)		1 (0.4)		4 (1.6)		0 (0.0)		0 (0.0)		1 (0.4)		0(0.0)
10–14 years	6 (1.8)		3(0.9)		1(0.3)		1 (0.3)		2 (0.6)		3 (0.9)		1(0.3)
15–19 years	43 (13.3)		15 (4.6)		5 (1.5)		6 (1.9)		3 (0.9)		1 (0.3)		2(0.6)
20–21 years	16 (11.9)		6 (4.5)		2 (1.5)		3 (2.2)		2 (1.5)		0 (0.0)		0(0.0)
RHA
EH	37 (5.0)	p = 0.091	11 (1.5)	◆	4 (0.5)	◆	7 (1.0)	◆	3 (0.4)	◆	2 (0.3)	◆	0(0.0)	◆
CH	14 (6.6)		6 (2.8)		0 (0.0)		0 (0.0)		0 (0.0)		0(0.0)		0(0.0)
WH	17 (9.3)		5 (2.7)		3 (1.6)		1 (0.5)		1 (0.5)		1(0.5)		0(0.0)
LGH	3 (2.9)		4 (3.8)		7 (6.7)		1(1.0)		2 (1.9)		1(1.0)		3(2.9)
Unknown/Other	1 (.)		1 (.)		0 (.)		1 (.)		1 (.)		2(.)		0(.)

Categorical Variables presented as n(incidence rate per 100,000)

◆ Chi square unable to be calculated due to sample size

* Statistical significance, p value <0.05

⍭ Fishers Exact Test

Mechanical trauma due to an MVC was the leading cause of death in NL accounting for 51.7% (72/139) of cases. MVCs in this study included collisions involving cars, cars and pedestrians, boats, snowmobiles, ATVs, and cyclists. Victims were noted to be passengers, drivers, and pedestrians. When considering body region injured, head injuries were the most common (Fig. 2). While MVC trauma affected those between 2- and 21- years of age, the majority of cases occurred in those between 15- and 19- years. Though not reported in the MVC group, further investigation showed that MVCs were contributory factors in deaths due to asphyxia, drowning and smoke inhalation/fire. When considering all trauma cases due to an MVC and cases where MVC was a contributory factor 59.7% (83/139) of cases involved an MVC.

Fig. 2.

Body regions injured in mechanical trauma due to motor vehicle collisions, 2004–2013

Male deaths were more common than female death in all cases, except smoke inhalation/fire related fatalities. Seven males, and seven females succumbed to smoke inhalation/fire related fatalities and sex differences were statistically insignificant (p = 0.925).

Deaths due to hypothermia were rare and restricted to the LGH region between January and March.

Drowning, asphyxia and drug toxicity related deaths tended to affect males in their pre-teen, teenage and young adult years. Drownings in those 0- and 18- years was infrequent with such deaths affecting those 19 years of age and older predominantly: cases in those 19 + years represent 55.55% (15/27) of all drownings. Cases of asphyxia did not affect anyone under 11 years. Drug toxicity fatalities affected those in mid-teenage years and above. Substances detected included ethanol, opioids, cocaine and derivatives, benzodiazepines, antihistamines, and inhalants. Only one had ethanol as the lone substance. It was more common for multiple substances to be detected where two (4/10), three (4/10) and four (1/10) classes were found. When multiple substances were present, 88.9% (8/9) involved an opioid in combination with another drug.

Comparison to MB

Comparison reflected cases in those 0- to 18-years between 2004 and 2010. No statistical significance was found between the incident rate of all accidental deaths, and smoke inhalation/fire-related fatalities in the NL and MB populations. NL had a statistically significant incidence rate higher than that of MB in mechanical trauma due to MVC (p < 0.001).

Comparison to NB

Comparison between NB and NL reflected cases in those 0 to 19 years between 2006 and 2013. NL had a statistically significant higher incidence rate per 100,000 than NB in smoke inhalation/fire-related causes of death (p = 0.042).NB had a higher incidence rate than NL in overall accidental death (p = 0.018), and mechanical trauma due to an MVC(p < 0.001). Among all accidental death no statistical significance was noted among the incidence of death in males between NL and NB, however NB had significantly more female cases than NL (p = 0.008).

Discussion

Sex differences

Differences in the occurrence of accidental death between males and females was present in the study population. In NL, males accounted for 75% of pediatric accidental fatalities, outnumbering females among all causes of death except for cases due to smoke inhalation/fire related causes. In accidental or injury-based fatalities literature, the difference in male and female death does not typically appear until after 4 years of age. From the age of 5 and up males outnumber females in injury- related fatalities. This gap becomes wider between the sexes until rates similar to those observed in the adult population occur between the ages of 15 and 17 years [25]. Existing literature has not fully explored causation for the male and female differences in accidental fatalities, but provides insight considering male and female behaviors. Overall, males are known to partake in risky behavior, largely related to operating motor vehicles which often begins in the teenage years. Their behaviors are described as reckless or risky and their view on consequences “fatalistic” [26]. Though not a definitive explanation this context may provide a hypothesis for why sex differences were observed in this study.

Males and females were affected by smoke inhalation/fire in equal proportions in NL. Reports of males and females succumbing to accidental death due to smoke inhalation or fire-related causes at equal proportions or with observed differences that are not statistically significant were present in multiple literature sources in Canada and the United Kingdom [16–23, 27, 28]. Contradictory publications are present and discuss findings of smoke or fire- related deaths where males are victims more so than females. It is important to note that these articles do not reflect cases that occurred during the time period of this study, and may be considered dated sources due to changing safety regulations and norms [29]. Rational for the absence of sex difference in smoke inhalation/fire related causes have been proposed by multiple authors and include the tendency for fires, in accidental cases, to be residential and occur predominately at night or early morning when residents are sleeping putting children of both sexes at equal risk [28, 30].

Regional differences

RHA differences are present in incidence of deaths in NL. NL has 40% of residents living in rural areas as defined by the Government of NL [31].Rural regions are known to have more pediatric injuries and deaths than urban areas [9, 32, 33]. Multiple factors may contribute to the rural-urban divide, which can include longer time before an injured person is discovered, less advanced on-scene care, longer transport times to hospital, lack of specific pediatric trauma care in the region, and environmental exposures which can include frequent travel on or over water and overall increased commuting times [9, 33–35].

An assessment of ethnic differences could not occur as that was not a variable in the original dataset.

Interestingly, there were no recorded cases in any RHA of deaths by firearm, sharp force injuries or electrocution.

MVC

Mechanical trauma due to an MVC was the leading cause of accidental death in the pediatric population in NL. This is consistent with the literature nationally and internationally. In NL, the mean age of death due to an MVC was 16.4 (+/- 4.2) years. The driving age is 16 years old in NL [36]. New drivers are known to be at higher risk for accidents and associated fatal motor vehicle events [37]. This study was not able to comment on the position of the deceased in a vehicle at the time of an accident. We might, assume that teens are often accompanied by other teens in the vehicle, be they passengers or drivers, in multiple situations.

MVCs play a role in cases where causes of death were attributed to something other than mechanical trauma. In NL, MVCs were known factors in 59.7% (83/139) of all cases. This is a conservative estimate as cases only recognized as an MVC were referenced.

Recreational and road vehicles have been associated with cases of drowning and asphyxia post-MVC in the published literature and in the presented study [38, 39].

In NL, in accidents of in all vehicle types, cases of asphyxia preceded by an MVC were not always clear. Specific causes of death due to asphyxia after an MVC were attributed to traumatic asphyxia, or asphyxia secondary to an MVC. Using stated causes of death it can be assumed that occupants were crushed or trapped under the vehicle causing restricted respiratory movements due to pressure on the chest and abdomen, leading to death [40, 41]. Circumstances of post-MVC drowning were provided in some cases and included recreational vehicles going through thin ice, or road vehicles landing in water.

Toxicology

Toxicology results were an important consideration in accidental pediatric death. While substances directly lead to death in drug toxicity cases, other causes of death were potentially influenced by the effect of substances, in the deceased. Among all cases of accidental death in NL, positive toxicology results were present in 34.5% (48/139) of cases.

Canadian literature discussing toxicology in accidental pediatric deaths is limited. Alcohol is known to be the most common substance used in a pediatric population followed by cannabis [42, 43]. Psychedelics, opioids, amphetamines, benzodiazepines, and cocaine are known to be used, but at varying and lower frequencies [44–46].

Alcohol and drug use is commonly discussed in relation to MVCs. The NL dataset does not reliably describe victims’ position in the vehicle in MVC-related fatalities. Therefore, only general issues regarding toxicology in a population where an MVC was a factor in the event may be addressed.

Alcohol was the most common substance in the context of MVCs. In 2017, after the study period, new laws prohibiting drivers under the age of 22 years from having a blood alcohol concentration over zero were brought into effect [36, 47]. This law affected the study population post the study time frame. The results from this study would thus be useful to detect if the change in law had any effect on toxicology results in MVCs in future assessments of pediatric accidental death.

Limitations

As this study utilized a dataset that was previously created with all available information on all cases from the OCME, it is likely that data collection is close to full ascertainment, in that all referred cases are included in this dataset. However, specific information regarding in depth circumstances for each death referred to the OCME is not available in all cases. Details included were not always extensive and at times lacked information about circumstances of the death or medical information of the participant. While in-depth details were reported in some cases, they were sporadically available. This limited discussion of the circumstances of various accidents. Impact was well illustrated in the limitation of analysis of MVCs as participants location in the motor vehicle (i.e. driver, passenger, etc.) was not regularly available and limited discussion about this topic. Having limited information about a persons position in a vehicle further limited discussion of toxicology results in this context. Further, information about ethnicity was not available in this database.

Literature for comparison excluded those 19-to 21-years. This is likely because in Canada the upper range for the pediatric population is 18 years. When data was available about the population over 19 years of age it was combined and reported in aggregate with ages outside the scope of this study leaving comparison of 19 to 21 -year-olds not fully discussed in the NL population.

Conclusion

Causes of death that occurred in NL were comparable to other regions in Canada. When considering incidences of NL compared to other provinces, accidental pediatric death occurred at similar incidences between NL and MB. It was noted that NB had a higher incidence than MB and NL. The incidence of accidents differed according to region.

This study is the first comprehensive description of OCME pediatric death in NL. In this study, differences in the incidence of accidents were observed among RHAs. These differences suggested a rural-urban divide. This information is significant as of April 2023 RHAs are no longer separate due to health system restructuring. The absence of RHAs can potentially impact future research into regional differences.

Understanding the epidemiology will be important in the care of children, teenagers, and young adults in NL. The identification of trends and occurrences of such death will help inform future public health policy and health care planning.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

NL

Newfoundland & Labrador

OCME

Office of the Chief Medical Examiner

NB

New Brunswick

MB

Manitoba

MVC

Motor Vehicle Collision

BC

British Columbia

QC

Quebec

RHA

Regional Health Authority

EH

Eastern Health

CH

Central Health

WH

Western Health

LGH

Labrador-Grenfell Health

THC

Tetrahydrocannabinol

Emily Jane Devereaux

The first author, EJD, completed this research study for fulfilment of her MSc thesis while attending Memorial University of Newfoundland. This study was completed in in 2023, and the graduate degree awarded in the same year. 

At present EJD is a PhD candidate at Dalhousie University.

Authors’ contributions

EJD contributed to the design, analysis, interpretation of data and both wrote and edited the manuscript. KH provided supervision, manuscript editing and dataset acquisition. RP provided supervision and contributed to the design of the study. ND contributed to interpretation of data and data acquisition. HM contributed to data interpretation. All authors read and approved the final manuscript.

Funding

This study was completed as a thesis project for the lead author (EJD). Funding for this project in fulfillment of a MSc degree was provided by a Janeway Foundation Research Grant.

Data availability

The dataset analyzed during the current study is not publicly available due to health ethics research board restrictions and privacy concerns. Researchers interested in accessing data or materials used in this study can contact Dr. Kathy Hodgkinson, [khodgkin@mun.ca] to discuss data sharing.

Declarations

Ethics approval and consent to participate

Ethics approval was granted by the Health Ethics Research Board of Newfoundland and Labrador (HREB #2012.199). The Health Ethics Research Board of Newfoundland and Labrador is guided by principles of the TCPS-2 which governs ethical conduct for research involving humans. This guiding document states that consent shall be given voluntarily and withdrawn at any time by the participant or designate, however alterations to consent can be granted under specified circumstances such as; If research involves no more than minimal risk to the participants, alternation to consent requirements in unlikely to adversely affect the welfare of participants, and it is impossible or impracticable to carry out the research and to address the research question properly if the prior consent of participants is required [48]. The presented study was approved by the health ethics research board under a non-consent model due to the above stated circumstances and applicable to those under, and over the age of 16 years.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.