Sociodemographic Factors and the Risk of Pediatric Out‐of‐Hospital Cardiac Arrest in Ontario, Canada: A Province‐Wide Case–Control Study

Abstract

Background

Pediatric out‐of‐hospital cardiac arrest (POHCA) is associated with significant mortality and poor neurological outcomes. We aimed to describe the association between sociodemographic factors and POHCA risk in Ontario, Canada.

Methods and Results

We conducted a province‐wide case–control study at ICES, where patient records are linked across administrative databases. The case group included children (aged 1 day to 17 years) who experienced an out‐of‐hospital cardiac arrest between 2004 and 2020. Controls were matched up to 1:4 on age, sex, index date, and key comorbidities. We used conditional logistic regression to measure the association between sociodemographic indicators and POHCA risk. The case and control groups included 1826 and 7254 children, respectively. Children living in areas with the highest levels of material deprivation (adjusted odds ratio [aOR], 2.35 [95% CI, 1.94–2.85]) and dependency (aOR, 1.22 [95% CI, 1.01–1.48]) had a higher odds of POHCA, relative to children living in regions with the lowest levels of material deprivation and dependency, respectively. Children living in neighborhoods with the lowest levels of ethnic diversity had a higher odds of POHCA (aOR, 1.62 [95% CI, 1.30–2.01]), relative to children living in neighborhoods with the highest levels of ethnic diversity. The odds of POHCA were lower in immigrants (aOR, 0.67 [95% CI, 0.47–0.95]), relative to the general population. Northern urban residence was associated with a higher odds of POHCA (aOR, 1.45 [95% CI, 1.13–1.87]), relative to southern urban residence.

Conclusions

Children living in neighborhoods with high levels of marginalization may have an elevated risk of experiencing POHCA. These findings highlight the importance of addressing disparities through targeted prevention and intervention efforts.

Nonstandard Abbreviations and Acronyms

LHIN

Local Health Integration Network

OHCA

out‐of‐hospital cardiac arrest

POHCA

pediatric out‐of‐hospital cardiac arrest

Clinical Perspective.

What Is New?

• In our province‐wide case–control study, we found that children living in neighborhoods with high levels of marginalization had an elevated risk of experiencing pediatric out‐of‐hospital cardiac arrest.

What Are the Clinical Implications?

• These findings highlight the importance of addressing disparities through targeted prevention and intervention efforts, including increased education and awareness around pediatric out‐of‐hospital cardiac arrest and its risk factors, and improved access to high‐quality cardiopulmonary resuscitation training and primary and urgent care services.

Pediatric out‐of‐hospital cardiac arrest (POHCA) is associated with significant mortality and poor neurological outcomes. An estimated 2% to 10% of patients with POHCA survive to hospital discharge, and only 1% to 5% are discharged with good neurological status. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ The incidence of POHCA varies regionally, with rates ranging from 2 to 18 cases per 100 000 person‐years. ¹ , ³ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ Recent evidence suggests that there may be sociodemographic disparities in POHCA survival and in the provision of upstream interventions, including bystander cardiopulmonary resuscitation, automated external defibrillator application, and emergency medical services response times. ²² , ²³ , ²⁴ There is limited literature on the source of these disparities, and it is unknown whether the populations impacted also have a higher likelihood of experiencing POHCA. This gap in the literature was further highlighted in a recent systematic review where there were no studies reporting on the association between socioeconomic status (SES), migrant status, and POHCA risk. ²⁵

The impact of sociodemographic factors on the risk of out‐of‐hospital cardiac arrest (OHCA) in adult populations has been well documented ²⁶ , ²⁷ ; however, these associations may not hold in the pediatric population due to differences in etiology and incidence. Importantly, most cases of POHCA are caused by respiratory failure, whereas adult OHCA is often secondary to myocardial infarction and of cardiac pathogenesis. ²⁸ As respiratory failure is largely preventable, the factors that elevate OHCA risk in children likely differ from those in the adult population. The identification of these factors could inform targeted prevention and intervention strategies, including increased awareness and education around POHCA and its risk factors, with a particular focus on recognizing the early signs of respiratory failure and ensuring emergency medical services providers are adequately trained in management techniques (eg, airway access and ventilation). We may also identify a need for increased access to health care, social supports, and high‐quality cardiopulmonary resuscitation and automated external defibrillator training.

This project aimed to mobilize information from population‐based health administrative data sets to describe the relationship between sociodemographic factors and POHCA risk in Ontario, Canada. We focused our analyses on the impact of neighborhood‐level marginalization, individual‐level immigration status, and geographic context.

Methods

We followed the Reporting of Studies Conducted Using Observational Routinely Collected Data guidelines, which are an extension of the Strengthening the Reporting of Observational Studies in Epidemiology guidelines (Table S1). ²⁹

Study Design and Setting

We conducted a province‐wide matched case–control study to describe the association between sociodemographic factors and the odds of experiencing POHCA in Ontario, Canada. The data sets used in this study were linked using unique encoded identifiers (ie, ICES Key Numbers) and analyzed at ICES (formerly known as the Institute for Clinical Evaluative Sciences). ICES is an independent, not‐for‐profit organization and data repository that collates information collected through the routine administration of publicly funded health services in Ontario, Canada. Institutional review board approval was not required for this study, as ICES is an entity with prescribed designations under section 45 of Ontario's Personal Health Information Protection Act, allowing data collection and analysis without individual consent or additional research ethics approval. The data set from this study is held securely in coded form at ICES. While legal data sharing agreements between ICES and data providers (eg, healthcare organizations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet pre‐specified criteria for confidential access, available at www.ices.on.ca/DAS (email: das@ices.on.ca).

We used the National Ambulatory Care Reporting System to identify cases of POHCA, and the Discharge Abstract Database to obtain information on select comorbidities for the purpose of matching. The National Ambulatory Care Reporting System houses data on hospital‐ and community‐based ambulatory care, including emergency department visits, day surgeries, and outpatient clinic visits. The Discharge Abstract Database is a national database providing clinical summaries of patient hospitalizations and discharge notes. Reporting to these databases is mandated in the province of Ontario.

We used the following databases to gather information on the sociodemographic characteristics of the study sample: the Registered Persons Database; the Ontario Marginalization Index; the Immigration, Refugees, and Citizenship Canada Permanent Residents Database; Local Health Integration Network (LHIN) data; and the Postal Code Conversion File.

Study Population

Cases were identified using a validated algorithm described by Gray et al. ³⁰ We included children (aged 1 day to 17 years) who experienced an OHCA and were transported to an emergency department in Ontario, Canada, between April 1, 2004, and March 31, 2020. Case ascertainment was based on International Classification of Diseases, Tenth Revision (ICD‐10) codes. The first documented POHCA event was denoted as the index event, which was subsequently used to define the index date. Children were excluded if they met any of the following criteria: (1) missing or invalid ICES Key Number; (2) missing or invalid age (<1 day or ≥18 years) at index date; (3) missing data on sex; (4) death before index event; (5) non‐Ontario resident; and (6) trauma code associated with index OHCA event.

The control group was identified through a 1‐to‐many matching strategy, with up to 4 controls selected for each case. The controls were matched on birth year, birth month (+/− 2 months) for the infant subgroup, sex, index date (+/− 2 months), and the presence of cardiac arrhythmias or congenital cardiac malformations, with a lookback window of 2 years from the index date. These comorbidities were selected due to their high prevalence in the case group. ²¹ We were unable to match for all potentially related comorbidities, due to the difficulty in finding suitable controls for conditions that are less common in the general population. The initial pool of potential controls included all children who were alive during the case accrual window. The pool was narrowed using the following exclusion criteria: (1) missing or invalid sex; (2) membership in case group; (3) invalid age (<1 day or ≥18 years) on index date; and (4) non‐Ontario resident on index date. Following the application of the first 2 exclusion criteria, individuals in the control pool were assigned pseudo index dates to match the distribution of index dates in the case group. After the application of the fourth exclusion criterion, individuals who had not been matched to a case were excluded.

Sociodemographic Variables

We examined the following social determinants of health: neighborhood‐level material deprivation, neighborhood‐level dependency, neighborhood‐level residential instability, neighborhood‐level ethnic diversity, individual‐level immigration status, and geographic context.

Neighborhood Marginalization

The Ontario Marginalization Index is a validated indicator that provides information on a wide range of social, demographic, and economic characteristics associated with an individual's area of residence. ³¹ The index uses census tract–level data to compare the level of inequality between neighborhoods, assigning quintile estimates across 4 dimensions of marginalization: material deprivation, dependency, residential instability, and ethnic diversity. The least marginalized regions (eg, areas with the lowest levels of material deprivation, dependency, etc) are represented by quintile 1, and the most marginalized areas (eg, areas with the highest levels of material deprivation, dependency, etc) are represented by quintile 5. The index values are composite measures derived from data on different constructs of marginalization. Material deprivation incorporates information on variables such as income, education, proportion of lone‐parent families, and housing quality. Dependency integrates data on the proportion of residents who are unemployed or in unpaid professions, and the proportion of individuals aged ≥65 years or <15 years. Residential instability describes various family and housing characteristics including the proportion of the population living alone, the average number of people per household, and the proportion of residents who relocated in the past 5 years, among other factors. Ethnic diversity includes the proportion of residents who are recent immigrants or identify as being part of an ethnic minority group.

Immigration Status

The Immigration, Refugees, and Citizenship Canada Permanent Residents Database provides information on all landed immigrants to the province of Ontario since 1985. This data set was used to ascertain immigration status for members of the study population. Immigration status was classified at the level of the child. Children who were born outside Canada and migrated thereafter were classified as immigrants. If the child was not an immigrant, then they were categorized as being a member of the general population.

Geographic Context

Children residing in areas with a community size of ≤10 000 people were categorized as rural residents, and children residing in communities with >10 000 people were categorized as urban residents.

LHINs were not‐for‐profit corporations designed to facilitate the coordination of in‐hospital, community‐based, and in‐home health care services in the province of Ontario, before the year 2021. The LHIN database at ICES was used to classify members of the study population as southern or northern Ontario residents. Children were categorized as northern Ontario residents if the postal code associated with their home address was matched to 1 of the 2 northern LHINs (North East and North West). Children were categorized as southern Ontario residents if their postal code was matched to 1 of the remaining 12 LHINs.

The LHIN data were combined with information on rurality to facilitate the creation of the geographic context variable, describing members of the study population as residents of a Rural‐North, Rural‐South, Urban‐North, or Urban‐South region.

Missing Data

We provided a breakdown of missing data in both the case and control group for descriptive purposes. We did not include children with missing data in the regression analysis.

Statistical Analysis

We used descriptive statistics to summarize the demographic characteristics of the case and control group. Categorical variables were summarized using frequencies and proportions, while continuous variables were summarized using means and SDs or medians and interquartile ranges. We compared the distribution of matched factors and sociodemographic characteristics between the case and control group, where a P value of <0.05 was indicative of a statistically significant difference. We used the chi‐square test to assess differences in the distribution of categorical variables between both groups and the Wilcoxon rank‐sum test to assess differences in continuous variables. We also compared the mean age between both groups using a 2‐sample t test.

We used conditional logistic regression to measure the association between sociodemographic factors and the odds of experiencing POHCA. We did not adjust for most of the matched factors to avoid redundancy in the adjusted model. Although age was matched as a categorical variable, we included continuous age in adjusted models to increase precision and reduce residual confounding. We did not adjust for other matched factors, as they were binary variables and balanced by matching. We assessed multicollinearity by computing variance inflation factors for variables in the adjusted model, with a variance inflation factor of ≥10 indicating multicollinearity in the analysis. ³² We also conducted 5 likelihood ratio tests to determine if categorical variables with >2 categories (ie, all neighborhood‐level marginalization variables and the geographic context variable) were significantly associated with the risk of experiencing POHCA. The purpose of this analysis was to determine whether the variable as a whole was associated with POHCA risk, rather than the specific categories within it. This was done by comparing the fully adjusted model to models where 1 of the categorical variables had been removed.

Findings are reported as unadjusted and adjusted odds ratios (ORs) alongside 95% CIs. The adjusted analysis includes all 4 neighborhood‐level marginalization variables, the immigration status variable, the geographic context variable, and age as a continuous variable. Associations were considered statistically significant if the 95% CI excluded the null value of 1. The OR can be used to approximate the risk ratio due to the rarity of the outcome in the general population. ¹ All analyses were completed using SAS Enterprise Guide 7.1 (SAS Institute, Cary, NC). ³³

Results

A total of 1839 children experienced an OHCA in Ontario between April 1, 2004, and March 31, 2020. Among these, 13 were not matched to a control and subsequently excluded from our analysis. The creation of the control group involved narrowing a pool of potential controls through sequential application of the exclusion criteria (Figure). In total, the case and control groups included 1826 and 7254 children, respectively.

Figure . Flowchart outlining the derivation of the control group.

The distribution of matched factors was relatively balanced between both groups (Table 1). The median age was 2 years (interquartile range, 0–12) in both groups, and there was a larger proportion of infants, relative to the other age categories. There were significant differences in the distribution of each sociodemographic variable between the case and control group (P values <0.0001) (Table 2).

Table 1.

Distribution of Matched Factors in Case and Control Group

Matched factor	Cases	Controls	P value
	(N=1826)	(N=7254)
Age at index, y
Mean (SD)	5.51 (6.4)	5.53 (6.4)	0.883
Median (IQR)	2 (0–12)	2 (0–12)	0.848
Age category, n (%)
<1 y	712 (39.0)	2807 (38.7)	0.973
1–11 y	610 (33.4)	2433 (33.5)
12–17 y	504 (27.6)	2014 (27.8)
Female, n (%)	710 (38.9)	2820 (38.9)	0.995
Arrythmia, n (%)	74 (4.1)	257 (3.5)	0.299
Congenital cardiac malformation, n (%)	133 (7.3)	494 (6.8)	0.476
Index year, n (%)
2004/2005 to 2009/2010	651 (35.7)	2554 (35.2)	0.931
2010/2011 to 2014/2015	525 (28.8)	2090 (28.8)
2015/2016 to 2019/2020	650 (35.6)	2610 (36.0)

IQR indicates interquartile range.

Table 2.

Distribution of Sociodemographic Factors in Case and Control Group

Sociodemographic factor	Cases	Controls	P value
	(N=1826)	(N=7254)
Material deprivation quintile, n (%)
1	268 (14.7)	1492 (20.6)	<0.0001
2	272 (14.9)	1340 (18.5)
3	286 (15.7)	1373 (18.9)
4	345 (18.9)	1293 (17.8)
5	601 (32.9)	1656 (22.8)
Unknown	54 (3.0)	100 (1.4)
Dependency quintile, n (%)
1	477 (26.1)	2365 (32.6)	<0.0001
2	385 (21.1)	1541 (21.2)
3	298 (16.3)	1275 (17.6)
4	301 (16.5)	1043 (14.4)
5	311 (17.0)	930 (12.8)
Unknown	54 (3.0)	100 (1.4)
Residential instability quintile, n (%)
1	322 (17.6)	1763 (24.3)	<0.0001
2	307 (16.8)	1343 (18.5)
3	312 (17.1)	1236 (17.0)
4	401 (22.0)	1268 (17.5)
5	430 (23.5)	1544 (21.3)
Unknown	54 (3.0)	100 (1.4)
Ethnic diversity quintile, n (%)
1	295 (16.2)	882 (12.2)	<0.0001
2	297 (16.3)	1013 (14.0)
3	310 (17.0)	1165 (16.1)
4	352 (19.3)	1510 (20.8)
5	518 (28.4)	2584 (35.6)
Unknown	54 (3.0)	100 (1.4)
Immigration status, n (%)
Immigrant	41 (2.2)	271 (3.7)	<0.0001
General population	1785 (97.8)	6983 (96.3)
Geographic context, n (%)
Rural‐North	47 (2.6)	156 (2.2)	<0.0001
Rural‐South	192 (10.5)	550 (7.6)
Urban‐North	113 (6.2)	245 (3.4)
Urban‐South	1459 (79.9)	6288 (86.7)
Unknown	15 (0.8)	15 (0.2)

In the regression analysis, we excluded 154 children (1.7% of the study population) who had missing Ontario Marginalization Index data (N=154) or missing geographic context data (N=30). There was no indication of multicollinearity between variables in the adjusted model (variance inflation factors <10). The odds of experiencing POHCA varied across different levels of neighborhood‐level marginalization (Table 3). Children living in regions with the highest levels of material deprivation had a higher odds of POHCA (adjusted odds ratio [aOR], 2.35 [95% CI, 1.94–2.85]), as compared with children living in areas with the lowest levels of material deprivation. Children living in regions with the highest levels of dependency also had a higher odds of POHCA (aOR, 1.22 [95% CI, 1.01–1.48]), as compared with children living in neighborhoods with the lowest levels of dependency. The odds of POHCA were higher for children living in regions with the highest levels of residential instability (OR, 1.52 [95% CI, 1.30–1.79]), relative to children living in areas with the lowest levels of residential instability. However, this association was no longer significant in adjusted analyses (aOR, 0.97 [95% CI, 0.80–1.16]). Children living in areas with the lowest levels of ethnic diversity had a higher odds of POHCA (aOR, 1.62 [95% CI, 1.30–2.01]), relative to children living in neighborhoods with the highest levels of ethnic diversity. The odds of POHCA were lower in immigrants (aOR, 0.67 [95% CI, 0.47–0.95]), relative to the general population.

Table 3.

Univariable and Multivariable Conditional Logistic Regression Models With Sociodemographic Factors (N=8926)^*

Sociodemographic indicator	Crude OR (95% CI)	Adjusted OR† (95% CI)
Material deprivation quintile
1	[Reference]	[Reference]
2	1.14 (0.95–1.38)	1.13 (0.94–1.37)
3	1.17 (0.97–1.40)	1.18 (0.97–1.43)
4	1.51 (1.26–1.80)‡	1.60 (1.32–1.94)‡
5	2.02 (1.72–2.37)‡	2.35 (1.94–2.85)‡
Dependency quintile
1	[Reference]	[Reference]
2	1.25 (1.08–1.45)‡	1.07 (0.91–1.25)
3	1.16 (0.99–1.37)	0.89 (0.74–1.06)
4	1.44 (1.22–1.69)‡	1.10 (0.91–1.31)
5	1.67 (1.41–1.96)‡	1.22 (1.01–1.48)‡
Residential instability quintile
1	[Reference]	[Reference]
2	1.25 (1.05–1.48)‡	1.09 (0.91–1.31)
3	1.38 (1.16–1.64)‡	1.05 (0.87–1.26)
4	1.74 (1.47–2.05)‡	1.13 (0.94–1.36)
5	1.52 (1.30–1.79)‡	0.97 (0.80–1.16)
Ethnic diversity quintile
1	1.68 (1.43–1.98)‡	1.62 (1.30–2.01)‡
2	1.47 (1.26–1.73)‡	1.56 (1.29–1.90)‡
3	1.34 (1.14–1.57)‡	1.53 (1.28–1.82)‡
4	1.18 (1.02–1.37)‡	1.37 (1.17–1.61)‡
5	[Reference]	[Reference]
Immigration status
Child immigrant	0.59 (0.42–0.83)‡	0.67 (0.47–0.95)‡
General population	[Reference]	[Reference]
Geographical context
Rural–North	1.23 (0.82–1.85)	0.80 (0.52–1.25)
Rural–South	1.46 (1.22–1.75)‡	1.22 (0.99–1.51)
Urban–North	2.05 (1.63–2.60)‡	1.45 (1.13–1.87)‡
Urban–South	[Reference]	[Reference]

OR indicates odds ratio.

^*Children with missing data were excluded from these analyses (n=154).

^† Adjusted for all 4 dimensions of marginalization, age, immigration status, and geographic context.

^‡ Indicates statistically significant findings.

The odds of experiencing POHCA were higher for children living in Urban‐North (OR, 2.05 [95% CI, 1.63–2.60]) and Rural‐South regions (OR, 1.46 [95% CI, 1.22–1.75]), as compared with children living in Urban‐South regions. The comparison of POHCA odds between Rural‐North and Urban‐South residents was not significant (OR, 1.23 [95% CI, 0.82–1.85]). The direction of effects and findings of significance were consistent in the adjusted model for the comparison involving Urban‐North (aOR, 1.45 [95% CI, 1.13–1.87]) and Urban‐South; however, the comparison between Rural‐South (aOR, 1.22 [95% CI, 0.99–1.51]) and Urban‐South was no longer significant.

The likelihood ratio tests indicate that 4 of the 5 sociodemographic categorical variables are significantly associated with the risk of experiencing POHCA (P values <0.05) (Table 4). The residential instability variable was not significantly associated with POHCA risk (P=0.394).

Table 4.

Likelihood Ratio Test Statistics

Model Comparison	LR	DF	P value
Full model vs full model without material deprivation variable	102.74	4	<0.0001
Full model vs full model without ethnic diversity variable	33.66	4	<0.0001
Full model vs full model without residential instability variable	4.09	4	0.3939
Full model vs full model without dependency variable	11.79	4	0.0190
Full model vs full model without geographic context variable	12.44	3	0.0060

DF indicates degrees of freedom; and LR, likelihood ratio.

While not the primary focus of this study, we conducted separate analyses on neighborhood‐level income quintile to facilitate future data synthesis, given its common use as a measure of SES. We also examined the distinct components of the geographic context variable (ie, rurality and Ontario region). In these analyses, we found that children living in low‐income communities had a higher risk for POHCA relative to children living in high‐income communities, and northern and rural residence was independently associated with increased POHCA risk (Table S2 and S3).

Discussion

In this study, we measured the impact of neighborhood‐level marginalization, immigration status, and geographic context on the odds of experiencing POHCA in Ontario, Canada. We found significant associations between these factors and POHCA risk.

We found that children living in areas with the highest levels of material deprivation had a higher odds of experiencing POHCA, as compared with children living in areas with the lowest levels of material deprivation. This is consistent with evidence in adult populations, where individuals living in low SES neighborhoods have an elevated risk of OHCA. ²⁷ , ³⁴ , ³⁵ These disparities may be due to a higher prevalence of some prearrest risk factors in areas with high levels of socioeconomic marginalization. Our sample of cases included children who experienced OHCA due to sudden infant death syndrome or drowning. There is evidence of a potentially elevated risk of sudden infant death syndrome in low SES groups. ³⁶ These disparities may stem from differences in health literacy skills or limited access to health care resources that promote sudden infant death syndrome prevention through education on prominent risk factors such as infant sleeping position or smoking during pregnancy. Improving health care access and tailoring prevention efforts toward vulnerable populations could prove effective in reducing disparities. Furthermore, prior research suggests that children in low‐income neighborhoods are less likely to know how to swim, potentially increasing their risk for drowning. ³⁷ , ³⁸ These differences may partially explain the observed disparities in POHCA risk; however, future studies should consider stratifying the analysis by etiology to ascertain whether the aforementioned conditions are overrepresented in socially and economically deprived groups.

There was also a higher odds of POHCA for children living in neighborhoods with the highest levels of dependency, as compared with children living in neighborhoods with the lowest levels of dependency. As this dimension of marginalization incorporates information on the proportion of residents who are unemployed or in unpaid professions, areas with high levels of dependency may also have lower median household incomes and higher levels of material deprivation. Thus, the decreased magnitude of the effect estimate in the adjusted model may be due to higher levels of material deprivation in these regions.

Children living in neighborhoods with the highest levels of instability had a higher odds of POHCA, as compared with children living in neighborhoods with the lowest levels of instability. Previous literature indicates that people living in communities with high residential turnover are more likely to have poor access to health care. ³⁹ There is also evidence that children who move frequently have a higher likelihood of experiencing negative health outcomes, including chronic illnesses and poor physical health. ⁴⁰ Importantly, the observed association was no longer significant in adjusted models, suggesting that other factors may contribute to the impact of residential instability on POHCA risk. Prior research indicates that low‐income families move often, either voluntarily or involuntarily, as they are more likely to face eviction or foreclosure. ⁴¹ , ⁴² , ⁴³ Thus, the attenuation of this effect may be driven by higher levels of material deprivation in neighborhoods with high levels of residential instability.

There was also a higher odds of POHCA for children living in areas with lower levels of ethnic diversity, as compared with children living in neighborhoods with the highest levels of ethnic diversity. These findings are in contrast with prior literature, where membership in a racialized group is often associated with an elevated risk of OHCA in both adult and pediatric populations. ¹⁹ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ Importantly, the majority of this evidence focuses on populations in the United States, where there is a different sociopolitical climate and where the health care system differs from that of Canada. However, the impact of neighborhood‐level ethnic diversity was consistent with the decreased risk of POHCA in immigrants. This association may be partially explained by the “healthy immigrant effect,” a phenomenon whereby recent immigrants to Canada often have better health relative to those born in the country. ⁵⁰ Prior research on the healthy immigrant effect has focused on adult populations, and evidence on its impact in children is limited and mixed. ⁵¹ , ⁵² , ⁵³ The findings from our study highlight the importance of fully understanding the impact of the healthy immigrant effect in pediatric populations.

The odds of POHCA were higher for children living in Urban‐North and Rural‐South regions, as compared with children living in Urban‐South regions. Importantly, the comparison between Rural‐South and Urban‐South was no longer significant in adjusted analysis, whereas the comparison between Rural‐North and Urban‐South was not significant in crude or multivariable models. The latter comparison was likely underpowered due to the small sample size in Rural‐North strata. There are long‐standing disparities in health outcomes and health care access in northern and rural communities in Canada. ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ There is conflicting evidence on the impact of geography on OHCA risk in adult populations, although several studies report disparities in survival, likely due to time‐to‐treatment delays in rural and remote regions. ⁵⁹ These differences also highlight the importance of ensuring adequate POHCA management training among emergency medical services responders in these regions.

There is also a lack of access to primary care and urgent care in rural and northern Ontario communities. A recent Ontario‐based study found that 68 rural communities were more than an hour away from a primary care practice, with 59 of these communities being in northern Ontario. ⁵⁴ Geographic barriers to care likely have a disproportionate impact on low‐income families that do not have access to reliable means of transportation. These factors can force families to forgo or delay treatment for persistent or emergent health issues, potentially leading to the aggravation of underlying chronic conditions and increasing the risk of preventable health outcomes.

Limitations

This study should be interpreted in context of the following limitations. The findings establish associations between variables; however, these relationships may not necessarily be causal, as there is a risk for unmeasured confounding bias. We did not have access to individual‐level measures of marginalization or SES. The use of neighborhood‐level proxies for these data may have led to misclassification. We focused our analysis in a specific region, and further research is needed to determine whether disparities exist across different geographic contexts. There were also limitations associated with the use of the Immigration, Refugees, and Citizenship Canada Permanent Residents Database. The portion of this database housed at ICES does not include immigrants who originally landed in another province or immigrants who landed in Ontario but relocated shortly thereafter. Although we analyzed the impact of immigration status at the level of the child, we were limited in our ability to identify parental migrant status in the health administrative data holdings. Thus, our findings may not reflect the risk of POHCA for children who are second‐ or third‐generation immigrants. Moreover, the definition of the geographic context variable may not have been sufficiently granular to capture the impact of place of residence, as we did not account for important factors such as proximity to major health care centers. The algorithm used to identify POHCA cases is characterized by excellent sensitivity (87.3%) and specificity (99%); however, we were unable to capture children who experienced OHCA but were not transported to a hospital. Importantly, children who are not transported to a hospital are often dead on arrival or palliated and expected to die. We also did not have access to arrest characteristics, but these were beyond the scope of the study. The representativeness of the sample may be measured by the incidence of POHCA in our cohort. We calculated a crude annual incidence of 4.2 cases per 100 000 children, ²¹ a figure that falls within the range derived from previous studies. ¹ , ³ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

Conclusions

The findings from our analyses indicate that children living in neighborhoods with high levels of marginalization have an elevated risk of experiencing POHCA. Additionally, residence in northern urban regions was associated with a higher odds of POHCA, relative to southern urban residence. These communities should be prioritized in POHCA prevention and intervention efforts, which could include targeted education, awareness, and cardiopulmonary resuscitation training initiatives, as well as improved access to primary and urgent care services. Future analyses should investigate the contextual factors that contribute to lower POHCA risk for immigrants, and larger studies are needed to better understand the impact of geographic context on POHCA risk.

Sources of Funding

None.

Disclosures

None.

Supporting information

Tables S1–S3