Prevalence and characterization of food insecurity in a Canadian paediatric emergency department

Abstract

Objectives

Food insecurity (FI) is associated with a number of adverse child health outcomes and increased emergency department (ED) use. The COVID-19 pandemic exacerbated the financial hardship faced by many families. We sought to determine the prevalence of FI among children with ED visits, compare this to pre-pandemic rates, and describe associated risk factors.

Methods

From September to December 2021, families presenting to a Canadian paediatric ED were asked to complete a survey screening for FI along with health and demographic information. Results were compared to data collected in 2012. Multivariable logistic regression was used to measure associations with FI.

Results

In 2021, 26% (n = 173/665) of families identified as food insecure compared to 22.7% in 2012 (n = 146/644) a difference of 3.3% (95% CI [−1.4%, 8.1%]). In multivariable analysis, greater number of children in the home (OR 1.19, 95% CI [1.01, 1.41]), financial strain from medical expenses (OR 5.31, 95% CI [3.45, 8.18]), and a lack of primary care access (OR 1.27, 95% CI [1.08, 1.51]) were independent predictors of FI. Less than half of families with FI reported use of food charity, most commonly food banks, while one-quarter received help from family or friends. Families experiencing FI expressed a preference for support through free or low-cost meals and financial assistance with medical expenses.

Conclusion

More than one in four families attending a paediatric ED screened positive for FI. Future research is needed to examine the effect of support interventions for families assessed in medical care facilities including financial support for those with chronic medical conditions.

Food insecurity (FI), defined as inadequate or insecure access to food due to financial constraints, is a key determinant of child health. FI is correlated with material hardship and family stress and was estimated to affect one in six Canadian children in 2017 to 2018 (1). A growing body of literature has described associations between FI and a number of health disparities, including increased obesity (2–4), asthma, depressive symptoms (5), early childhood developmental risk, behaviour dysfunction, and poor academic performance (6–13). FI has also been associated with poor access to primary care, increased emergency department (ED) visits and hospital admissions, and challenges with chronic disease management (5,14,15).

The COVID-19 pandemic has exacerbated the financial hardship faced by many families, particularly those experiencing lower household income and FI (16,17). There has been an unprecedented increase in the use and government funding of charitable food assistance programs, with persistent unmet needs (16,18). Evidence suggests the majority of families experiencing FI do not utilize food charity, but may compromise other expenses or seek help from friends or family (19). The Hospital for Sick Children in Toronto showed a household FI prevalence of 34.2% among families admitted to the hospital during the COVID-19 pandemic (20). Given the association of FI with paediatric ED use, this clinical setting is well-positioned to identify families.

We sought to determine the prevalence of FI in our paediatric ED population in the context of the COVID-19 pandemic, to compare this to pre-pandemic rates, and to describe associated social and health-related risk factors. Moreover, we aimed to examine the utilization of community services and identify families’ preferred means for future support interventions within the hospital.

METHODS

This was a cross-sectional study to determine the prevalence of FI and associated risk factors among families visiting our paediatric ED in 2021, compared with similar data collected in 2012.

McMaster Children’s Hospital (MCH), in Hamilton, Ontario, is a tertiary Canadian children’s hospital serving a diverse catchment of 2.2 million people. The region has been documented to have higher rates of poverty than national and provincial averages (19,21).

Surveys were distributed from September to December 2021. All families presenting to the ED with a paediatric patient (<18 years) were eligible to be included. Families were approached in the ED waiting room between 0700 and 0100 daily, subject to the availability of a research team member. Participants were offered paper surveys in English, Arabic, French, or Spanish, based on previous institutional data of these being the most commonly requested languages for interpretation in the ED. Families were provided with study information by a member of the research team and a participant information form. All participants received a handout of community food resources and were offered the option to speak with an ED social worker. Surveys contained no identifying material and were returned voluntarily by participants upon completion to labelled collection boxes in the ED.

Caregivers of critically ill patients and those who were not primary caregivers of the child were excluded. Families with language or literacy barriers that prohibited completion of the written survey in English, Arabic, French, or Spanish could self-exclude. The study was approved by the Hamilton Integrated Research Ethics Board (HiREB) #2021-10718-GRA. Due to the anonymous and voluntary nature of the survey, the letter of information that every family received stated explicitly that the return of the survey implied consent to participate in the study.

Household FI status was assessed using the Hunger Vital Sign (HVS) tool, which has been validated in American ED and primary care contexts against the 18-item gold standard Household Food Security Survey Module (sensitivity 96.7%, specificity 86.2%) (22,23). The HVS identifies households at risk for FI by responses of ‘sometimes true’ or ‘often true’ to either of the following statements: (1) within the past 12 months, we worried whether our food would run out before we got the money to buy more; or (2) within the past 12 months, the food we bought just didn’t last and we didn’t have money to get more (22). Additional survey questions exploring information on household demographics, child health and food assistance were collated by the project advisory team and reviewed by the Pediatric Emergency Medicine Research Oversight Committee and the family advisory council, for internal validity (Supplementary Appendix 1).

As a comparator group, results were compared to a screening study using the HVS tool conducted by a study team member (QN) in 2012, within the same institution with similar methods and sample size (n = 640) (24).

Statistical analysis

Primary outcome.

The prevalence of FI among families presenting to the ED was calculated with 95% confidence intervals and compared with that found in 2012 and that of the Ontario and Canadian population using χ² analysis. Based on the most recent available provincial prevalence (2020) of 16.1% (25), and 2012 data from the MCH ED of 22.6% (24), an estimated sample size of 272 was needed to determine the prevalence of FI with a 5% risk of type 1 error.

Secondary outcomes.

Child health characteristics, household demographics, use of community food assistance and preferred means of support in-hospital were assessed using descriptive statistics and examined using χ². Using backward stepwise elimination, multivariable logistic regression analysis was performed, with FI as the outcome variable and emergency department use, lack of primary care, caregivers born outside of Canada, status other than Canadian citizenship, having a child with a chronic medical condition, having financial strain pertaining to child’s medical expenses and the reason for the ED visit being the child’s chronic condition as the predictor variables. Results were reported as odds ratios (OR) and 95% confidence intervals (CI). A P value < 0.05 was considered significant.

RESULTS

A total of 1312 surveys were distributed with 665 (50.7%) completed. The prevalence of FI in the study population was 26% (95% CI [22.6%, 29.4%], (173/665 families)). Compared to a prevalence of 22.7% at MCH ED in 2012 (n = 146/640), the prevalence of FI in 2021 was 3.3% higher [−1.4%, 8.1%] (Figure 1). Demographic and child health characteristics are summarized in Table 1.

Figure 1.

Prevalence of food insecurity (%).

Table 1.

Child health and self-reported household demographic data in relation to food security status

	Food-insecure, n (%)	Food-secure, n (%)	P value
Total families	173 (26%)	492 (74%)
Primary care by family doctor or paediatrician	117 (67.6%)	436 (88.6%)	<0.001
Primary care by ED, walk in clinic, or no primary care	56 (32.4%)	56 (11.4%)	<0.001
Child with chronic medical condition	75 (43.4%)	138 (28%)	<0.001
Medical expenses for child causing financial strain (all comers)	96 (55.5%)	71 (14.4%)	<0.001
Child with chronic medical condition and medical financial strain	66/75 (88%)	55/138 (40%)	<0.001
ED visit reason related to child’s chronic medical condition	75 (43.4%)	138 (28%)	<0.001
Child with special diet due to medical condition	45 (16.2%)	49 (10%)	NS
Number of children in household (mean)	2.35 +/-	2.05 +/-	0.002
Household receiving social assistance	87 (50.3%)	81 (16.5%)	<0.001
Caregiver born in Canada	112 (64.7%)	379 (77%)	0.002
Caregiver Canadian citizen	128 (74%)	434 (88.2%)	<0.001
Caregiver with status other than Canadian citizena	45 (26%)	58 (11.8%)	<0.001

^aIncludes permanent resident, refugee, student visa, work visa or temporary worker, and primary identification as First Nations, Métis or Inuit

In the univariate analysis, FI was associated with having a child with a chronic medical condition (OR 1.5, [1.2–1.9]), having financial strain related to a child’s medical expenses (OR 3.8, [3.0, 5.0]) and the reason for the ED visit being the child’s chronic condition (OR 1.5, [1.2–1.9]). FI was also associated with a lack of primary care (OR 2.8, 95% CI [−2.1, −3.9]). The prevalence of FI was higher among caregivers born outside of Canada (OR 1.54, [1.19–1.99]), and status other than Canadian citizenship (OR 2.2, [1.3–3.1]). While the sample size was small, all caregivers who identified as refugees (n = 6/6) screened positive for FI.

Multivariable analysis identified that having a greater number of children at home (OR 1.19, 95% CI [1.01, 1.41]), financial strain from the child’s medical expenses (OR 5.31 [3.45, 8.18]) and a lack of primary care access (OR 1.27, [1.08, 1.51]) were independently associated with FI. Details of the logistic regression can be found in Supplementary Appendix 2.

Less than half (39%) of families experiencing FI reported use of food charity, most commonly food banks (32%), while one-quarter (26%) reported receiving help from family or friends. Almost one in ten families (9%) screened negative for FI yet reported use of food charity. Table 2 summarizes families’ utilization of community food resources.

Table 2.

Reported use of community food assistance by families with and without food insecurity.

	Food-insecure, n (%)	Food-secure, n (%)
Use of charitable food assistance (excluding help from family/friends)	67 (39%)	44 (9%)
Food banks	55 (32%)	16 (3%)
Community kitchen	2 (1%)	2 (0.4%)
Community meals	3 (2%)	3 (1%)
Good food box	7 (4%)	11 (2%)
School meals	24 (14%)	19 (4%)
Food assistance from family members or friends	45 (26%)	16 (3%)
Other	7 (4%)	2 (0.4%)
Prefer not to say	19 (11%)	15 (3%)

Families experiencing FI expressed a variety of preferred means of support in the hospital, summarized in Table 3. The most popular were free or low-cost meals, followed by financial assistance with their child’s medical expenses, access to a kitchen to prepare meals and the presence of a food pantry in the hospital.

Table 3.

Preferred means of in-hospital material support expressed by families experiencing food insecurity

Free or low cost meals provided by hospital	108 (62%)
Financial help for child’s medical expenses	75 (43%)
Food bank or pantry in hospital	70 (40%)
Access to kitchen in hospital	68 (39%)
Information on local food banks/programs	49 (28%)
Meeting with social worker	48 (28%)
Not needed	5 (3%)

DISCUSSION

In this cross-sectional study, we found that more than one in four families presenting to the MCH ED report FI. The prevalence of FI in our sample was higher than the most recently reported provincial (16.1%) and national averages (15.8%) from 2020 (25). When compared to our setting in 2012, FI rates demonstrate a trend towards increasing prevalence (26% compared to 22.7%) (24). Among other factors, we demonstrated that FI was independently associated with financial strain from children’s medical expenses, emphasizing that FI is a major issue among the families we serve. As health care providers and institutions, we must advocate for better financial support for low-income families and food access in health care settings.

The timing of the study during the COVID-19 pandemic may have contributed to the high prevalence of FI observed. Households with lower incomes have been particularly affected by pandemic-related financial hardship (16). Yet, the high FI rate in the 2012 comparator data suggests there are longstanding, pandemic-independent, systemic issues perpetuating inequality. Rising food prices and inflation suggest vulnerable families will continue to face greater difficulty (26,27).

Our study found higher rates of FI amongst caregivers born outside of Canada and those with a status other than Canadian citizenship. There is a known higher prevalence of FI among recent immigrants, which may explain our finding of particularly increased risk among those who identified as refugees (1). Newcomers to Canada face unique challenges, which may include housing, employment and financial instability, language barriers and fewer social supports. The COVID-19 pandemic exacerbated disparities through social isolation and reliance on language and literacy to navigate government restrictions and supports (28). We need to ensure we work with communities to develop culturally appropriate resources.

We found a significant association between households with FI and childhood chronic diseases, including physical, neuro-behavioural, and psychiatric conditions (reported in 43.4% of households with FI vs. 28% of food-secure households). Among families with FI and a child with chronic illness, the vast majority (88%) reported difficulty with medical expenses. Moreover, a significant number (40%) of food-secure households with a child with a chronic condition also reported medical-related financial strain, which suggests the economic impact of having a child with a chronic illness is not limited to more vulnerable families. While Canada has a public healthcare system, with a provincial program that covers the majority of basic medications for anyone under 25 years of age, there are certain drugs, medical equipment, and services that are not covered and rely on families to pay out of pocket or through private benefit plans. Furthermore, having a child with a chronic condition often requires parents to attend frequent medical appointments or admissions, which may interfere with employment and household income. Despite having higher rates of childhood chronic disease, we found families with FI were less likely to receive the majority of their medical care through a family doctor or paediatrician, and instead reported relying on the ED or a walk-in clinic. Lack of primary care, compounded by financial hardship, presents challenges to chronic disease management, which could contribute to the observed increased rate of ED presentations due to the child’s underlying chronic condition among families with FI. Previous literature has highlighted associations between material hardship, higher rates of ED visits and greater unmet health care needs (15,29). These findings suggest that families with FI require more support in order to improve health outcomes for children, especially in the context of childhood chronic disease.

Existing food assistance programs in Canada leave unmet needs (30). A minority of families with FI reported use of food banks with a similar proportion receiving help from family and friends. FI literature demonstrates the tendency towards seeking assistance from one’s social network or compromising other essentials, including medical expenses, over using food charity (16). Existing research suggests FI screening in health care settings leading to food charity referral is not necessarily desired or effective, due to limited impact on household FI and low uptake by patients (31). Interventions that address underlying causes of household FI, such as poverty reduction strategies, are needed. Accessing food while admitted to the hospital presents an additional challenge, with expensive onsite options and expectations for caregiving at the bedside (20). Strategies that allow for self-referral to targeted programs for hospital-associated FI may be preferred by patients. A program trialled at MCH placed a poster in each patient room asking ‘Caregivers, have you eaten today?’ Patients who self-identified to their RN would be provided cafeteria gift cards. In our study, the majority of caregivers with FI expressed a preference for the provision of free or low-cost meals in the hospital. Options include meal vouchers, hospital cafeteria gift cards, or an additional meal tray from the hospital kitchen. Programs that offset costs related to childhood chronic illness could help address FI by targeting the underlying cause. Caregiver access to a kitchen during their child’s admission could be helpful both financially and for dietary limitations. In-hospital food pantries provide low-barrier access to healthy food, including medically-specific diets, and collaborate with allied health (32,33). The two least popular assistance strategies selected by families were to receive information about community resources or meet with a social worker. The literature describes that limited utilization of food charity is not typically due to a lack of knowledge (30). Families may be hesitant to meet with a social worker for a variety of reasons (34), though the skills and expertise of a medical social worker position them well to assess all aspects of social determinants of health. Integrating family preferences can help inform institutional programs that best support families with FI based on actual rather than perceived needs.

Limitations

Our study has several limitations. We observed almost one in ten families screened negative for FI yet reported the use of charitable food assistance. Previous literature has described a minority of families with FI using food charity, and typically those most severely affected (30). If those with reported use of food banks, community kitchens or meals from the 2021 data are included with positive screens, the prevalence of FI increased to 29.2%. This may suggest sensitivity limitations of the screening tool, while also acknowledging the phrasing of our question surrounding the utilization of food assistance did not specify a timeline, unlike the screening questions, which may have led to overestimation. Second, surveys were available in written format in English, Arabic, Spanish, and French. Families with literacy challenges or alternate language preferences self-excluded, and based on known risk factors, could bias our results toward an underestimation of FI. Third, our sampling method may produce selection bias, as survey distribution depended on research team presence. We endeavoured to mitigate this by having screening occur at various times of the day and week. Our response rate was approximately 50%, in keeping with reported averages for survey studies in the ED setting (35), however, may also incur selection bias. Finally, the study was based at a tertiary paediatric hospital and relied on the research team to deliver surveys. Existing workflows challenge the ability of ED staff to distribute surveys which may limit generalizability to other centres.

CONCLUSION

More than one in four families attending a Canadian paediatric ED screened positive for FI. Compared to a decade ago, the incidence of FI is trending higher. FI was independently associated with a lack of access to primary care, financial strain related to medical expenses, and greater number of children in the household. Further study is needed into the effect of support interventions, such as financial assistance for children’s medical expenses and providing meals for caregivers while in health care facilities.

PRIOR PRESENTATION

Poster of Abstract at Canadian Paediatric Society (CPS) Conference May 2022.

ETHICS APPROVAL

Hamilton Integrated Research Ethics Board #2021-10718-GRA.

FUNDING

This project was funded by a Pediatric Resident Research Grant, Department of Pediatrics, McMaster University, Hamilton, Ontario.

POTENTIAL CONFLICTS OF INTEREST

QNN holds a grant from eCampus Ontario for a separate project designing a virtual emergency department simulation space. SM’s research work is supported by the McMaster University Ronald Barr Professorship in Pediatric Oncology and a Hamilton Health Sciences Early Career Award. She also holds grants for other projects from the Canadian Institutes of Health Research. GW is supported by the Hamilton Health Sciences Early Career Award. She also holds grants with the Canadian Institutes of Health Research. EM is the recipient of the Pediatric Resident Research Award that funded this project. AP received a stipend for testimony at a coroner’s inquest unrelated to the topic of this manuscript. There are no other disclosures.