Food Insecurity and Experiences of Discrimination Among Caregivers of Hospitalized Children Free

This study was conducted at a 155-bed academic children’s hospital in Chicago, Illinois. The population in the hospital’s primary service area is predominantly African American or Black (74% of 626 264 in the hospital’s service area) and more than one-quarter of people (27%) in this area live below the Federal Poverty Level.¹⁸  The hospital has ∼5300 admissions each year.¹⁹ 

Caregivers of children admitted to the hospital’s general inpatient and ICUs and enrolled in the control arm (“usual care”) of the CommunityRx for Hunger double-blind randomized controlled trial between November 2020 and June 2022 were included in this analysis.²⁰ 

Eligible English- or Spanish-speaking caregivers lived in a 42-zip-code region, including the hospital’s 12-zip-code primary service area, on Chicago’s South and West sides and adjacent suburbs and self-identified as the primary caregiver of a hospitalized child. Eligible caregivers also consented to receive text messages from the research team. Caregivers of non-NICU newborns, children hospitalized <24 hours or >30 days, children with a diagnosed eating disorder, and those who recalled participating in a previous Community-Rx study were excluded from the study. Eligible caregivers were consented and enrolled after their child’s hospital admission, before discharge. All caregivers provided written documentation of informed consent, documented electronically. During the consent process, participants were informed that the study was about food security and related needs. The study protocol was approved by the University of Chicago’s Institutional Review Board (IRB20-0324).

The 319 caregivers enrolled in the control arm of the CommunityRx for Hunger trial and included in this analysis received usual clinical care at the time of admission. Usual clinical care included information in the standard “Caregiver FYI Admissions Packet” from hospital admission staff about all available retail food options in the hospital and self-serve food pantries operated by Feed1st, a longstanding 24/7/365 self-serve food pantry program operating at multiple sites in the children’s hospital.^13,21  Usual clinical care also included a referral to social work at the discretion of the health care team. Usual clinical care did not include screening for FI.

Validated survey measures were used when possible. Sociodemographic characteristics were assessed in the baseline survey administered during the child’s admission. Gender identity was assessed with the following item recommended by the 2014 Best Practices for Asking Questions to Identify Transgender and Other Gender Minority Respondents on Population-Based Surveys.²²  While the response options provided for this item were “Male” vs “Female”, indicating sex, the item also included the following options: Trans male/Trans man, Trans female/trans woman, Genderqueer/gender non-conforming, and Different identity.²²  We also assessed sex as a biological variable in accordance with NIH policy, and found no significant differences in sex vs gender. Relationship with the hospitalized child was assessed with an item sourced from the 2016 Consumer Assessment of Healthcare Providers and Systems Child Hospital Survey.²³  Caregiver race, caregiver ethnicity, and partner status were assessed with items from the 2018 Behavioral Risk Factor Surveillance System (BRFSS).²⁴  Race and ethnicity were assessed with 2 items: “Which one or more of the following would you say your race is?” (American Indian or Alaskan Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, White, or other)” and “Are you Hispanic, Latino/a, or of Spanish origin?” (yes or no).²⁴  Because of small numbers, participants who selected American Indian or Alaskan Native, Asian, Native Hawaiian or Pacific Islander, other race, or multiple races were categorized as “Other.”

Economic and household characteristics were also assessed in the baseline survey. Annual income, educational attainment, and insurance status were assessed with items from or adapted from the 2018 BRFSS.²⁴  The median number in the household was assessed with an item from the 2018 American Community Survey.²⁵  The median number of children <18 years old in the household was assessed with an item from the 2018 BRFSS.²⁴  Household receipt of Supplemental Nutritional Assistance Program (SNAP) and Special Supplemental Nutritional Program for Women, Infants, and Children (WIC) were assessed with items from the 2014 Hunger in America Survey.²⁶ 

Household FI in the 30 days before the child’s hospitalization was assessed in the baseline survey with the 18-item US Household Food Security Survey Module.²⁷  Affirmative responses were summed by using standard scoring to generate a raw score of 0 to 18.²⁷  Raw scores were first used to categorize households into 4 food security statuses as specified by the US Household Food Security Survey Module: food security (0 items endorsed), marginal food security (MFS; 1–2 items endorsed), low food security (3–5 items endorsed), and very low food security (≥6 items endorsed).²⁸  However, because of a small sample size of participants from households affected by very low food security, we collapsed households experiencing low and very low food security and categorized households into 3 food security statuses: food security (0 items endorsed), MFS (1–2 items endorsed), food insecure (≥3 items endorsed).²⁹ 

Experiences of perceived discrimination in the children’s hospital were assessed in the survey administered 7 days after the child’s discharge. Experiences of discrimination were measured as a proxy for stigma by using the 7-item Discrimination in Medical Settings (DMS) Scale.³⁰  The DMS Scale was developed to assess the frequency of individual-level discrimination in the health care setting attributed to “race, ancestry, or national origin.”³⁰  To capture discrimination during the child’s hospitalization, the measure’s introduction was adapted to say, “Now I’m going to ask you about your experiences during your child’s hospitalization. Thinking about your experiences when you child was hospitalized last week, would you say…”. In this introductory statement, no reference was made to any specific reason for discrimination, such as race, ancestry, national origin, or other factors. Standard scoring with a 5-point Likert scale was used (1: never, 2: rarely, 3: sometimes, 4: most of the time, and 5: always). Preliminary analyses revealed skewed distributions in responses to each DMS item, with 59% to 90% of participants responding “never,” as consistent with findings from the 2011 validation study.³⁰  Therefore, responses were dichotomized according to whether caregivers had any experiences of discrimination (raw score >7) or no experiences of discrimination (raw score = 7) during their child’s hospitalization, as outlined by Benjamins and Middleton et al.³¹ 

Descriptive statistics were calculated for sample sociodemographic characteristics and responses to DMS items by food security status, including medians with interquartile ranges (IQRs; continuous variables with skewed distributions) and response frequencies (categorical variables). Bivariate analyses were performed to assess statistically significant differences in sample characteristics by food security status. Logistic regression, both unadjusted and adjusted, was used to test the relationship between food security status and discrimination in medical settings as reported overall (dichotomized any vs no experiences of discrimination during child’s hospitalization) and per-item on the DMS scale. Analyses were repeated by using Poisson regression with robust standard errors to estimate adjusted and unadjusted relative risks. Models were adjusted for variables previously reported to be associated with FI and discrimination: gender (“female” vs “male”), race and ethnicity (non-Hispanic, African American or Black vs all others), partner status (married or in a relationship vs not), and income (<$50 000 vs ≥$50 000). These variables were selected because they are well-documented in the literature to be related to both FI²  and experiences of discrimination.⁵  Race and ethnicity were combined to create a single covariate in these models because caregivers who identified as non-Hispanic, African American or Black comprised the vast majority in our sample. Because a small percentage of participants identified as Hispanic, including ethnicity and race as separate covariates affected model convergence and decreased precision.

Participants who did not complete all relevant items were excluded from these models. The significance level was set at α = 0.05. All analyses were conducted with Stata version 16 (StataCorp LLC, College Station, TX).

The majority of participants were African American or Black, women, and the parent of the hospitalized child (Table 1). Participants had a median age of 33.5 years (IQR = 28.3–40.1). Most caregivers were insured through Medicare/Medicaid (74.4%) and lived in a household enrolled in government food assistance through SNAP and/or WIC (75.0%).

At baseline, >1 in 4 caregivers were food insecure in the 30 days before their child’s hospitalization. An additional 15.1% were marginally food secure. Most sample characteristics were similar by food security status. Differences by food security status were observed for annual household income (overall P < .001), with caregivers from food-insecure households reporting annual incomes of <$50 000 at the highest rate (92.3% for caregivers from food-insecure households vs 89.4% from marginal food-secure households and 70.8% from food-secure households). Also, differences in insurance status by food security status were observed (P = .03), with caregivers from marginal food-secure households reporting being insured with Medicaid/Medicare at the highest rate (89.4% for caregivers from marginally food-secure households vs 78.5% from food-insecure households and 69.0% from food-secure households).

More than one-half of caregivers reported experiencing discrimination in the medical setting during their child’s hospitalization (Fig 1). Experiences of discrimination were highest for caregivers from food-insecure households, followed by those from marginally food secure and food-secure households (59.7% vs 57.1%, and 47.4%, respectively, P = .20). Caregivers from food-insecure households had the highest rate of discrimination for 5 of 7 DMS items, with feelings of “being treated with less respect than other people” being the most prevalent: 51.5% for caregivers from food-insecure households (P = .001) versus 42.9% from marginally food-secure households (P = .4), compared with 28.7% from food-secure households.

In logistic models adjusted for baseline characteristics, caregivers from food-insecure households had twice the odds of experiencing any discrimination during their child’s hospitalization compared with those from food-secure households (adjusted odds ratio [aOR] = 2.0, 95% confidence interval [CI] 1.1–3.6, P = .03; Fig 2). The odds of experiencing discrimination during the child’s hospitalization were not significantly higher among caregivers from marginally food-secure households compared with caregivers from food-secure households (aOR = 1.6, 95% CI 0.7–3.4, P = .25).

In adjusted logistic models, caregivers from food-insecure households had significantly higher odds of experiencing 5 of the 7 types of discrimination assessed during their child’s hospitalization compared with caregivers from food-secure households. The adjusted odds of being “treated with less respect than other people” were >3-fold higher (aOR = 3.4, 95% CI 1.8–6.5, P < .001; Fig 2). In addition, caregivers from marginally food-secure households had significantly higher odds than caregivers from food-secure households of reporting that they “receive[d] poorer service than other people” at the child’s hospital (aOR = 3.0, 95% CI 1.3–6.6, P < .01). These results are consistent with adjusted relative risk estimates (Supplemental Fig 3).

More than 1 in 4 caregivers of a child hospitalized during the COVID-19 pandemic, mostly lower-income African American or Black women, lived in a household that was food insecure in the 30 days before their child’s hospitalization. Household FI, especially during a pandemic that has disproportionately impacted African American and Black people,⁹  presents a major risk for poor outcomes after a child’s hospitalization. Without access to a sufficient diet, a child may not be able to adhere to medication regimens, healing can be compromised by poor nutrition, and caregivers can struggle to make critical caregiving and other medical decisions for their child. As hypothesized, we found that household FI in the month before the child’s hospitalization was a predictor of experiences of discrimination for caregivers during their child’s hospitalization.

Our data were collected in parallel with unprecedented federal interventions to address a massive food security crisis in the United States throughout the COVID-19 pandemic.³²  Even with these interventions, the proportion of households experiencing MFS or FI in our sample during the pandemic (40.2%) was similar to the rate in a 2011 study at the same site (43.6%, using a 12-month recall period)¹³  and were several-fold higher than national averages from mid-November to mid-December 2020 for households with children (7.2%) and with heads of households who identified as non-Hispanic, African American or Black (11.8%).³³  The high rates of FI among families with a hospitalized child can be attributed in part to hospitalization itself¹³  and supports emerging policies and calls from the Centers for Medicare and Medicaid Services, the Joint Commission, the National Quality Forum, and others to move health systems toward integrating social with medical care.^34–36 

More than one-half of caregivers in this study reported experiencing discrimination in the medical setting during their child’s hospitalization. In addition, caregivers from food-insecure households were at significantly higher odds of experiencing 5 of the 7 types of discrimination assessed compared with caregivers from food-secure households. In this study, participants were not asked to indicate the perceived cause of discrimination. It is possible that parents and other caregivers in our study would have attributed their experiences of discrimination to race, gender, socioeconomic, or other circumstances. As described in the Stigma and Food Inequity Conceptual Framework, manifestations of stigma (like the perception of discrimination) have harmful effects on the psychosocial processes and behavior of people with FI, ultimately worsening food inequities and health.⁵  The link between experiences of discrimination and food assistance use is well-documented in the case of SNAP; perceived discrimination by SNAP caseworkers triggers feelings of stigma that deter eligible participants from accessing food assistance.^10–12  To our knowledge, the current study is the first to assess the relationship between food security status and caregivers’ experiences of discrimination in the children’s hospital setting.

The finding that caregivers from food-insecure households had twice the odds of experiencing discrimination during their child’s hospitalization compared with caregivers from food-secure households has implications for efforts to intervene in FI in the children’s hospital setting. As reported in a 2019 cross-sectional study, the majority of African American or Black patients and adult caregivers of pediatric patients (79%) saw health-care-based screening for social risks as appropriate.³⁷  At the same time, consistent with the current study findings, qualitative studies of adult patients and caregivers of pediatric patients highlight stigmatization, including discrimination, as a barrier to both the disclosure of need and the use of referral resources.^38–42  In settings in which FI is highly prevalent, the benefits of a screening-based approach to assistance versus a universal approach (for example, availing everyone in the clinical setting access to emergency food assistance and information about community-based food supports) ought to be weighed against risks. The authors of the CommunityRx-Hunger randomized controlled trial are evaluating the impact of a universal approach to intervention in a setting in which everyone has 24/7/365 access to emergency food support.^13,20,21 

In addition, this study sheds new light on experiences of discrimination in the hospital setting among caregivers from marginally food-secure households. Specifically, this group had significantly higher odds of reporting receiving poorer service than other people during their child’s hospitalization. Households with MFS are regarded as similar to those with food security in most food security research and in practice.⁴³  However, the experience of MFS is increasingly being recognized in the literature as having a unique psychosocial profile.^29,44–46  This study offers more evidence for consideration of marginally food secure caregivers as a distinct group.

Findings should be interpreted in light of certain limitations. The informed consent process did explain to study participants that the study was about food security and related needs, a factor that could have affected or primed their responses regarding both FI and discrimination.⁴⁵  Preenrollment screening for health-related social risks for the purpose of randomization into the randomized controlled trial may have impacted perceptions of discrimination. Although, with this study, we find a relationship between FI and experiences of health care discrimination, we do not discern perceived causes of discrimination. The single-site design and focus on caregivers of hospitalized children may limit this study’s generalizability to other populations. Rapid cycle temporal trends in food access and food support related to the COVID-19 pandemic likely contributed to observed rates of FI throughout the study period. In addition, contemporaneous increases in societal awareness and action on issues of racism (health care- and community-based) and increased rates of health care worker burnout⁴⁷  may have increased perceptions of discrimination during the study period.

Experiences of discrimination in the children’s hospital setting were prevalent overall and were associated with household FI in the 30 days before a child’s hospitalization. Interventions in the children’s hospital setting to mitigate FI should be delivered in a manner that optimizes the care experience and minimizes stigma.

We would like to thank Eva Shiu, MPH, for overseeing data collection and Charlie Fuller and Veera Anand for verifying our citations. We would like to acknowledge the Feed1st Community Advisory Board for their ongoing guidance, commitment and collaboration on this work.

aOR

adjusted odds ratio

BRFSS

Behavioral Risk Factor Surveillance System

CI

confidence interval

COVID-19

coronavirus disease 2019

DMS

Discrimination in Medical Settings

FI

food insecurity

MFS

marginal food security

SNAP

Supplemental Nutritional Assistance Program

WIC

Special Supplemental Nutritional Program for Women, Infants, and Children