Food Insecurity Screening and Intervention in United States Children’s Hospitals

We used the Children’s Hospital Association directory to compile a database of 190 United States children’s hospitals, excluding specialty hospitals and institutions which exclusively provided newborn care.²⁸  We determined hospital size (bed number) by first assessing the Children’s Hospital Association directory. In cases where the bed number was not listed in the directory, it was determined via an internet search. We characterized hospital region based on United States census data and determined county-specific poverty rates using the County Health Rankings and Roadmaps.^30–31 

The survey was developed and refined by our research team and piloted with content experts with knowledge in survey methodology, epidemiology, and health services research who provided feedback. From the expert feedback, we modified the survey to include mostly multiple-choice (rather than free text) questions and to include specific items regarding FI screening procedures such as type of screener, when or where it is administered, and assessments of the types or qualifications for food provisions provided (Supplemental Table 4).

Before the COVID-19 pandemic, in February 2020, we initially distributed the survey to 10 individuals from different children’s hospitals in the United States from multiple departments including social work, nutrition, patient relations, and nursing. We specifically spoke to these care team members, rather than physicians, nurse practitioners, and physician assistants, because of known inpatient clinical workflows. After the 10 surveys were administered, we found that social work team members demonstrated the most consistent expertise in answering survey questions and willingness to participate in the study. Additionally, other groups outside of social work would often defer to the social work team at their institution when asked to complete survey questions.

Because of the COVID-19 pandemic and clinical responsibilities, research activities were paused until February 2021. We then redistributed the survey, informed by the prepandemic responses, between February 2021 through July 2021 using the following workflow¹ : We completed a Web site search for either a social work department phone number or the general phone number for the hospital.²  We then called and asked to connect with the inpatient social work department and to speak with either a social work manager or social worker. If neither was available, we left a message requesting a call-back. At least 2 attempts were made to reach each hospital.³  If we connected with someone from the social work department, we asked to speak with someone who had a global perspective of institutional FI screening and interventions such as a social work manager. If the social work team member shared that someone else at their institution could better answer questions regarding FI screening practices, we would attempt to connect with that individual even if they were from a different department.⁴  Once a person was identified and they agreed to participate in the study, respondents chose either E-mail or phone administration by member of the research team. E-mail respondents received 1 reminder E-mail if needed. We also redistributed the survey to the 10 institutions which completed survey before the survey modifications. There was no incentive for survey participation.

After contacting all institutions, 40.0% (76 of 190) submitted at least 1 response and 9.2% (7 of 76) of those institutions submitted multiple responses (Fig 1). We consolidated multiple responses, and when there were discrepancies from the same institution, we included the affirmative responses (example: answered yes to screening for FI), recognizing that an individual may not always have a global perspective of protocols and services. Descriptive statistics were performed by using Stata 15 (StataCorp, College Station, Texas), and between-group differences were assessed by using χ² test. This study was deemed exempt by the university institutional review board.

There was no significant difference in the hospital demographic characteristics for hospitals which had an individual complete the survey. Social work team members constituted the majority of participants in this study (79.0%). Most hospitals had between 100 to 300 staffed beds and were located in the Midwest or Southern regions of the United States. Most (68.4%) hospitals were located in a county with a child poverty rate between 10% to 20% (Table 1).

The majority (67.1%) of hospitals screened at least some individuals for FI. About one-third of institutions universally screened for FI on admission. The most frequently reported locations where FI screeners where administered included the inpatient unit (58.8%), multiple settings (33.3%), and the emergency department (13.7%). Individuals administering screeners included social workers (35.5%), nurses (34.2%), and physicians (4.0%). The most frequently reported method used to administer the screener was via oral interview (38.2%), followed by paper (10.5%), and electronic device (7.9%). Screener type included institution-specific (30.3%), “other” (15.8%), and the Hunger Vital Sign (2.6%). Interventions and responses to address identified FI included placing a social work consultation (51.3%), providing connections to community resources (47.4%), and/or offering food or meals (43.4%) (Table 2).

Most respondents (84.2%) reported administering food/meals to some caregivers. Conditional qualifications for food or meals included need-based determination (31.6%) and presence of a breastfeeding mother (30.3%) (Table 3, Fig 2). Need-based was further defined by participant free-text responses and included food insecurity (25.0%), medical team assessment (18.4%), and being out of town (11.8%).

In this cross-sectional survey of representatives of United States children’s hospitals, we found that, although many hospitals screened for FI, the majority did not conduct universal screening. Screening practices and interventions for identified FI varied among institutions. Most hospitals provided some food or meals to the caregivers of admitted pediatric patients, but there were often conditional qualifications.

Although the majority of institutions performed some FI screening, universal screening was conducted at one-third of institutions. In hospitals that did not perform universal screening, conditional qualifications for screening included admission to certain units and/or medical team discretion. These results are consistent with the previous study of 4 large United States children’s hospitals, in which it was found that universal screening was less common than screening for FI when deemed clinically relevant.³²  As one-fourth of recently hospitalized children experience household FI and around 40% of pediatric caregivers experienced inpatient FI, transitioning from a conditional to a universal screening approach may present an opportunity for practice change to address unmet needs.^11,33 

We found that FI screening was conducted in multiple clinical settings, with the majority completed on inpatient units. Other studies have similarly found that screening for social stressors occurs in multiple clinical settings such as inpatient units, emergency departments, and primary care clinics.^32,34–36  Within these various clinical settings, we found that social workers and nurses accounted for the majority of individuals who administered screeners, whereas physicians made up a minority of screening. In 1 analysis of 4 large academic United States children’s hospitals, it was found that nurses were the most common providers to screen for social stressors, although hospitalists also conducted many of the screens.³²  This variation in findings of physician-administration of screeners may be related to survey methodology (our respondents did not include physicians), the breadth of institutions included, and/or the type of screeners administered.

Respondents identified that an oral interview was the most common screening method used, whereas few reported administering screeners with paper or electronic devices. Most hospitals used their own institutional screener, with only a small proportion using the Hunger Vital Sign, a validated tool.³⁷  Screened individuals have been found to be more likely to disclose health related social needs when using electronic tablets or written questionnaires, compared to oral interview.^38–40  Hospitals could consider refining screening practices by selecting validated, standardized screening tools and implementing evidence-based modalities of administration.

When there was a positive screen for FI, study participants shared that some hospitals provided caregivers with food or meals. Other hospitals provided food or meals to caregivers without specific FI screening but had conditional qualifications such as need-base and/or presence of a breastfeeding mother. These findings are similar to a 2007 study of United States and Canadian children’s hospitals that found that 39% of institutions provided some caregivers with food or meals and 14% provided provisions without conditional qualifications.¹⁸  These collective findings indicate that hospitals may meet the needs of some caregivers experiencing inpatient FI, but, because of conditional qualifications, others may be missed. Past research has shown that the caregivers of admitted pediatric patients may not only experience exacerbations of household FI, but also encounter isolated inpatient FI.^11–12  Our results present a potential opportunity for institutions to target food or meal provisions to those who are in need but may be missed currently, allowing caregivers to fully engage in their child’s care.

There were several limitations to this study. Most hospitals (91%) had 1 representative respondent who may not have had a complete knowledge of all screening protocols and interventions. Our study workflow did attempt to identify the key stakeholder with knowledge about food insecurity screening and interventions at each institution. Of note, the most frequent response provided to identified FI was “social work consult.” Social workers, who made up the majority of survey respondents, thus may represent a key stakeholder group to evaluate institutional screening and food provision practices. Respondents may have been affected by social desirability bias, although an online survey platform offered an opportunity for confidential responses that may have helped to mitigate bias. We had a 40% response rate; thus, the reported results may reflect a response bias with participants being more likely to respond if their institutions had practices for screening for FI and providing food. It is also possible that institutions with fewer resources, specifically without dedicated social workers, may have been more represented in the nonrespondents, and thus, we may be underrepresenting available services especially in areas with limited resources. We were not able to connect with an initial contact at 30% of hospitals; thus, our results may not represent all institutional practices. Although components of the survey were not validated, we field tested and refined the instrument with content and survey experts. Despite the discussed limitations, our results still highlight areas where many hospitals can make improvements to FI screening practices and interventions to identified needs.

In this survey of inpatient United States children’s hospital representative, we found that the majority of respondents reported at least some institutional inpatient screening for FI. A lack of universal screening may lead to missed opportunities to address needs. Institutions had variable screening practices and interventions to address both inpatient and household FI, and there are opportunities to improve screening protocols and interventions to ensure that families in need receive support, especially during and around hospital admission.

We thank Drs Eugene Shapiro, Magna Dias, Mona Sharifi, Adam Berkwitt, Melissa Langhan, and Marc Auerbach and for their feedback on our survey. We would like to thank Drs Kristin Reese and Nishant Pandya for help with data collection. We would like to thank Denine Baxter MHA, BSN, RN, CNML and Drs Jaspreet Loyal and Magna Dias for working to improve food provisions for families admitted to our children’s hospital. Finally, we would like to thank the parents and caregivers who inspired this work.