Using malnutrition and food insecurity screening to identify broader health-related social needs amongst older adults receiving emergency department care in the Southeastern United States: A cross-sectional study

Abstract

Unmet health-related social needs are common amongst older US adults and impact both quality of life and health outcomes. One of the ways that unmet health-related social needs impact health is through malnutrition, an imbalance in a person’s intake of energy and/or nutrients. Lack of reliable access to a sufficient quantity of nutritious food is a specific health-related social need that can be assessed rapidly and, when unmet, is a direct risk factor for malnutrition and may be indicative of a broader range of unmet health-related social needs. We conducted a cross-sectional study to characterise malnutrition and food insecurity amongst older adults receiving emergency department (ED) care using brief, validated measures and to assess the burden of a broader range of health-related social needs amongst these patients. Patients were asked about their need for and willingness to receive a range of social services. The study was conducted in an academic ED serving a racially and socioeconomically diverse population in the Southeastern United States. A convenience sample of noncritically ill adults aged 60 years and older was approached between November 2018 and April 2019. Study patients (n = 127) were predominantly non-Hispanic white (67%), community dwelling (91%) and urban residents (66%) with 28% screening positive for malnutrition risk, 16% for food insecurity and 5% for both. Of those at risk for malnutrition, 25 (69%) reported ≥2 unmet health-related social needs and 14 (38%) were receptive to social services. Amongst food insecure patients, 18 (90%) reported additional unmet health-related social needs and 13 (65%) were receptive to receiving social services. In conclusion, a brief set of questions can identify subgroups of older ED patients who are food insecure or at risk for malnutrition. Individuals who screen positive for food insecurity have a high burden of unmet health-related social needs.

1 |. INTRODUCTION

Malnutrition and food insecurity are common problems amongst older adults and increasingly recognised as having important secondary effects on health. As many as one in three older adults is at risk for malnutrition (Kaiser et al., 2010), a disorder resulting from inadequate consumption of health nutrients. Vulnerable subpopulations including hospitalised elderly patients and older adults in rehabilitation facilities experience much higher rates of malnutrition, up to 38.7% and 50%, respectively than the 6% estimated in the community dwelling older adults (Kaiser et al., 2010). In addition to adding an estimated $51 billion (Burks et al., 2017) of disease-attributable economic costs, malnutrition amongst older adults contributes to poor health, lower quality of life and premature death (Agarwal et al., 2013; Ferguson et al., 1998; Pereira et al., 2015; Saka et al., 2010; Spaccavento et al., 2009). Clinical risk factors for malnutrition include chronic medical conditions, frailty, polypharmacy, depression, poor oral health and impaired swallowing (Burks et al., 2017; Thomas et al., 2016). Several unmet health-related social needs (HRSN), most obviously food insecurity, defined as lacking reliable access to affordable, nutritious food, but also social isolation and lack of transportation, act synergistically with these clinical risk factors to exacerbate the risk and severity of malnutrition amongst older adults. Food insecurity often coincides with other unmet social needs as well as medical problems such as diabetes, heart disease and depression. (Berkowitz et al., 2017) Food insecurity can also lead to poor health outcomes via poor diet quality (Berkowitz, Basu, et al., 2018) and through associated adverse health behaviours like medication non-adherence (Bengle et al., 2010).

Emergency departments (EDs) are an important site of care for older adults, who make over 20 million ED visits annually (Pines et al., 2013). Older adults with limited access to a primary care provider are particularly dependent on EDs for care (Cheung et al., 2012; Pines et al., 2013; Rust et al., 2008). Factors that limit access to a primary care provider, such as lack of transportation and functional decline, may also contribute to malnutrition and food insecurity. Consistent with this, malnutrition affects an estimated 16% of older adults who present to the ED compared to 6% of all US older adults (Pereira et al., 2015). One third of ED patients who are malnourished reports food insecurity (Burks et al., 2017).

Whilst malnutrition is a treatable condition, the large number of risk factors suggests a need for multifaceted and personalised interventions. For example, clinical approaches, such as registered dietician consultations or dental care, may address clinical contributors to malnutrition risk whilst community-based social services may be enlisted to address food insecurity or other HRSN (Brewster et al., 2019). Cross-sectoral partnerships between health care systems that can identify malnutrition and food insecurity and community-based organisations (CBOs) that can address HRSN are likely an essential component of a comprehensive response to malnutrition and food insecurity in older adults (Brewster et al., 2018). In the United States, medical care spending represents a higher proportion of gross domestic product than in other developed nations, but proportional spending on community services is relatively low (Berchick et al., 2019) (Garber & Skinner, 2008). However, the ongoing transition in the United States from fee-for-service care, in which providers are reimbursed for each medical service they provide, to capitated or value-based care models for healthcare delivery, in which a provider or health system is paid per patient year or paid and penalised based on patient outcomes, has the potential to substantially shift the priorities of healthcare payers and providers towards addressing these HRSN, because in many cases the value of doing so exceeds the value of direct medical care (National Academies of Sciences E. & Medicine., 2019). Of particular relevance to food insecurity are the Home Health Value-based Purchasing Model and the Hospital Readmission Reduction Programme, both of which seek to improve health outcomes for older adults with chronic health conditions (Centers for Medicare & Medicaid Services, 2020, Centers for Medicare and Medicaid Services, 2021). However, a better understanding of the extent to which malnourished and food insecure older ED patients need and are receptive to services from CBOs is unknown. The primary question for this study is as follows: Amongst older adults receiving emergency care who are either food insecure or malnourished, what is the burden of other HRSN that are addressable through community-based services? The goal of this study is to answer this question by identifying and characterising patients at risk for malnutrition and/or food insecurity, assessing receptivity to receiving community-based services and to assessing the HRSN of older adults visiting the ED. The results of this study will inform the implementation of a programme to address food insecurity amongst older adults seeking ED care.

2 |. METHODS

2.1 |. Study design

We conducted a cross-sectional observational study in an academic ED serving a racially and socioeconomically diverse population of older adults. The study was conducted as part of the preparation for the feasibility assessment for an ED-based food insecurity intervention.

2.2 |. Study setting and participants

The study was conducted in an academic ED in the Southeastern United States serving a socioeconomically diverse population. The ED has approximately 65,000 visits annually, 18% of which are by adults aged 60 years and older (older adults). A convenience sample of patients was recruited between 9:00 a.m. and 5:00 p.m. on weekdays between November 2018 and April 2019. All patients aged ≥60 years who presented to the ED during these times were screened for enrollment. Potential patients were identified using the ED track board feature of the electronic health record. Patients were deemed ineligible if they were on a psychiatric hold, non-English speaking, prisoners or critically ill as defined by an Emergency Severity Index (ESI) of 1. Patients were also excluded if they had already been flagged for admission or discharge, were not in the room after at least two visits to the room, were actively receiving treatment or sleeping when a research assistant (RA) visited the room. Eligible patients who provided verbal consent to participate in the study were included. Cognitive impairment was assessed with the Abbreviated Mental Test (AMT-4) (Jitapunkul et al., 1991). If a patient’s cognitive impairment was such that they were judged by the RA to be unable to provide useful information, caregivers or family members accompanying the patient in the ED were offered the opportunity to consent on behalf of the patient and complete the interview on behalf of the patient. All patients who screened positive for either food insecurity or malnutrition were given a handout with contact information for local community-based services. The study was reviewed and approved by the hospital’s Institutional Review Board.

2.3 |. Data collection

RAs reviewed the ED track board daily to identify potential participants. RAs approached all eligible patients and obtained assent to describe the study. After obtaining verbal consent, the RA conducted the interview using a structured data collection instrument with a scripted introduction and questions. Data were collected on paper forms which were entered into a secure web-based database (REDCap) that meets or exceeds NIH security requirements. REDCap data were reviewed for completeness and accuracy on a periodic basis by the study coordinator and investigator.

2.4 |. Measures

Patient demographic characteristics were assessed using standard measures. Rural residence was defined by ZIP code, using population classifications from USDA Rural-Urban Continuum Codes (USDA Economic Research Services (2019)). To identify patients with malnutrition risk and food insecurity, we used validated screening tools that are recommended for use in health care settings.

2.4.1 |. Malnutrition

The Malnutrition Screening Tool (MST) is a three-item questionnaire that identifies individuals at risk for malnutrition and in need of nutrition support and interventions. Patients are asked the following questions: “Have you recently lost weight without trying?,” “If yes, how much weight have you lost?” and “Have you been eating poorly because of a decreased appetite?” (Ferguson et al., 1999). Consistent with the recommended scoring interpretation, malnutrition risk was determined by a score of 2 or greater on the MST.

2.4.2 |. Food insecurity

We assessed food insecurity with the Hunger Vital Sign (HVS) questionnaire, which uses level of agreement with two statements about perceived possibility and experience of food scarcity: “We worried whether our food would run out before we got money to buy more” and “The food we bought just did not last and we did not have money to get more.”(Hager et al., 2010) A response of “sometimes” or “often” to either question is considered indicative of food insecurity.

HRSN

Health-related social needs were assessed to characterise the prevalence of needs that may intersect with food insecurity. HRSN questions were selected from validated instruments in consultation with experts in geriatric care. Five domains were examined: transportation (Burks et al., 2017), social isolation (Russell et al., 1980), financial strain (Goldberg & Mawn, 2015), housing stability and difficulty paying for medication (Sattler et al., 2014). Each of these domains has been discussed in the literature as risk factors for food insecurity and/or malnutrition. Additionally, each of these domains can, to some degree, be addressed using community-based services. Transportation was assessed using a single question with a 5-point Likert scale. Social isolation and financial strain and need for financial support were assessed using multiple questions. Housing stability and difficulty paying for medication were assessed with single yes/no response questions. The social needs assessment also included questions related to ability to perform instrumental activities of daily living (IADLs) and activities of daily living (ADLs) (Lawton & Brody, 1969). Specifically, we asked about the need for assistance with grocery shopping, preparing meals and using the telephone. For all the HRSN questions, a positive response was defined by yes (housing stability and difficulty paying for medication) or by sometimes, often or almost always (transportation, social isolation, financial support and IADLs/ADLs).

2.4.3 |. Receptivity to services

Patient receptivity to receive social services was first assessed by asking patients about a hypothetical question regarding their willingness to speak with a social worker or care navigator by phone following the ED visit about being connected to community resources. Patients who were willing to speak to a social worker or care navigator were then asked about their desire to be connected with specific examples of community-based resources: home-delivered meals, congregate meals, Supplemental Nutritional-Assistance Programme (SNAP; i.e., Food Stamp Programme), transportation assistance, homemaker services, home health services, utility assistance and insurance and benefits assistance. These programmes were selected based on a review of the services typically provided by local Area Agency on Ageing (AAA) programmes (National Association of Area Agencies on Aging, 2017).

2.5 |. Data analysis

The study population was stratified into subgroups of patients by malnutrition risk and food insecurity, based on their scores on the MST and HVS. Descriptive statistics were used to describe prevalence of these two conditions, patient characteristics, self-reported HRSN and willingness to receive services. To evaluate inter-rater reliability of the survey administration, 16 surveys were performed with a second assessor concomitantly filling out the form throughout the interview. For these cases, each assessor was blinded to the other’s collected responses. Percent agreement and kappa were calculated. Assuming a prevalence of food insecurity of 10%, a sample size of 120 patients would be sufficient to provide lower and upper bounds of the 95% confidence interval for the prevalence of malnutrition and food insecurity within 5% of the point estimate.

3 |. RESULTS

Of 511 ED patients aged 60 years and older approached during the period of enrollment (November 2018 to April 2019), 203 met inclusion criteria. The most common reasons for ineligibility were that the patient was not present in room (n = 90) or was with a provider (n = 81; Figure 1). Of the 203 patients who were eligible for participation, 76 were excluded because they declined participation or were unable to communicate. Of the 127 consenting participants, 51% were male, 68% White and 25% Black. The mean age was 68 years old, and about one third of patients (32%) were older than 75. The majority of study participants (66%) lived in urban ZIP codes, and 91% of participants were community dwelling in a private residence (Table 1).

FIGURE 1.

Flow diagram of enrollment process

TABLE 1.

Characteristics of study patients

	Overall n (%)		At-risk for malnutrition (MST+) n (%)a		Food insecure (HVS+) n (%)b		Neither HVS + nor MST + n (%)
All patients	127		36c		20c		77
Age, years
60–69	53	42%	15	42%	14	70%	29	38%
70–79	55	43%	12	33%	5	25%	39	51%
>80	19	15%	9	25%	1	5%	9	12%
Sex
Female	62	49%	17	47%	16	80%	34	44%
Male	65	51%	19	53%	4	20%	43	56%
Race
White	86	68%	21	58%	10	50%	58	75%
Black	32	25%	13	36%	9	45%	12	16%
Other	8	6%	2	6%	1	5%	6	8%
Ethnicityd
Hispanic	4	3%	2	6%	1	5%	2	3%
Non-Hispanic	121	95%	34	94%	19	95%	73	44%
Education
Not college graduate	88	69%	27	75%	19	95%	48	62%
College graduate	39	31%	9	25%	1	5%	29	38%
Employment statuse
Employed	17	13%	1	3%	0	0%	16	21%
Unemployed	3	2%	2	6%	1	5%	1	1%
Retired	92	72%	30	83%	12	60%	53	69%
On disability	14	11%	3	8%	7	35%	6	8%
Insurance type
Medicare	36	28%	11	31%	9	45%	17	22%
Medicare Plusf	60	47%	14	39%	5	25%	43	56%
Medicaidg	11	9%	6	17%	3	15%	4	5%
Private/other	16	13%	3	8%	2	10%	11	14%
Uninsured	4	3%	2	6%	1	5%	2	3%
Living arrangement
Private residence	115	91%	29	81%	18	90%	74	96%
Assisted living	11	9%	6	17%	2	10%	3	4%
Population densitye,h
Urban	84	66%	21	58%	11	55%	54	70%
Rural	42	33%	15	42%	9	45%	22	29%
Value-based carei
ACO	20	16%	4	11%	1	5%	15	19%
Non-ACO	104	82%	30	83%	19	95%	61	79%

Abbreviations: ACO, Accountable Care Organisation; HVS, Hunger Vital Sign; MST, Malnutrition Screening Tool.

^aMeasured by positive screen on malnutrition screening tool.

^bMeasured by positive screen on hunger vital sign.

^cSix patients were both insecure and malnourished. Data from these patients are included in both the food insecure and malnourished columns.

^dMissing = 2.

^eMissing = 1.

^fMedicare enrollment supplemented by a Medicare Advantage Plan or other supplemental private insurance.

^gIncludes Medicare and Medicaid dual eligible participants (n = 8) and Medicaid only participants (n = 3).

^hPopulation density measured by USDA Rural-Urban Continuum Codes.

ⁱParticipant in next-generation Accountable Care Organisation (ACO) plan. Missing = 3.

For the 16 cases with two blinded assessors, percentage agreement was 93.7% for the measure of malnutrition risk (MST), with a kappa 0.79. Percentage agreement was 100% for the measure of food insecurity (HVS), with a kappa of 1.0.

3.1 |. Prevalence of malnutrition risk and food insecurity

Of the 127 study participants, 50 (39%) were either at risk for malnutrition (28%; n = 36) or food insecure (16%; n = 20). Six of these 50 patients were both at risk for malnutrition and food insecure and are included in the counts of both categories. Malnutrition risk was more prevalent amongst Black patients compared to other races, those with less than a college degree compared to those with a college degree, and Medicaid patients compared to other insured patients. Food insecurity was more common amongst patients who were younger (age <70 years), female, Black, receiving social security disability insurance (SSDI) and/or patients with less than a college degree compared to other subgroups within each of these categories. Rural patients were more likely to be at risk for malnutrition (35% vs. 25%) and more likely to be food insecure than those living in urban areas (21% vs. 13%). Nearly all patients surveyed had insurance (97%). Both malnutrition risk and food insecurity were more prevalent amongst patients enrolled only in traditional Medicare than those with Medicare Advantage or supplemental insurance (31% vs. 23% and 25% vs. 8%, respectively). Of the 20 patients who were members of a Next-Generation Accountable Care Organisation (ACO), 20% (n = 4) screened positive for malnutrition risk and 5% were food insecure. Functional limitations regarding the ability to shop for food or prepare meals were reported by about one third of patients with malnutrition risk or food insecurity. The prevalence of these limitations was similar or slightly higher in patients who screened negative for both malnutrition risk and food insecurity (Table 2).

TABLE 2.

Ability to perform selected independent activities of daily living among study patients

	Overall (n = 127)	At-risk for malnutrition (MST+) n = 36 a,b		Food insecure (HVS+) n = 20b,c		Neither HVS + nor MST+ n = 77
Sometimes, often or almost always needs assistance
Grocery shopping	35	12/35	34%	11/35	31%	12/35	34%
Preparing meals	34	13/34	38%	9/34	26%	14/34	41%
Using the telephone or looking up numbers	14	5/14	35%	8/14	57%	2/14	14%

Abbreviations: HVS, Hunger Vital Sign; MST, Malnutrition Screening Tool.

^aMeasured by positive screen on malnutrition screening tool.

^bCategories are not mutually exclusive.

^cMeasured by positive screen on Hunger Vital Sign.

4 |. HRSN

A majority (74%) of patients who either screened positive for risk of malnutrition or food insecurity reported at least one of the five additional HRSN: social support, transportation, housing stability, financial support and medication purchasing (Table 3). The need for financial and social support was the most frequently reported social needs both amongst those at risk for malnutrition (27% and 19%, respectively) and amongst those who were food insecure (85% and 50%, respectively). The presence of one or more HRSN was more common amongst those with food insecurity (90%) than in those at risk for malnutrition (69%). In addition, food insecure patients were more likely to report having to choose between paying for medication or food than those at risk for malnutrition (30% vs. 14%).

TABLE 3.

Social needs and receptivity to health-related social services among study patients

	At-risk for malnutrition (MST+) n = 36a,b		Food insecure (HVS+) n = 20b,c		Neither HVS + nor MST+ n = 77
Receptive to services
Yes	14	38%	13	65%	23	30%
Past attempts to access social services
Yes	12	33%	14	70%	9	12%
Health-related social needb
Social supportd	7	19%	10	50%	3	4%
Transportation	2	6%	6	30%	2	3%
Housing stability	2	6%	3	15%	0	0%
Financial support	10	28%	17	85%	12	16%
Medication purchasing	5	14%	6	30%	3	4%
Any of the above	25	69%	18	90%	15	19%
Services desiredb
Meals on wheels	4	11%	7	35%	2	2%
Congregate meals	0	0%	0	0%	0	0%
SNAP	5	14%	5	25%	1	1%
Transportation	1	3%	4	20%	4	5%
Home care	3	0%	2	10%	3	4%
Home health	2	8%	1	5%	3	4%
Utility assistance	2	5%	2	10%	2	2%
Insurance assistance	1	3%	2	10%	5	6%
Any of the above	8	22%	13	65%	15	19%

Abbreviations: HVS, Hunger Vital Sign; MST, Malnutrition Screening Tool.

^aMeasured by positive screen on Malnutrition Screening Tool.

^bCategories are not mutually exclusive.

^cMeasured by positive screen on Hunger Vital Sign.

^dRussell DW. UCLA Loneliness Scale (Version 3): Reliability, validity and factor structure. Journal of personality assessment.1996 Feb 1;66(1):20–40.

4.1 |. Receptivity to services

Patients who did not screen positive for either malnutrition risk or food insecurity had a substantially lower frequency of reporting a desire for health-related social services. Amongst patients with malnutrition risk, 38% were receptive to being contacted at home regarding health-related social services and 22% reported a desire for one or more specific services (Table 3). Amongst patients who were food insecure, 65% were receptive to being contacted at home regarding health-related social services and 65% reported a desire for a specific type of service. Amongst patients with malnutrition risk, the specific services that were most frequently desired were SNAP, Meals on Wheels and home care. Amongst food insecure patients, the most frequently desired services were Meals on Wheels, SNAP and transportation services.

5 |. DISCUSSION

Malnutrition risk and food insecurity were common amongst older adults in this study setting: in this single-centre prospective study of older adults receiving ED care, 28% of patients screened positive for malnutrition risk and 16% screened positive for food insecurity, with 5% positive for both. The majority of patients in both groups were receptive to being contacted at home following the ED visit to discuss additional health-related social service needs that could be compounding their malnutrition risk or food insecurity. These findings support the feasibility of using a short screener in the ED to identify patients with a high need for social services. Malnutrition risk and food insecurity were more common amongst females, minorities and individuals receiving SSDI; those with limited higher education and those living in rural areas. Additional social service needs were associated with similar demographic and socioeconomic characteristics: women, minorities, individuals receiving SSDI and those with less formal education reported more HRSN.

Social services can help address factors contributing to malnutrition and food insecurity. Financial strain and lack of social support were predominant concerns amongst food insecure patients and those at risk for both. In this study, 30% of food insecure patients reported having had to choose between paying for medication and paying for food, a prevalent issue amongst food insecure adults that contributes to poor health outcomes (Knight et al., 2016; Silverman et al., 2015). Participation in programmes such as SNAP has been shown to increase senior household financial security by providing income specifically for food. Similarly, home-delivered meal programmes may offer considerable benefits to individuals who are home-bound; participation in these programmes can improve older adults’ well-being by improving their nutritional status and reducing isolation (Thomas et al., 2016). Home-delivered meal programmes have also been shown to results in net savings for healthcare systems (Thomas & Mor, 2013). Our results suggest that many older ED patients require social services to help them age in place, given 74% of malnourished or food insecure patients reported one or more additional social need.

Findings from this study suggest that an ED-based screening and referral intervention amongst older adults with food insecurity and/or malnutrition risk is acceptable to this population. Patients who were food insecure and patients who were both food insecure and at risk for malnutrition were more likely to report an interest in social services, particularly nutrition programmes, whilst patients who screened negative for both tools had a lower level of social need and a correspondingly low interest in social services overall.

These findings lay the groundwork for an intervention that will test an ED-based screening and referral intervention to identify food insecure older adults who are at risk for malnutrition and link them to community resources. The results of the study will inform the development of the referral pathway and community partners involved. Specifically, in this sample, patients who screened positive for food insecurity had a higher burden of HRSN than those with malnutrition risk, and we anticipate using food insecurity as the primary screening method for the implementation effort. The optimal community partner will be one who can provide a range of services to meet patient needs, such as an AAA. With federally funded ageing networks found in every state, AAAs can link patients to an array of programmes. This type of partnership may benefit individual patients, who gain streamlined access to necessary social services, and may improve health at the community level. There is evidence that well-networked AAAs with a range of health care and community partnerships can lead to lower hospital admissions and nursing home use (Brewster et al., 2018).

As US health systems transition to value-based care, defined as models of care in which hospitals and providers are compensated based on health outcomes, understanding how to meet the needs of patients with high levels of social need, such as those at risk for malnutrition and those who are food insecure, is essential. This transition is ongoing in the United States and is manifesting through Accountable Care Organisations, policy changes in the private health insurance sector, specific value-based payment mechanisms specified by the Centres for Medicare and Medicaid Services and Medicare Advantages plan construction. In contrast to traditional Medicare plans that pay providers a fee for each health service provided, Medicare Advantage insurers typically receive a monthly fee per enrollee—creating a capitated system in which the Medicare Advantage plan is incentivised to identify interventions and services that optimise patient health. Prior work has shown that programmes such as medically tailored home-delivered meals are associated with reduced health care expenditures amongst patients discharged from the hospital (Berkowitz, Terranova, et al., 2018). Implementing a screening and referral process to identify and address food insecurity and malnutrition could connect older adults to needed support from community-based services and may reduce their overall health care costs. Partnering with an AAA or another well-networked community entity that can serve as a resource hub may streamline the process of identify and addressing HRSN for this population (Brewster et al., 2019). This type of partnership has the potential to be successful and sustainable in both the fragmented US health care system as well as in more centralised or single-payer health systems (Loopstra, 2020).

Awareness about the need to address social determinants of health for individual patients has increased rapidly over the past 5 years, and the COVID-19 pandemic has brought issues of poverty, hunger and isolation to the fore in the United States, United Kingdom, and EU nations alike (Connors et al., 2020) (Barker & Russell, 2020). The work presented here provides initial data for the feasibility of one approach to addressing social needs (i.e., screening for food insecurity in the ED with subsequent referral to CBOs). In subsequent work we hope to implement, evaluate and scale this approach. As pointed out by Silverstein et al., optimal solutions to some problems of social determinants of health will require both individual level solutions, as well as policy or public health solutions (Silverstein et al., 2019). For example, changes in tax structure, funding of food assistance programmes, provision of public transportation and protections for older adults from financial elder abuse all have the potential to influence the burden of malnutrition and food insecurity in a community. However, as it stands, most senior nutrition programmes in the United States are underfunded, resulting in lengthy waitlists for services and whilst recent changes to Medicare reimbursement regulations may allow health plans to cover home-delivered meals and other services, uptake of this option has been slow (Meals on Wheels America, 2019) (Meyers et al., 2020).

5.1 |. Limitations

The current study did not examine the impact of malnutrition and food insecurity on health outcomes. However, it is known from other work that both malnutrition and food insecurity have significant effects on patient health (Berkowitz, Basu, et al., 2018; Berkowitz et al., 2019). Amongst older adults, malnourished individuals are at an increased risk of falls, poor wound healing and premature death (Ferguson et al., 1998; Pereira et al., 2015; Spaccavento et al., 2009). Food insecurity is associated with higher rates of depression, diabetes and heart disease, and older adults who are food insecure tend to have worse overall health outcomes (Pooler et al., 2019; Thomas et al., 2016). For patients who receive care in the ED, identifying malnutrition risk and food insecurity is the first step to intervening on these conditions. Further research is needed to evaluate the effect of identifying food insecurity and malnutrition and making connections to social services on health outcomes and utilisation. A second phase of this study will evaluate both the feasibility of running an ED-based screening and referral programme and the effect of receiving community services on patient health, well-being and ED utilisation.

Patient responses to questions about receptivity to social services provide an approximation of willingness to receive these services. However, responses may be subject to social desirability bias, which could lead patients to either deny a need for service when a need is present or express a willingness to receive a service when willingness is not actually present. Actual rates of patient acceptance of services may be different than reported rates. The receipt of such services, particularly food assistance, by individuals within the study sample, may have affected the responses to the screener questions. We did not ask about receptivity of patients to receiving cash transfers. Available data from low- and middle-income countries suggest that cash transfers are an effective means of reducing food insecurity and may be an acceptable and effective intervention for some food insecure older adults (Owusu-Addo et al., 2018). The sample was comprised of English-speaking patients only, due to research staff limitations, cultural considerations and a small population of non-English speakers; we do not know about nutrition status and needs of non-English speakers nor about what resources they would want to receive. Further, the study was conducted in a single, academic ED in the Southeastern United States. In the state where the study was conducted, 8.5% of older adults are below the official poverty threshold and 13.1% are below the supplemental poverty threshold; these values are slightly below the national averages of 9.3% and 14.5%, respectively (Cubanski et al., 2018). Accordingly, food insecurity, malnutrition risk and the need for health-related social services will differ in other settings and regions.

6 |. CONCLUSION

Amongst older adults receiving ED care, brief screening tools for malnutrition risk and food insecurity can identify subgroups of patients who are in need of and willing to receive health-related social services. Food insecure patients had particularly high rates of HRSN, with over 90% reporting at least one such need, suggesting that screening for food insecurity may be an efficient method of identifying patients with a need for referral to social services. Additional research is needed to assess the feasibility of an ED-based screening and referral programme for these patients and assess the impact of these referrals on subsequent health outcomes and health care utilisation.

7 |. ETHICAL STANDARDS DISCLOSURE

This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the Institutional Review Board at University of North Carolina-Chapel Hill. (IRB #18–1559). Verbal informed consent was obtained from all patients. Verbal consent was witnessed and formally recorded.

What is known about this topic:

• Food insecurity is a significant risk factor for malnutrition in older adults.

• Food insecurity has its own aetiology including lack of transportation, social isolation, and financial strain.

• Connections with community-based resources can reduce the related problems of food insecurity and malnutrition risk.

What this paper adds:

• Both malnutrition risk and food insecurity were common amongst females, minorities, and disability insurance recipients, those without a college degree, and those dwelling in rural areas.

• Brief screens for malnutrition and food insecurity can identify patients with a range of unmet social needs, with food insecure patients having a particularly high burden of unmet needs.

• Patients who are malnourished and food insecure are willing to receive community services.

• Understanding the prevalence and type of needs amongst older adults in a specific health care setting is a key first step before implementing new programmes to address social needs.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.