The Prevalence and Impact on Patient Outcomes of Malnutrition Among Older Adults Presenting at an Irish Emergency Department: A Secondary Analysis of the OPTIMEND Trial

Malnutrition is common among older adults and is associated with adverse outcomes but remains undiagnosed on healthcare admissions. Older adults use emergency departments (EDs) more than any other age group. This study aimed to determine the prevalence and factors associated with malnutrition on admission and with adverse outcomes post-admission among older adults attending an Irish ED. Secondary analysis of data collected from a randomised trial exploring the impact of a dedicated team of health and social care professionals on the care of older adults in the ED. Nutritional status was determined using the Mini Nutritional Assessment- short form. Patient parameters and outcomes included health related quality of life, functional ability, frailty, hospital admissions, falls history and clinical outcomes at index visit, 30-day and 6-month follow up. Aggregate anonymised participant data linked from baseline to 30-days and 6-month follow-up were used for statistical analysis.

In 2018, the number of people worldwide who were over the age of 65 years outnumbered children under the age of ve years for the rst time in history. It is predicted that this cohort will more than double (from 9-16%) by 2050 (1). In 2016, 19% of the population of Ireland was 65 years or older (2). This is projected to reach 29% by 2041 with the greatest growth in the number of people aged 85 years or older (3). This will result in an increase of older adults at risk of malnutrition due to the effects of ageing and the accumulation of diseases and impairments (4) that impact on the ability to maintain adequate energy and protein via food intake. The prevalence of malnutrition among older adults increases with deteriorating functional and health status. Reported prevalence rates greatly depend on the de nition used, but are generally below 10% in independently living older persons and rates increase up to two thirds in older people in acute care and rehabilitation hospitals (4)(5)(6).
Protein-energy malnutrition, often referred to simply as malnutrition, is a condition resulting from inadequate intake or an inability to absorb and/or digest adequate energy and/or protein (7). Nutritional problems are commonplace in acute and chronic illness. A recent systematic review of prospective studies identi ed thirty potentially modi able determinants of malnutrition among older adults across seven domains (oral, psychosocial, medication and care, health, physical function, lifestyle, and eating) (8). Age-related changes in appetite regulation can affect food intake which may lead to malnutrition if additional risk factors such as physical, cognitive, environmental, economic and/or social problems also occur (9,10).
A reduced dietary intake in combination with underlying in ammatory processes rapidly leads to malnutrition (11). An older adult who has reduced physical reserves is nutritionally vulnerable and limited in their resilience and ability to recover from acute health threats or stressors (10). A strong association between malnutrition and adverse health outcomes among older adults is well documented including increased morbidity and mortality (6,12,13). Nutritional vulnerability contributes to more medical complications, longer hospital stays, increased likelihood of nursing home admission and poorer quality of life (10,11,14). Total costs associated with malnutrition among institutionalised and communitydwelling older adults are reported as considerably higher than those among well-nourished older adults, predominately due to higher use of health care resources -GP consultations, hospitalisations, health care monitoring, and treatments (15).
The increasing ageing population and higher number of individuals with multimorbidity are some of the main demographic drivers of incremental increases in Emergency Department (ED) attendances (16,17).
A high incidence of ED visits by older adults, with over a three-fold difference among those aged 80 years or over compared to that of those aged less than 80 years has been reported (18,19). There is limited information about the prevalence of malnutrition among older adults presenting at EDs. Those that exist show a prevalence of 15-29% (20-23) associated with an increase in short-term mortality (21,24).
Serious signs of malnutrition are the presence of either an unintended loss of body weight (> 5% in three months or > 10% inde nite of time) or a markedly reduced body mass index (i.e. BMI < 20 kg/m 2 or < 22 kg/m 2 if younger or older than 70 years, respectively) or low fat free mass index (i.e. <15 and < untreated (9,20,(26)(27)(28). Clinical staff can miss malnutrition in older adults due to barriers that affect nutritional screening including a lack of routine nutritional screening in many hospitals (9,29), and a belief by clinical staff that individual judgement of nutritional status is more superior (30,31). Older adults who are nutritionally at risk often do not receive nutritional support or counselling (6,9,20,26,28). Interdisciplinary care in the ED has been shown to enhance decision-making and contribute to timely and safe patient care, particularly for older adults (32). Malnutrition screening in the ED, therefore can capture a nutritionally vulnerable population that could otherwise be overlooked as not all individuals who attend the ED are admitted to hospital, where screening for malnutrition more usually takes place (10,20).
Given the dearth of information on the prevalence of malnutrition among older adults presenting to EDs, the aims of this study are threefold: to determine: (i) the prevalence of malnutrition; (ii) factors associated with malnutrition; and (iii) the association of malnutrition with adverse outcomes post-discharge from ED in older adults.

Study design
This study represents secondary analysis of a single-centre randomised controlled trial, OPTI-MEND, that explored the impact of a dedicated health and social care professional (HSCP) team on the quality, safety, timeliness and cost-effectiveness of care of older adults in the ED. The protocol for the trial is published elsewhere (33). The OPTI-MEND study received ethical approval from the Health Service Executive (HSE) Mid-Western Regional Hospital Research Ethics Committee (ref. 103/18). Written informed consent has been obtained from all study participants.

Participants
Adults aged 65 years or older who presented to the ED at the University Hospital of Limerick (UHL, between December 2018 and May 2019 (inclusive), were considered eligible for the OPTI-MEND study.
Inclusion criteria required the capacity and willingness to provide written informed consent and presenting with any of the complaints presented in Table 1 as per the Manchester Triage System 2-5 (33,34). The OPTI-MEND study received ethical approval from the Health Service Executive (HSE) Mid-Western Regional Hospital Research Ethics Committee (ref. 103/18). Written informed consent has been obtained from all study participants. Ear and facial problems *The health and social care professional (HSCP) team will proactively treat these individuals without prior assessment by a physician **The HSCP team will await medical clearance prior to assessment and intervention.

Tools and Procedures
Participants were assessed by one or more trained members of the dedicated HSCP team (senior physiotherapist, senior occupational therapist and senior medical social worker). This included a holistic assessment of mobility, functional, cognitive and psychosocial abilities. Patient's baseline information was categorised as follows: Demographic -gender and age; Psychological/social factors -marital status and health-related quality of life by the EuroQoL's 5-level of the EQ-5D (EQ-5D-5 L) and a global self-assessment of health by the EuroQol-visual analogue scales (EQ-VAS) tool (35); Environmental/economic factors -residential status, mode of transport to the ED and source of referral; Physiological/biomedical factors -falls within the past 3 months; hospitalisation within the past 6 months; presenting complaint; functional ability, rated by the Barthel Index (36); frailty using the Identi cation of Seniors at Risk (ISAR) (37); and the Clinical frailty scale (38).
Malnutrition in older adults was measured with the short form of the Mini Nutritional Assessment (MNA®) by a member of the dedicated HSCP team. This is a valid nutritional screening tool recommended for use by the European Society for Clinical Nutrition and Metabolism (ESPEN) guideline on clinical nutrition and hydration in geriatrics (6,39). It takes into account physical and mental functional impairments that regularly contribute to the development of malnutrition and thus, considers an existing risk of malnutrition (6,9). Further, the revised MNA-Short Form (MNS-SF) allows for the use of either body mass index (BMI) or calf circumference. This allows it to be used with individuals who are immobile or in situations where weight and height cannot be measured (40). Previous research has established the criterion validity of this tool in all healthcare settings (community, rehabilitation, residential care, and hospital) among older people. The MNA-SF is comprised of six individual components which include documenting food intake, weight loss, mobility, stress or disease, neuropsychological problems and either Body Mass Index (BMI) or calf circumference (CC). It is scored out of a possible total of 14 points with the following categories: 0-7 points: 'Malnourished', [8][9][10][11] points: 'at risk of malnourishment' and 12-14 points: 'normal nutritional status'.

Outcomes measures
Adverse outcomes recorded from the index visit included ED outcome (admission, discharge home, transfer to other hospital), the hospital length of stay (measured in days), and patient's length of stay in the ED (measured in hours).
Follow-up at 30-day and 6-month outcomes included ED revisit, number of ED visits, hospital admissions, hospital length of stay, healthcare use, frequency of healthcare use, and nursing home admission. At 30days, functional decline was measured by the change in the Barthel index from baseline to 30-days (no change, reduced function, improved function), and decline in quality of life (QOL) was measured by the change in EQ-5D (no change, reduced QOL, improved QOL). At 6-months further decline in function and QOL was measured by the change from 30-days to 6-months.

Statistical analysis
Aggregate anonymised participant data linked from baseline to 30-days and 6-month follow-up were used for statistical analysis. Descriptive statistics of the patients presenting to the ED were conducted.
Categorical data were described by counts and percentages. Continuous data that approximated a normal distribution were described using means and standard deviations. Skewed data was described using medians and interquartile ranges. Differences between patient's demographic, psychological/social, environmental/economic and physiological/biomedical informationand the MNA tool were tested using Pearson's Chi-square test (or Fisher's exact test if appropriate) for categorical data.
For continuous data, differences were tested using the one-way ANOVA test or Kruskal Wallis tests where appropriate. Eta 2 was used to measure effect size for three or more groups, where 0.01, 0.06 and 0.14 represent a small, medium and large effect. Cramer's V was used to measure the size of the effect between categorical variables, with V = 0.1, 0.3 and 0.5 for a small, medium and large effect, respectively.
Hierarchical logistic regression models were used to further analyse associations between the assigned MNA categories and the observed decline in functional ability and quality of life at follow up. A 5% level of signi cance was used for all statistical tests. All statistical analysis was undertaken using SPSS Version 24.

Malnutrition Classi cation
A total of 353 patients participated with a mean patient age of 79.6 (SD = 7.0 years); 48.2% (n = 144) of patients were female. Using the MNA screening tool, 7.6% (n = 27) older adults attending the ED were classi ed as malnourished, 28.0% (n = 99) were classi ed as at risk of malnutrition and 64.3% (n = 227) had a normal nutritional status. Table 2 presents the component information of the MNA screening tool.  Differences in Patient Pro les by MNA Categories Table 3 shows the descriptive characteristics for the patients admitted to the ED and the factors associated with malnutrition in this study. Results suggest those who were classi ed as being malnourished had poorer QOL scores (EQ-5D median 15 (IQR 6.0) vs 12 (7)  2) and 17.2 (14.4) hours, p < 0.001) and were more likely to be discharged home from the ED (33.3% (n = 9) vs 40.4% (n = 40) and 20.3% (n = 46), p < 0.001) compared to older adults who were at risk of malnutrition or had normal nutritional status, respectively. MNA: mini nutritional assessment; HoLOS: hospital length of stay in days; PET: patient length of stay in ED in hours; EQ-5D total: quality of life score; EQ VAS: visual analogue scale score; ISAR score: identi cation of seniors at risk. Table 3 here Differences in 30-day follow-up outcomes by MNA categories Table 4 presents the outcomes measured at 30-day follow-up since the initial ED attendance and differences between MNA categories. In general, those who were malnourished compared to the older adults who were at risk of malnutrition or had normal nutrition at the initial ED attendance, were more likely to have reported a hospital admission (29.6% (n = 8) vs 12.1% (n = 12) and 10.6% (n = 24), p = 0.02), a nursing home admission (33.3% (n = 9) vs 24.4% (n = 24) and 8.4% (n = 19), p < 0.001), a reduced quality of life (40% (n = 10) vs 15.1% (n = 13) and 13.3% (n = 28), p = 0.02) and reduced functional ability (52% (n = 13) vs 36% (n = 31) and 24.8% (n = 52), p = 0.02), respectively.  Table 4 here While statistically signi cant differences were not found between 6-month outcomes and MNA categories, a reported further decline in functional ability was more likely among those older adults who were malnourished compared to those who were at risk of malnutrition or had normal nutritional status (56.5% (n = 13) vs 32.1% (n = 25) and 27.7% (n = 56), p = 0.05), respectively (Appendix 1).

Malnutrition status as a predictor of Quality of Life and Functional Ability
Hierarchical logistic regression models were used to further analyse associations between the assigned MNA categories and the observed decline in functional ability and quality of life at 30-day follow up.  OR: odds ratio; CI: con dence interval; MNA: mini nutritional assessment; EQ-5D total: quality of life score; EQ VAS: visual analogue scale score; ISAR score: identi cation of seniors at risk. Table 5 here Table 6 shows the models of associations between the assigned MNA categories and declining quality of  OR: odds ratio; CI: con dence interval; MNA: mini nutritional assessment; EQ VAS: visual analogue scale score; ISAR score: identi cation of seniors at risk. Table 6 here Discussion This is the rst Irish study to examine the prevalence of malnutrition in a cohort of older adults presenting to a large urban ED in Ireland, the factors associated with malnutrition and the association of malnutrition with 30 day and 6 month adverse patient outcomes post-discharge from the ED. Over one in three older adults presenting to this Irish ED were at risk of malnutrition or classed as malnourished.
These older adults were more likely to be frail and at risk of adverse outcomes, to have experienced a fall in the previous three months, had hospital admissions in the previous six months and reported a poorer functional ability and quality of life. In the short term, malnutrition was associated with a longer stay in the ED but a greater likelihood to be discharged home. At 30-day follow-up, being malnourished was associated with further decline in functional ability, quality of life, greater risk of hospital admissions and a greater likelihood of admission to nursing home facilities. Similar ndings were evident at 6 months.
There have been few published studies documenting the prevalence of malnutrition amongst hospitalised older adults in Ireland. An observational multi-site study reported the prevalence of malnutrition as 18% and at risk of malnourishment as 45% among older adults on admission to hospital (27). The higher prevalence of malnutrition on the hospital ward in this study from that observed in the ED setting could be attributed to further decline in nutritional status resulting from the hospital stay (9,10,13). This can be related to a number of factors including acute and chronic in ammation, in exible mealtimes, fasting, insu cient energy and protein content of meals and a lack of recognition of the increasing risk of malnutrition in the hospital setting (9,26). Among community-dwelling Irish older adults, incidence of malnutrition over two years has been documented to occur at a rate of 10.7% (41). Bardon and colleagues reported that the main predictors of incident malnutrition were being unmarried/divorced/separated, hospitalised in the year previous, and mobility limitations that indicate declining physical function (41). In each study, the authors conclude that the prioritisation of nutritional screening to identify and manage older adults vulnerable to malnutrition in clinical practice should be incorporated into public health policy (26,27,41).
Screening for malnutrition is an important step in recognising and identifying risk of or diagnosis of malnourishment (6). However, it is often not completed owing to perceived barriers of screening implementation in hospital settings including time, competence and resources (26,42). Most of the studies advocating for screening, do so at the ward level, thus potentially missing those who present via ED and are not admitted to the ward (20). We found that over half of the patients discharged from the ED had a risk of malnutrition or were malnourished. Health and social care professionals have both the skills and opportunity to assess older adults and to co-ordinate a care plan to minimise the risk of adverse events after discharge from the ED. The approach to managing nutritional risk needs to be multi-faceted and include the management of co-morbidities, the provision of home and social supports to encourage and facilitate food intake and the implementation of dietary modi cations to improve diet quality (43). When energy intake is insu cient to meet the demands of the body, either due to starvation, acute illness, or chronic disease/disability, then malnutrition becomes the driver that leads to further functional decline and an inability to recover from disease and a continuum of nutritional vulnerability (10).
Malnutrition affects quality of life through adverse outcomes, impaired cognitive and affective functioning and physical disability as people age (14). Assessment of quality of life is recognised as a clinically relevant outcome measure when evaluating new treatment strategies in patient populations, particularly older adults (14). We report a decline in quality of life associated with malnutrition at 30-day follow up, however no statistical differences were observed at the6 months follow-up. However, we noted a clinically signi cant decline in functional ability at 6-months follow up that was associated with being malnourished. Loss of functional independence results in loss of quality of life (44). Low muscle strength, a factor related to undernutrition, can compound the functional decline from chronic disease or conditions related to ageing (45).
Research to date demonstrates that risk of malnutrition is associated with increased morbidity and mortality (25). Therefore, it is crucial to commence nutritional therapy as early as possible. It is imperative to recognise both malnutrition and risk of malnutrition as early as possible during the hospital admission process (6,25). It has previously been shown that comprehensive, individualised nutritional care combining different measures according to individual needs improves the dietary intake, nutritional status, clinical course and quality of life of older patients with malnutrition or at risk of malnutrition (46).
Intervention coordinated by an ED registered dietitian providing nutrition support to older adults identi ed at nutritional risk at presentation to a hospital setting may lead to improved patient outcomes (20).
Consistent with this, interventions designed to improve and sustain optimal nutritional status can also lead to signi cant improvements in quality of life, for both physical and mental aspects (9,14).
A more recent and rapidly growing phenomenon is undernourishment concomitant with obesity in older adults (10). Our study was pragmatic in nature and, re ecting the realities of clinical practice in the ED, body weight and height were measured when the equipment was available and it was feasible to remove the patient from the bed. Therefore, only forty-nine participants had measurements of weight and height taken to measure BMI despite the majority indicating an ability to get out of bed/chair. Nonetheless, 13% (n = 4) of those who were classi ed as overweight/obese based on BMI were at risk of malnutrition. If BMI was relied upon as a sole indicator of nutrition, it would fail to identify nutritional issues in these individuals. This has also been reported in other studies highlighting that older people can be at nutritional risk although they may be overweight or obese (6,43,47).

Limitations
A limitation of this study is the lack of clinical information relating to the conditions of sarcopenia and cachexia. Loss of body mass is a common phenotype for these different conditions that can be caused by a combination of reduced food intake, excessive requirements, altered metabolism, sepsis, trauma, ageing and inactivity (11,48). Cachexia can be de ned as a multifactorial syndrome characterized by severe body weight, fat and muscle loss and increased protein catabolism due to underlying disease (6).
Therefore, malnutrition in older adults may occur due to a combination of cachexia (disease-related) and malnutrition (inadequate consumption of nutrients) as opposed to malnutrition alone. Sarcopenia is a muscle disease common among older adults that is rooted in adverse muscle changes that accrue across a lifetime. It is assessed by measures of muscle quality and is characterised by low muscle strength as a principal determinant (6,49). Determining the primary aetiology (starvation, cachexia, and/or sarcopenia) of malnutrition is critical for the implementation of appropriate nutrition support as responsiveness to dietary modi cations differ (11). This information is also critical in developing an individualised targeted intervention to address malnutrition with the goal of reducing hospital readmission or institutionalisation in this high risk group.

Conclusions
This study found over one in three older adults presenting to an Irish emergency department were at risk of malnutrition or classed as malnourished. The older adults found to be malnourished could be characterised as vulnerable and more likely to suffer adverse outcomes at follow up including declining health and increased use of health care services. The ndings add support to the prioritisation of nutritional screening in clinical practice and public health policy for older adults, particularly targeted towards high risk groups with frailty and multi-morbidity, at increased risk of functional decline. Future research should assess the feasibility and value of integrating ED-based dietetic involvement and multifaceted interventions that would reduce readmission rates and institutionalisation among this high risk group with malnutrition.