Nutrition 29 (2013) Contents lists available at ScienceDirect. Nutrition. journal homepage:

Transcription

1 Nutrition 29 (2013) Contents lists available at ScienceDirect Nutrition journal homepage: Applied nutritional investigation Malnutrition screening tools: Comparison against two validated nutrition assessment methods in older medical inpatients Adrienne M. Young B.Hlth.Sci. a,b, *, Sarah Kidston B.Hlth.Sci. b, Merrilyn D. Banks Ph.D. a,b, Alison M. Mudge Ph.D. a,c, Elisabeth A. Isenring Ph.D. b,c,d a Royal Brisbane and Women s Hospital, Brisbane, Queensland, Australia b Queensland University of Technology, Brisbane, Queensland, Australia c University of Queensland, Brisbane, Queensland, Australia d Princess Alexandra Hospital, Brisbane, Queensland, Australia article info abstract Article history: Received 19 October 2011 Accepted 16 April 2012 Keywords: Undernutrition Diagnosis Triage Hospitalization Aged Objective: Although several validated nutritional screening tools have been developed to triage inpatients for malnutrition diagnosis and intervention, there continues to be debate in the literature as to which tool/tools clinicians should use in practice. This study compared the accuracy of seven validated screening tools in older medical inpatients against two validated nutritional assessment methods. Methods: This was a prospective cohort study of medical inpatients at least 65 y old. Malnutrition screening was conducted using seven tools recommended in evidence-based guidelines. Nutritional status was assessed by an accredited practicing dietitian using the Subjective Global Assessment (SGA) and the Mini-Nutritional Assessment (MNA). Energy intake was observed on a single day during first week of hospitalization. Results: In this sample of 134 participants (80 8 y old, 50% women), there was fair agreement between the SGA and MNA (k ¼ 0.53), with MNA identifying more at-risk patients and the SGA better identifying existing malnutrition. Most tools were accurate in identifying patients with malnutrition as determined by the SGA, in particular the Malnutrition Screening Tool and the Nutritional Risk Screening The MNA Short Form was most accurate at identifying nutritional risk according to the MNA. No tool accurately predicted patients with inadequate energy intake in the hospital. Conclusion: Because all tools generally performed well, clinicians should consider choosing a screening tool that best aligns with their chosen nutritional assessment and is easiest to implement in practice. This study confirmed the importance of rescreening and monitoring food intake to allow the early identification and prevention of nutritional decline in patients with a poor intake during hospitalization. Crown Copyright Ó 2013 Published by Elsevier Inc. All rights reserved. Introduction Despite the high prevalence and negative health consequences, protein energy malnutrition in elderly hospitalized patients continues to be under-recognized and undertreated [1,2]. Malnutrition screening is recommended as the first step in nutritional This study was supported by Queensland Health Strengthening Aged Care Funding and Queensland University of Technology Vacation Research Experience Scholarship. * Corresponding author. Tel.: þ ; fax: þ address: Adrienne_Young@health.qld.gov.au (A. M. Young). care to allow early identification and treatment malnutrition [2 4]. A screening tool needs to be quick and simple and accurately identify patients with possible malnutrition to allow the efficient targeting of resources for nutritional assessment [5,6]. Ideally, such a tool would identify all malnourished patients for assessment (high sensitivity), with a positive screen identifying no well-nourished patients (high positive predictive value) [7]. Many validated screening tools have been recommended for use in elderly and/or hospital populations (Table 1). The Malnutrition Screening Tool (MST) [8] is the most common screening tool used in Australian hospitals [9], whereas the Nutrition Risk Screening (NRS 2002) [10] tool has been successfully /$ - see front matter Crown Copyright Ó 2013 Published by Elsevier Inc. All rights reserved. doi: /j.nut

2 102 A. M. Young et al. / Nutrition 29 (2013) Table 1 Malnutrition screening tools for elderly and/or hospitalized populations [2,3] Screening tool Parameters Development study Validation studies Malnutrition Screening Tool (MST) [8] weight change, recent intake, at-risk score inpatients (mean age 58 y); standard for comparison: SGA; sensitivity 93%, specificity 93% SGA: sensitivity 92%, specificity 61% [15]; MNA: sensitivity 92%, specificity 72% [15] Mini-Nutritional Assessment Short Form (MNA-SF) [12] Nutritional Risk Screening (NRS 2002) [10] Malnutrition Universal Screening Tool (MUST) [22] Short Nutritional Assessment Questionnaire (SNAQÓ) [24] Simplified Nutritional Appetite Questionnaire (SNAQ) [13] Rapid Screen [14] weight change, recent intake, BMI, acute disease, mobility, dementia/depression, at-risk score 11 weight change, recent intake, BMI, acute disease, age, at-risk score 3 weight change, recent/predicted intake, BMI, acute disease, high-risk score 2 weight change, appetite, supplements/tube feeding, at-risk score 2 recent intake, appetite, satiety, taste change, at-risk score <14 weight change, BMI, at-risk score community-dwelling elders (mean age 79 y); standard for comparison: physician assessment of nutritional status; sensitivity 98%, specificity 100% (MNA-SF cutpoint 10) 8944 inpatients, review of 128 trials (mean age not reported); standard for comparison: nutritional support trials demonstrating improved clinical outcomes; sensitivity 75%, specificity 55% adapted from Malnutrition Advisory Group screening tool 291 inpatients (mean age 58 y); standard for comparison: BMI <18.5 or weight loss >5%; sensitivity 86%, specificity 89% 352 community-dwelling elderly (age range y); standard for comparison: future weight loss 5%; sensitivity 82%, specificity 85% 65 medical and orthopaedic inpatients, subacute residents (mean age 80 y); standard for comparison: standard nutritional assessment; sensitivity 79%, specificity 97%. BMI, body mass index; MNA, Mini-Nutritional Assessment [26]; N/A, not available; SGA, Subjective Global Assessment [25] MNA: sensitivity 90%, specificity 88% (MNA-SF cutpoint 11) [16]; MNA: sensitivity 89%, specificity 82% (MNA-SF cutpoint 11) [17]; nutritional assessment : sensitivity 100%, specificity 38% (MNA-SF cutpoint 10) [18] SGA: sensitivity 74%, specificity 87% [19]; SGA: sensitivity 62%, specificity 93% [20]; MNA: k ¼ 0.39 [19], k ¼ 1.00 [21] SGA: sensitivity 61%, specificity 79% [20]; SGA: sensitivity 72%, specificity 90% [19]; MNA: k ¼ 0.39 [19], k ¼ 0.55 [23] BMI <18.5 or recent weight loss >5%: sensitivity 79%, specificity 83% [24] N/A N/A implemented throughout Europe [11]. The Mini-Nutritional Assessment Short Form (MNA-SF) [12] has been recommended for screening elderly people across settings [4]. The Simplified Nutritional Appetite Questionnaire (SNAQ) [13] and the Rapid Screen [14] were developed in community-dwelling populations but have not yet been validated in the hospital setting. There are distinct similarities among tools, with most including recent changes in weight and food intake and some accounting for body mass index (BMI) and acute disease (Table 1) [8,10,12 26]. A major limitation in validating malnutrition screening tools is the absence of a single objective measurement or gold standard for diagnosing malnutrition [27]. The Subjective Global Assessment (SGA) [25] and the Mini-Nutritional Assessment (MNA) [26] are widely used validated nutritional assessments that use a range of parameters to make a nutritional diagnosis and initiate treatment, have been recommended as outcomes in clinical trials [28], and predict health outcomes in elderly hospitalized patients [29,30]. There is a key difference between nutritional assessment and nutritional screening [31]: the SGA and MNA are comprehensive nutritional assessments used by trained professionals (e.g., dietitians, physicians, trained nurses, or research assistants) to diagnose malnutrition and initiate a nutritional intervention. In contrast, nutrition screening tools (such as those listed in Table 1) are intended as a quick and easy method for identifying possible malnutrition and to triage patients for comprehensive nutritional assessment and intervention. Although the SGA and MNA have been recommended for use in diagnosing malnutrition in the elderly [2], there are substantial differences between the parameters of these assessments, meaning that different at-risk groups may be identified [21,32]. Although previous research has compared existing nutritional screening tools, no study has concurrently investigated the validity of these tools against the SGA and MNA. Studies have shown that nutritional status declines during hospitalization [33] and that nutritional intake is suboptimal [1, 34,35]. In elderly hospitalized patients, it has become clear that malnutrition on admission does not necessarily predict a poor nutritional intake during hospitalization [1]. In fact, many well-nourished inpatients eat poorly, presenting a second group of patients who should be identified early in their hospital admission to prevent malnutrition. This highlights the importance of screening and rescreening elderly patients to identify not only those with existing malnutrition but also those at risk of poor intake during hospitalization. Although malnutrition screening tools are commonly used to identify existing malnutrition [9,11], there are no screening tools to proactively identify patients at risk of poor nutritional intake during hospitalization. This study aimed to 1) compare the assessment of malnutrition using the SGA and MNA in elderly medical inpatients, 2) compare the accuracy of seven nutritional screening tools in identifying patients with malnutrition as assessed by the SGA and MNA, and 3) compare the predictive accuracy of screening tools to identify patients with poor energy intake during the first week of hospitalization. Materials and methods This was a prospective cohort study conducted in the medical wards of the Royal Brisbane and Women s Hospital, a large metropolitan public teaching hospital in Brisbane, Australia, and was part of a larger observational study of nutritional intake in older medical patients [1]. The study was approved by the hospital human research ethics committee. Consecutive patients at least 65 y old with a hospital stay longer than 2 d were recruited from November 2007 through March From day 3 through day 7 of admission, a single trained dietitian (A. M. Y.) screened each patient with the MST, MNA-SF, Malnutrition Universal Screening Tool (MUST), NRS 2002, Short Nutritional Assessment Questionnaire (SNAQÓ), SNAQ, and Rapid Screen (presented in Table 1). We draw the reader s attention to the differences between the two SNAQ tools in Table 1. These screening tools were selected because they have been recommended for use in evidence-based practice guidelines [2,3]. Each tool was performed separately and according to the authors instructions. The same dietitian assessed each participant using the SGA [25] and MNA [26].

3 A. M. Young et al. / Nutrition 29 (2013) Table 2 Participants characteristics (n ¼ 133) Men 67 (50%) Women 66 (50%) Diagnosis Chronic cardiorespiratory disease 36 (27%) Acute infection 27 (20%) Gastrointestinal disease 13 (10%) Cancer 10 (8%) Other 47 (35%) Living situation Community living 114 (86%) Residential care 19 (14%) Body mass index (kg/m 2 ) <21 27 (20%) (57%) (23%) Values are numbers of subjects (percentages) Nutritional screening and assessment data were available for all participants, with the exception of the SNAQ (missing data for two participants). Dietary intake was measured at breakfast, lunch, and dinner on the same day of the nutritional screening and assessment. The plate waste of each meal component (e.g., soup, meat, and vegetables) was visually estimated, which correlates closely to weighed methods [36]. Mid-meal intake, including snacks and/or nutritional supplements (ordered for 20% of participants according to the existing nutritional support protocol), was estimated by observation and/or patient recall. The food intake on a single day during hospitalization has been shown to correlate closely with intake over 2 or 3 d [1,37]. Energy intake was determined using the known food composition data of each meal component and an analysis of standardized recipes in FoodWorks Professional 3.02 (Xyris, Brisbane Australia, 2004). Resting energy expenditure was calculated as 18.4 kcal/kg of bodyweight per day for patients with a BMI higher than 21 kg/m 2 and 21.4 kcal/kg per day for those with a BMI no higher than 21 kg/m 2 [38]. Nutritional assessments were categorized as well nourished (SGA A; MNA score 24) or malnourished or at risk of malnutrition (SGA B or C; MNA score <24). Scores for each screening tool were also categorized into no/low risk or at risk of malnutrition using recommended cutpoints (Table 1). Inadequate energy intake was defined as a measured energy intake lower than the resting energy expenditure. Participants characteristics were summarized using mean standard deviation for continuous variables or categorized according to validated cutoffs and clinical meaning. To compare the performance of the two nutritional assessments (SGA and MNA), k statistics were calculated and interpreted using the criteria by Shrout [39]. To compare the accuracy of each screening tool to detect malnutrition as diagnosed using each nutritional assessment, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. Sensitivity was defined as the proportion of malnourished correctly identified as such, whereas specificity was the proportion of wellnourished who were correctly identified as well nourished. The PPV was the proportion of patients with a positive screen result who were malnourished. Conversely, the NPV was the proportion of patients with a negative screen result who were well nourished. These were calculated for the three outcomes of interest: 1) malnutrition assessed using the SGA, 2) malnutrition assessed using the MNA, and 3) inadequate energy intake. In further analysis, raw scores for each tool (except the Rapid Screen, which produces dichotomous data) were used to construct receiver operating characteristic (ROC) curves, where the sensitivity was plotted against the false positive rate (1 specificity) for each outcome of interest. Area under the curve (AUC) values for each ROC curve were interpreted as follows: acceptable ( ), excellent ( ), or outstanding (>0.90) [40]. ROC analysis was also used to explore instrument cutpoints. Results Participants and nutritional assessments Over the 16-wk study period, 134 patients (mean age 80 8 y, 50% women, mean weight kg, mean BMI 26 6 kg/m 2, median length of stay 8 d, interquartile 8 d) consented to participate in the study (38% consent rate). One participant was excluded because of incomplete data. Participants characteristics are presented in Table 2. Non-participants had similar demographic characteristics and lengths of stay but were more likely to be discharged to residential aged care (24% versus 13%). There was fair agreement between the SGA and MNA (k ¼ 0.53, 95% confidence interval ; Table 3). More participants were assessed as at risk or malnourished using the MNA (68%) than using the SGA (47% malnutrition). Malnutrition as determined by SGA The performance of each screening tool to identify malnutrition determined by the SGA (rating of B or C) is presented in Table 4. The MST, NRS 2002, MUST, and SNAQÓ had a high sensitivity and PPV, with the MST and NSR 2002 achieving a slightly better NPV. Although the MNA-SF and SNAQ were highly sensitive, they had a lower specificity and PPV, meaning more well-nourished patients would be identified for assessment. Conversely, the Rapid Screen was highly specific but had a very low sensitivity (29%), indicating that many malnourished patients could be missed using this tool. All tools (excluding the Rapid Screen) showed an excellent to outstanding discrimination between those who were and those who were not malnourished using the AUC analysis. The cutoff point for the MNA-SF with the highest accuracy in this sample was 8 (compared with the published cutoff of 11), with a sensitivity and specificity of 89%. Malnutrition risk as determined by MNA When malnutrition was determined using the MNA (score <24), the MNA-SF was most sensitive, with a good PPV (Table 5). All other tools tested were highly specific but were less sensitive. In particular, the Rapid Screen had a very low sensitivity in this sample (20%). Using AUC analysis, the MNA-SF was outstanding at discriminating between those who were and those who were not at risk of malnutrition with the MNA. All other tools showed an excellent discrimination, with ROC curves demonstrating that using lower cutpoints for the MST, NSR 2002, and SNAQÓ and a higher cutpoint for the SNAQ could increase the sensitivity of these tools to identify malnutrition risk determined by MNA. Table 3 Comparison of malnutrition assessments using the SGA and MNA (n ¼ 133) SGA MNA Total Well nourished (score 24) At risk (score ) Malnourished (score <17) Well nourished (SGA A) 41 (30.8%) 28 (21.1%) 2 (1.5%) 71 (53.4%) Malnutrition (SGA B þ C) 2 (1.5%) 22 (16.5%) 38 (28.6%) 62 (46.6%) Total 43 (32.3%) 50 (37.6%) 40 (30.1) 133 (100%) MNA, Mini-Nutritional Assessment; SGA, Subjective Global Assessment Values are numbers of subjects (percentages); k ¼ 0.53 (95% confidence interval ) for agreement between the SGA and MNA (well nourished [SGA A, MNA 24] versus at risk or malnourished [SGA B or C; MNA <2])

4 104 A. M. Young et al. / Nutrition 29 (2013) Table 4 Accuracy of screening tools at identify malnutrition (as determined by Subjective Global Assessment rating B or C; malnutrition) Screening tool Sensitivity (95% CI) Specificity (95% CI) PPV (95% CI) NPV (95% CI) AUC (95% CI) AUC discrimination * Malnutrition Screening Tool (MST) 90.3% ( ) 84.7% ( ) 83.6% ( ) 91.0% ( ) 0.92 ( ) outstanding Mini Nutritional Assessment Short 100% ( ) 52.8% ( ) 64.6% ( ) 100% ( ) 0.95 ( ) outstanding Form (MNA-SF) Nutritional Risk Screening 90.3% ( ) 83.3% ( ) 82.4% ( ) 90.9% ( ) 0.89 ( ) excellent (NRS 2002) Malnutrition Universal Screening 87.1% ( ) 86.1% ( ) 84.4% ( ) 88.6% ( ) 0.89 ( ) excellent Tool (MUST) Short Nutritional Assessment 79.0% ( ) 90.3% ( ) 87.5% ( ) 83.3% ( ) 0.93 ( ) outstanding Questionnaire (SNAQÓ) Simplified Nutritional Appetite 86.9% ( ) 78.9% ( ) 77.9% ( ) 87.5% ( ) 0.87 ( ) excellent Questionnaire (SNAQ) Rapid Screen 29.0% ( ) 100% ( ) 100% ( ) 62.1% ( ) N/A AUC, area under the curve; CI, confidence interval; N/A, not applicable; NPV, negative predictive value; PPV, positive predictive value * Classification of AUC (range 0 1): acceptable , excellent , outstanding >0.90 [40]. Inadequate energy intake Most participants (59%) had an inadequate energy intake to meet the estimated resting energy expenditure. All screening tools had a low sensitivity and specificity for predicting patients with an inadequate energy intake on a single day during their first week of hospitalization (Table 6). The AUC analysis showed that no screening tool adequately discriminated between those who had adequate versus inadequate energy intake. The SNAQ obtained the highest level of discrimination (0.66) but did not reach an acceptable level. Discussion This study compared the accuracy of validated malnutrition screening tools against two commonly used nutritional assessments (SGA and MNA) in a sample of elderly medical inpatients. Only a fair agreement was found between the SGA and MNA (k ¼ 0.53), indicating that these nutritional assessments identify different at-risk groups. Velasco et al. [19] reported a similar agreement between the SGA and MNA (k ¼ 0.49) in their study of 400 hospitalized patients, as did Persson et al. [32] and Martins et al. [21] who proposed that, because of its holistic approach, the MNA identifies those at risk and those with existing malnutrition. In contrast, the SGA identifies existing malnutrition only. The choice of nutritional assessment tool should be guided by the goal of therapy; that is, whether the goal is prevention or treatment [21,32,41]. This suggests that the MNA may be better suited when a service aims to prevent malnutrition or where there is a well-resourced dietetic workforce, whereas the SGA may be more useful in the acute setting for identifying existing malnutrition to be prioritized for treatment during the short timeframe of hospitalization. The primary objective of this study was to compare the performance of seven screening tools to identify patients with malnutrition. When nutritional status was assessed using the SGA, most tools performed with a high sensitivity and specificity, with the MST and NRS 2002 having the highest accuracy. In the present study, the increased complexity of the NRS 2002 (which includes medical condition and BMI) did not improve the accuracy compared with the simpler MST. A similar accuracy between simple screening tools (MST and SNAQÓ) and more comprehensive tools (MUST and NSR 2002) has been reported previously [42]. Because it is important that nutritional screening be quick and easy and completed by anyone (e.g., nursing staff, medical staff, allied health assistants, or patients themselves) [3], the MST has been recommended as a highly accurate and userfriendly malnutrition screening tool [6,42]. Other reviews have found the NSR 2002 and MUST to have high accuracy [19,20]. As outlined in Table 1, the screening tools compared in this study included similar parameters, so it is not unexpected that they have a similar performance. As reported previously, the MNA-SF had a high sensitivity and specificity when used with the MNA [16,17]; however, a poor specificity and PPV were observed when compared with the SGA. The MNA-SF was designed to identify patients requiring further assessment with the MNA, and the poor performance against the SGA likely is to be due to the different focus of the MNA and SGA, Table 5 Accuracy of screening tools at identify malnutrition (as determined by Mini-Nutritional Assessment score <24; at risk of malnutrition/malnourished) Screening tool Sensitivity (95% CI) Specificity (95% CI) PPV (95% CI) NPV (95% CI) AUC (95% CI) AUC discrimination * Malnutrition Screening Tool (MST) 67.7% ( ) 88.3% ( ) 92.4% ( ) 56.7% ( ) 0.87 ( ) excellent Mini Nutritional Assessment Short 95.6% ( ) 79.1% ( ) 90.5% ( ) 89.5% ( ) 0.96 ( ) outstanding Form (MNA-SF) Nutritional Risk Screening 72.2% ( ) 95.3% ( ) 97.0% ( ) 62.1% ( ) 0.90 ( ) excellent (NRS 2002) Malnutrition Universal Screening 67.8% ( ) 93.0% ( ) 95.3% ( ) 58.0% ( ) 0.82 ( ) excellent Tool (MUST) Short Nutritional Assessment 62.2% ( ) 100% ( ) 100% ( ) 55.8% ( ) 0.89 ( ) excellent Questionnaire (SNAQÓ) Simplified Nutritional Appetite 69.3% ( ) 83.7% ( ) 89.7% ( ) 57.1% ( ) 0.83 ( ) excellent Questionnaire (SNAQ) Rapid Screen 20.0% ( ) 100% ( ) 100% ( ) 37.4% ( ) N/A AUC, area under the curve; CI, confidence interval; N/A, not applicable; NPV, negative predictive value; PPV, positive predictive value * Classification of AUC (range 0 1): acceptable , excellent , outstanding >0.90 [40].

5 A. M. Young et al. / Nutrition 29 (2013) Table 6 Accuracy of screening tools at predicting inadequate energy intake (measured energy intake below estimated resting energy expenditure) Screening tool Sensitivity (95% CI) Specificity (95% CI) PPV (95% CI) NPV (95% CI) AUC (95% CI) AUC discrimination * Malnutrition Screening Tool (MST) 50.0% ( ) 50.0% ( ) 58.2% ( ) 41.8% ( ) 0.52 ( ) poor Mini Nutritional Assessment- Short 71.8% ( ) 28.6% ( ) 58.3% ( ) 42.1% ( ) 0.52 ( ) poor Form (MNA-SF) Nutritional Risk Screening 50.6% ( ) 48.2% ( ) 58.0% ( ) 40.9% ( ) 0.53 ( ) poor (NRS 2002) Malnutrition Universal Screening 43.6% ( ) 46.4% ( ) 53.1% ( ) 37.1% ( ) 0.45 ( ) poor Tool (MUST) Short Nutritional Assessment 40.5% ( ) 55.4% ( ) 56.1% ( ) 39.7% ( ) 0.53 ( ) poor Questionnaire (SNAQÓ) Simplified Nutritional Appetite 61.8% ( ) 62.5% ( ) 69.1% ( ) 54.7% ( ) 0.66 ( ) poor Questionnaire (SNAQ) Rapid Screen 10.3% ( ) 82.1% ( ) 44.4% ( ) 39.7% ( ) N/A AUC, area under the curve; CI, confidence interval; N/A, not applicable; NPV, negative predictive value; PPV, positive predictive value * Classification of AUC (range 0 1): acceptable , excellent , outstanding >0.90 [40]. as discussed earlier. A recent study of 275 hospitalized patients also reported a low specificity of the MNA-SF in identifying existing malnutrition [42]. To improve the specificity of the MNA-SF, clinicians could consider lowering the cutoff of the MNA-SF score, as suggested by the original authors [12]. Although the revised MNA-SF (where BMI is substituted for calf circumference) was not tested in this study, this tool could be expected to perform similarly to the original MNA-SF, as shown previously [17]. The Rapid Screen had a very poor sensitivity against the SGA and MNA, suggesting it may identify only the most severely malnourished. Further validation studies are recommended before using this tool in elderly hospitalized patients. In summary, with the exception of the Rapid Screen, all tools (including simple tools such as the MST) were accurate in identifying malnutrition using the SGA and therefore can be recommended for use in elderly hospitalized patients. Although the MNA-SF was accurate, it identified a larger number of at-risk patients, as also reported by Raslan et al. [43] and, therefore, should be chosen only where health services have sufficient resources to provide nutritional assessment and intervention to all at-risk patients. When choosing which screening tools to use in practice, clinicians should consider which tool is simple to implement and the resources available to provide nutritional care to all at-risk patients. In this study, most participants had inadequate energy intakes. However, no screening tool accurately discriminated between those with adequate and those with inadequate intake. Although all screening tools, with the exception of the Rapid Screen, include a brief assessment of recent dietary intake, this study has found that they do not adequately identify those with a poor intake during hospitalization. This finding may reflect the other important predictors of poor nutritional intake, such as delirium and feeding dependency [2], which are not all adequately covered in these screening tools. Barriers to nutritional intake also may be related to the hospital environment and culture, e.g., quality of hospital food, interruptions during mealtimes, and lack of mealtime assistance [44]. This study demonstrates an absence of existing screening tools to proactively identify patients at risk of poor nutritional intake and supports the concept of two discrete nutritionally at-risk groups for which different nutritional care processes are required: malnutrition screening to identify existing malnutrition and close monitoring of food intake to identify inadequate nutritional intake. This is the first study to compare a range of screening tools against two recommended nutritional assessments in elderly hospitalized patients. It is also the first study to consider the accuracy of these tools to identify poor nutritional intake in the hospital, which is common in this patient group. We do recognize some study weaknesses. Although the assessment tools (SGA and MNA) are widely used by health professionals and the research community to diagnose malnutrition, there is no single objective measurement of malnutrition against which to validate the screening tools. An important part of assessing the performance of a screening tool is to consider the reliability of the tool. Because one dietitian performed all the screenings and assessments in this study, the reliability of measurements is increased. However, we are unable to comment more generally on the reliability of the tools or the performance of the tools when used by non-dietetic staff. However, the high inter-rater reliability of the tools has been reported previously [8,22,24]. Dietary intake was measured only on a single day, but we previously showed a close correlation of intake between day 3 and day 7 of hospitalization [1]. A further limitation of the assessment of dietary intake was the estimation of the energy requirements of individual participants, rather than measurement using indirect calorimetry. The low consent rate may have resulted in an under-representation of the frailest group of patients because fewer participants were discharged to residential aged care compared with the general elderly medical population. However, this is not likely to have affected the comparison of the screening tools. Conclusion With the exception of the Rapid Screen, all screening tools were accurate in identifying malnutrition (as assessed by common clinical assessment tools) and therefore can be recommended for use in elderly hospitalized patients. No tool predicted a poor nutritional intake during hospitalization, highlighting importance of rescreening and monitoring intake to allow the early identification and prevention of nutritional decline. Acknowledgments The authors thank Dr. Lynda Ross, Maria Cenita, Dianne Jones, and the staff and patients of Internal Medicine and Aged Care. References [1] Mudge AM, Ross LJ, Young AM, Isenring EA, Banks MD. Helping Understand Nutritional Gaps in the Elderly (HUNGER): a prospective study of patient factors associated with inadequate nutritional intake in older medical inpatients. Clin Nutr 2011;30: [2] Watterson C, Fraser A, Banks M, Isenring E, Silvester C, Hoevenaars R, et al. Evidence based practice guidelines for nutritional management of

6 106 A. M. Young et al. / Nutrition 29 (2013) malnutrition in adult patients across the continuum of care. Nutr Diet 2009;66:S1 34. [3] Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. ESPEN guidelines for nutrition screening Clin Nutr 2003;22: [4] Mueller C, Compher C, Ellen DM. A.S.P.E.N. clinical guidelines: nutrition screening, assessment, and intervention in adults. JPEN J Parenter Enteral Nutr 2011;35: [5] Elia M, Stratton RJ. An analytic appraisal of nutrition screening tools supported by original data with particular reference to age. Nutrition 2012;28: [6] van Venrooij LMW, de Vos R, Borgmeijer-Hoelen AMMJ, Kruizenga HM, Jonkers-Schuitema CF, de Mol BAMJ. Quick-and-easy nutritional screening tools to detect disease-related undernutrition in hospital in- and outpatient settings: a systematic review of sensitivity and specificity. e-spen 2007;2: [7] Anthony PS. Nutrition screening tools for hospitalized patients. Nutr Clin Pract 2008;23: [8] Ferguson M, Capra S, Bauer J, Banks M. Development of a valid and reliable malnutrition screening tool for adult acute hospital patients. Nutrition 1999;15: [9] Ferguson M, Banks M, Bauer J, Isenring E, Vivanti A, Capra S. Nutrition screening practices in Australian healthcare facilities: a decade later. Nutr Diet 2010;67: [10] Kondrup J, Rasmussen HH, Hamberg O, Stanga Z. Nutritional risk screening (NRS 2002): a new method based on an analysis of controlled clinical trials. Clin Nutr 2003;22: [11] Sorensen J, Kondrup J, Prokopowicz J, Schiesser M, Krähenbühl L, Meier R, Liberda M. EuroOOPS: an international, multicentre study to implement nutritional risk screening and evaluate clinical outcome. Clin Nutr 2008;27: [12] Rubenstein LZ, Harker JO, Salva A, Guigoz Y, Vellas B. Screening for undernutrition in geriatric practice: developing the short-form mininutritional assessment (MNA-SF). J Gerontol A Biol Sci Med Sci 2001;56:M [13] Wilson MMG, Thomas DR, Rubenstein LZ, Chibnall JT, Anderson S, Baxi A, et al. Appetite assessment: simple appetite questionnaire predicts weight loss in community-dwelling adults and nursing home residents. Am J Clin Nutr 2005;82: [14] Visvanathan R, Penhall R, Chapman I. Nutritional screening of older people in a sub-acute care facility in Australia and its relation to discharge outcomes. Age Ageing 2004;33: [15] Correia JR, Martins CA, Amaral TF. Efficiency of MSTdMalnutrition Screening Tooldin elderly hospitalised patients. Clin Nutr 2003;22:S10 1. [16] Lei Z, Qingyi D, Feng G, Chen W, Hock RS, Changli W. Clinical study of mininutritional assessment for older Chinese inpatients. J Nutr Health Aging 2009;13: [17] Kaiser MJ, Bauer JM, Ramsch C, Uter W, Guigoz Y, Cederholm T, et al. Validation of the Mini Nutritional Assessment Short-Form (MNA-SF): a practical tool for identification of nutritional status. J Nutr Health Aging 2009;13: [18] Ranhoff AH, Gjoen AU, Mowe M. Screening for malnutrition in elderly acute medical patients: the usefulness of MNA-SF. J Nutr Health Aging 2005;9: [19] Velasco C, Garcia E, Rodriguez V, Frias L, Garriga R, Alvarez J, et al. Comparison of four nutritional screening tools to detect nutritional risk in hospitalized patients: a multicentre study. Eur J Clin Nutr 2011;65: [20] Kyle UG, Kossovsky MP, Karsegard VL, Pichard C. Comparison of tools for nutritional assessment and screening at hospital admission: a population study. Clin Nutr 2006;25: [21] Martins CP, Correia JR, do Amaral TF. Undernutrition risk screening and length of stay of hospitalized elderly. J Nutr Elder 2005;25:5 21. [22] Elia M. The MUST report. Nutritional screening of adults: a multidisciplinary responsibility. Redditch: British Association for Parenteral and Enteral Nutrition; [23] Stratton RJ, Hackston AJ, Price S, Joseph K, Elia M. Concurrent validity of the newly developed Malnutrition Universal Screening Tool ( MUST ) with Mini Nutritional Assessment and Subjective Global Assessment tools. Proc Nutr Soc 2004;63:17A. [24] Kruizenga HM, Seidell JC, de Vet HC, Wierdsma NJ, van Bokhorst-de van der Schueren MA. Development and validation of a hospital screening tool for malnutrition: the Short Nutritional Assessment Questionnaire (SNAQ). Clin Nutr 2005;24: [25] Detsky AS, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA, Jeejeebhoy KN. What is subjective global assessment of nutritional status? JPEN J Parenter Enteral Nutr 1987;11:8 13. [26] Guigoz Y, Vellas B, Garry PJ. Assessing the nutritional status of the elderly: the Mini Nutritional Assessment as part of the geriatric evaluation. Nutr Rev 1996;54:S [27] Meijers JMM, Vanderschueren MAE, Schols J, Soeters PB, Halfens RJG. Defining malnutrition: Mission or mission impossible? Nutrition 2010;26: [28] Salva A, Corman B, Andrieu S, Salas J, Vellas B. Minimum data set for nutritional intervention studies in elderly people. J Gerontol A Biol 2004;59: [29] Covinsky KE, Martin GE, Beyth RJ, Justice AC, Sehgal AR, Landefeld CS. The relationship between clinical assessments and nutritional status and adverse outcomes in older hospitalised medical patients. J Am Geriatr Soc 1999;47: [30] Van Nes M, Herrmann FR, Gold G, Michel J, Rizzoli R. Does the Mini Nutritional Assessment predict hospitalization outcomes in older people? Age Ageing 2001;30: [31] Charney P. Nutrition screening vs nutrition assessment: how do they differ? Nutr Clin Pract 2008;23: [32] Persson MD, Brismar KE, Katzarski KS, Nordenstrom J, Cederholm TE. Nutritional status using mini nutritional assessment and subjective global assessment predict mortality in geriatric patients. J Am Geriatr Soc 2002;50: [33] Braunschweig C, Gomez S, Sheean PM. Impact of declines in nutritional status on outcomes in adult patients hospitalized for more than 7 days. J Am Diet Assoc 2000;100: [34] Barton AD, Beigg CL, Macdonald IA, Allison SP. High food wastage and low nutritional intakes in hospital patients. Clin Nutr 2000;19: [35] Sullivan DH, Sun S, Walls RC. Protein energy undernutrition among elderly hospitalized patients: a prospective study. JAMA 1999;281: [36] Odlund Olin A, Osterberg P, Hadell K, Armyr I, Jerstrom S, Ljungqvist O. Energy-enriched hospital food to improve energy intake in elderly patients. JPEN J Parenter Enteral Nutr 1996;20:93 7. [37] Breslow RA, Sorkin JD. Comparison of one-day and 3-day calorie counts in hospitalized-patientsda pilot-study. J Am Geriatr Soc 1993;41: [38] Alix E, Berrut G, Bore M, Bouthier-Quintard F, Buia JM, Chlala A, et al. Energy requirements in hospitalized elderly people. J Am Geriatr Soc 2007;55: [39] Shrout PE. Measurement reliability and agreement in psychiatry. Stat Methods Med Res 1998;7: [40] Hosmer D, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons; [41] Christensson L, Unosson M, Ek AC. Evaluation of nutritional assessment techniques in elderly people newly admitted to municipal care. Eur J Clin Nutr 2002;56: [42] Neelemaat F, Meijers J, Kruizenga H, van Ballegooijen H, van Bokhorst-de van der Schueren M. Comparison of five malnutrition screening tools in one hospital inpatient sample. J Clin Nurs 2011;20: [43] Raslan M, Gonzalez MC, Dias MC, Nascimento M, Castro M, Marques P, et al. Comparison of nutritional risk screening tools for predicting clinical outcomes in hospitalized patients. Nutrition 2010;26: [44] Ross LJ, Mudge AM, Young AM, Banks M. Everyone s problem but nobody s job: staff perceptions and explanations for poor nutritional intake in older medical patients. Nutr Diet 2011;68:41 6.