Abstract/Resume
The purpose of nutrition screening is to identify individuals at high nutritional risk. Given that dietitians cannot always carry out screening in health-care facilities, tools should be simple and based on data obtained from the nursing admission questionnaire. This study was conducted to develop timely and valid tools for screening protein-energy malnutrition (PEM). A dietetic technician administered an initial screening tool to 160 subjects recruited from two settings. This tool comprised nine PEM risk factors. The sample included 54 adults in acute care, 57 elderly adults in acute care, and 49 elderly adults in long-term care. Dietitians performed comprehensive nutritional assessments to determine the validity of this screening tool. Step-%rise regression analysis revealed significant risk factors among those included in the initial screening. These risk factors were considered during development of the first simple screening tool, which encompassed body mass index (BMI) and percentage of weight loss, and classified subjects as having low or high PEM risk levels. A second tool using BMI and albumin level was tested in cases where an albumin measurement was available upon admission. These simple tools had validity indices of 75.9% or higher, except in adults in acute care; sensitivity was low in this group. The tools proved helpful in establishing dietitians' priorities for involvement and in initiating early nutritional care.
(Can J Diet Prac Res 2001; 62:26-34)
Le depistage nutritionnel identifie les individus a risque nutritionnel eleve. Puisque les ressources en nutrition sont limitees dans les etablissements de soins de same pour depister, l'outil doit etre bref, base sur les donnees du questionnaire d'admission. Cette etude visait a developper un systeme simple et valide pour depister le risque de malnutrition proteino-energetique (MPE). Une technicienne en dietet6t]que a applique un outil de depistage initial, comprenant neuf facteurs de risque, chez 160 sujets repartis en trois categories: 54 adultes et 57 personnel agees hospitalises (soins de courte duree) et 49 personnel agees requerant des soins de Tongue duree. Deux dietetistes ont effectue des evaluations nutritionnelles pour evaluer la validite des outils. Une regression logistique a identifie les facteurs de risque significatifs inclus dans l'outil initial. Ceux-ci composent un premier outil simple, comprenant Vindice de la masse corporelle (IMC) et le pourcentage de perte de poids. Un second outil comprenant PIMC et l'albumine a ete evalue au cas ou l'albumine serait disponible a l'admission. Ces outils simplifies indiquent des mesures de la validite de 75,9 % et plus, excepte chez les adultes hospitalises ou la sensibilite est faible, et permettent d'initier tot les soins nutritionnels. (Rev can prat rech dieter 2001; 62:26-34)
INTRODUCTION
Studies over the past decade have continued to show a high prevalence of malnutrition in health-care facilities. At least 40-50% of adult and elderly inPATIENTS ARE MALNOURISHED (1-4). The incident of protien-energy malnutrition (PEM) has been reported to be a high as 60% in American nursing homes (5). Malnutrition forewarns of complications (3) and increases risks of morbidity, mortality, and longer hospitalization; these factors increase medical costs (6). mOREOVER, PEM exacerbates functional dependency in the elderly population (7). Fortunately, nutritional support can improve patients' clinical outcome (8-10).
To pinpoint individuals at risk of malnutrition or with undetected malnutrition, a systematic nutrition screening is required in health-care facilities; screening will help determine the need to assess the patient's nutritional status (11). Given the current financial and staff restrictions in healthcare facilities, simple and valid tools are necessary for timely and cost-effective screening. The criterion-related validity of a screening tool is generally determined by evaluating sensitivity, specificity, and overall predictive value (PV) (12). Sensitivity is defined as the tool's ability to identify correctly individuals truly at nutritional risk, i.e., true positives. Specificity measures the tool's ability to identify correctly persons who are not at nutritional risk, i.e., true negatives (13). A tool's overall PV is defined as its ability to predict correctly the presence or absence of nutritional risk (14).
Although various nutrition screening tools are available, their validity is not always documented (15,16). One of the screening tools validated (although simple) targeted only surgical patients, and its validity indices are not clearly defined (8). This tool's specificity is also questionable when it is applied to elderly subjects (17). With another simple screening tool, the authors did not specify whether the elderly inpatients participated in the validation study (18). The true criterion used to assess the validity of some of these tools is also questionable. It is based only on pre-albumin levels (19) or is mainly subjective, based on an overall clinical judgment of subjects' nutritional status; there is little detail about the nutritional assessment protocols (18,20).
Some screening tools have been validated specifically with the elderly population in health-care facilities. However, most tools are complex, given the many risk factors they evaluate. Some tend to blur the line between screening and assessment (21-23). One of them, the Nutrition Screening Initiative (NSI), showed low validity indices (21,24). Although a tool developed for nursing home residents proved to be reliable and valid, i.e., it had convergent and construct validity, its criterion validity was not assessed. This was justified by the absence of a standardized instrument for comprehensive assessment of the older adults' nutritional status (23).
There is still a need for valid, reliable, and simple nutrition screening tools for routine use throughout health-care facilities. Such tools will help ensure that patients receive nutritional care as early as possible.
OBJECTIVES
The goal of this study was to develop simple and valid tools, i.e., tools showing sensitivity, specificity, and overall PV of 80% or greater, for screening PEM risk among adult and elderly inpatients, and among the elderly in long-term care facilities. More specifically, the goal was to:
1. identify significant risk factors included in an initial screening tool comprising nine PEM risk factors, and
2. assess the validity of two simple screening tools comprising the significant risk factors.
METHODS
Subjects
The study included 160 subjects divided into three categories: 54 acute care adults, aged 64 or younger (ACA); 57 acute care elderly adults, aged 65 or older (ACE); and 49 long-term care elderly adults, aged 65 or older (LTCE). The subjects were recruited from two settings between May and August 1996. Those in the ACA and ACE categories were included in the study following admission to the Campbellton Regional Hospital (CRH). Patients admitted to the obstetrics/gynecology and psychiatry units or with a diagnosis of myocardial infarction were excluded. The subjects in the LTCE category were randomly enlisted from 100 residents at the Village Campbellton Nursing Home.
The research protocol was approved by the Ethics Committee of the Faculty of Graduate Studies and Research, Universite de Moncton, and by representatives of each research facility. Each subject eligible to take part in the study signed an informed consent form. For cognitively impaired elderly patients, consent was obtained through a family member.
INSTRUMENTS AND DATA COLLECTION
Initial PEM screening tool: The initial PEM screening tool included nine risk factors described in Table 1. Risk factors were chosen from those proposed by the American Dietetic Association to determine nutritional risk (25), as well as from those included in the reviewed nutrition screening tools. The risk factors' documented or clinically observed potential for identifying PEM and their practicality in a clinical setting also were considered.
A dietetic technician administered this initial screening tool. The ACA and ACE subjects were screened within 72 hours of hospital admission. Most of the information was obtained from patient interviews. Biochemical data and diagnoses/medical conditions were identified through chart reviews. The elderly nursing home residents were screened as they were selected to participate in the study. In the nursing home, usual and current weights were those documented monthly in each resident's chart. Because some residents were cognitively impaired, some information was gathered by interviewing the nursing staff.
According to Baden et al. (31), body mass index (BMI) is the most commonly used measure to interpret weight for height, even in the elderly, in whom a BMI of less than approximately 24 kg/mz may indicate lowered nutritional reserves (21,22,32-34).
For the screening tools as well as for the in-depth nutritional assessment protocol described below, detailed BMI zones were developed to determine the four PEM risk levels in the elderly, because no additional documented information on BMI zones was available at that time. The percentage of ideal weight was used as a reference (30) to develop the BMI zones. The ideal weight was considered to be the average of the healthy BMI zone in the elderly (24-29 kg/mr (33)), namely 26.5. Thus, 80-90% of this ideal weight (BMI 2123.9 kg/m^sup 2^) was considered a mild nutritional risk; 70-79% of the ideal weight (BMI 18.4-20.9 kg/m2) was associated with a moderate PEM risk; and, finally, 69% or less of the ideal weight (BMI 18.3 kg/m2 or less) was considered a severe nutritional risk. These BMI zones correspond roughly with the 25th, tenth, and fifth percentiles documented by Kubena et al. (35), and are similar to the BMI zones included in the Mini Nutritional Assessment tool (22).
In-depth assessment of nutritional status: A comprehensive evaluation of subjects' nutritional status was the standard for assessing the screening tool's validity. Two dietitians conducted the assessments within 24 hours of screening. The full nutrition assessment included the use of anthropometric and biochemical indicators, taking a dietary history, and conducting a physical exam (Table 2). Before the data collection, the dietitians participated in a reliability study on anthropometric measures, dietary history-taking, and physical exams.
Weight was measured using regularly calibrated chair scales. Height was determined by measuring knee height, using the method developed by Chumlea et al. (42) and a Ross caliper (Ross Laboratories, Columbus, OH). Skinfold, were measured using a Harpenden caliper (Hemco Corporation, Holland, MI). The arm circumference measurement used to calculate the mid-upper-arm muscle area (MAMA) was determined with a flexible, non-stretch tape measure. The measures were taken according to the methods documented by Gibson (38), and interpreted with the data compiled by Frisancho (43,44). Blood samples for testing (30,36,37) were taken according to CRH laboratory methods The dietary assessment was based on a modified dietary history developed by Staveren et al. 1985, and described by Gibson, 1993 (38). Protein intake below 80% of Recommended Nutrient Intakes for Canadians (RNIs) (39) was considered a risk for dietary deficiency among the hospitalized or institutionalized population (40); for the purposes o this study, energy intake below 80% of RNIs was also a deficiency risk.
For each subject, the dietitians diagnosed PEM when at least four nutritional indicators were abnormal, including
1. two anthropometric and two biochemical indicators;
2. two anthropometric, one biochemical, and one dietary indicator; or
3. one anthropometric, two biochemical, and one dietary indicator.
After PEM was confirmed, the dietitians ranked each subject as being at a mild, moderate, or severe PEM level, determined according to the data obtained with anthropometric and biochemical indicators. However, if subjects diagnosed as malnourished showed physical signs of PEM, they were classified at the severe level.
Data analysis: Subjects' energy and protein intakes were evaluated using Nutrient Analysis Program software (Version 4.0.5, 1994, Warwick E, Cornwall, PE). Systat for Windows software (Version 7.0.1, 1997, SYSTAT Inc., registered trademark of SPSS Inc.) was used for the various statistical analyses, including a logistic backward stepwise regression for all subjects. The regression identified significant (p <= 0.05) risk factors to be used in a simplified version of the initial screening tool.
The augmented model of regression was defined as follows: manifestational variable = constant + explanatory variables. The manifestational variable was the presence
of PEM determined by the dietitians' assessments. The explanatory variables were all risk factors included in the initial tool. Contingency tables on the classifications of the screening tools and nutritional assessments were developed to identify the true positives, true negatives, false positives, and false negatives, and to calculate the sensitivity, specificity, and overall PV (13,14). Finally, a receiver operating characteristic (ROC) (45) curve was drawn for each simple screening tool developed, although only the ROC curve for simple screening tool #1 is presented. This curve determined the screening tool cutoff point that provided the best sensitivity and specificity values and thus the best validity.
RESULTS
Subjects' Characteristics
Table 3 shows the study subjects' characteristics. The dietitians' assessments show that 61.2% of LTCE subjects had a BMI below 24 kg/m^sup 2^, compared with 28% of ACE subjects. A large percentage of subjects in each category showed weight loss (42.9% to 57.9%, depending on the category). The albumin level was 35 g/L or lower in 79.6% of LTCE subjects, compared with 70.2% of ACE subjects and 50% of ACA subjects.
Prevalence of PEM
Table 3 also shows the prevalence of PEM. The dietitians' assessments indicated that 29.6% of hospitalized adults (ACA category) were malnourished at mild to severe levels. The ACA subjects mainly displayed mild PEM (18.5%). Of the elderly people admitted to the hospital (ACE category), 40.3% were malnourished at mild to severe levels. Of these, 19.3% displayed severe PEM. Finally, 61.2% of the elderly subjects in the LTCE category were found to have PEM; 34.7% were severely malnourished.
Simplification of the Initial PEM Screening Tool
The logistic backward stepwise regression revealed four significant risk factors that best identified subjects' nutritional status: BMI (p = 0.000), diagnoses/medical conditions associated with PEM (p = 0.006), albumin level (p = 0.025), and percentage of weight loss over time (p = 0.032). These four risk factors explained 53% of the variance in PEM existence. However, for practical reasons the diagnoses/medical PEM-associated conditions that required an extensive chart review were not included in the final screens. Thus, a simple screening tool (#1) comprising BMI and percentage of weight loss over time was evaluated.
Simple screening tool #1 consists of anthropometric risk factors only, due to the lack of serum albumin measurements at admission to health-care facilities. However, there might be situations in which the weight history cannot be known. In such situations, laboratory data such as albumin level would be useful for screening. Thus, a second simple screening tool (#2), comprising BMI and albumin level, was also tested (Table 4). To simplify the screening tools, the four levels of nutritional risk classification were reduced to two: low (none or mild) and high (moderate or severe) PEM risk. For comparison purposes, the number of PEM levels in the in-depth nutritional assessment protocol was reduced as well.
According to the ROC curves drawn for each simple screen, a score of two points provided the best validity for both tools. Figure 1 shows the ROC curve for simple screening tool #1.
The validity of a screening tool comprising the three risk factors (BMI, weight loss, and albumin level) was assessed. The ROC curve showed that for this tool, the best validity measures were obtained with a total score of three points (instead of two). In this case, most of the validity results were similar to those obtained with the two previous simple screens. The authors' clinical and research experience indicated that it was not necessary to use the three risk factors for a valid screening, as similar validity results were obtained with either of the two simple tools.
Validity of Simple PEM Screening Tools
Table 5 presents the validity results of each simple screening tool. In the ACA category, both screening tools had high specificity and overall PV (75.9% or higher), but low sensitivity. In the elderly population, those hospitalized for acute care can be screened efficiently with BMI and albumin level (tool #2) because validity indices are greater than 80%. However, if albumin level is not available upon admission, screening with BMI and percentage weight loss (tool #1) has good validity results (validity indices of 76.7% or higher). Finally, in the LTCE category, tool #1 provides the most valid results (validity indices of 75.9% or higher). Because of its low specificity (44.8%), screening with BMI and albumin level (tool #2) is not as valid in long-term care facilities.
DISCUSSION
The prevalence of PEM in this study is similar to that documented previously. There was a high prevalence of malnutrition, particularly among the elderly. This finding confirms the need for efficient PEM screening in health-care facilities.
The four risk factors significantly associated with PEM have also been identified in other studies. In a validation of the Subjective Global Assessment of Nutritional Status, Hirsch et al. (46) revealed that weight loss and underlying illness had the greatest influence on the final nutritional classification. One of the NSI tools identified the nutritional risk based on BMI or weight loss over time (21). Elsewhere, weight loss and albumin level have been identified as the best indicators of hospital readmission one month after discharge (47), and these factors were correlated with the subsequent risk of complications (48).
In the current initial screen, the lack of other significant risk factors might be explained by colinearity between some of them. For example, the "loss of appetite" risk factor was not identified as significant; this can be explained by the "BMI" risk factor, which accounted for much of this variance. However, the colinearity between the risk factors was not assessed.
Although the simple tools developed had high specificity and overall PV values for the ACA category, they are not sensitive enough. Many hospitalized adults won't be identified as being at high PEM risk when they are at risk. This might be explained by the low number of true positives and by the group's heterogeneity. Perhaps other risk factors should be considered to screen this population more sensitively. One possibility might be to combine albumin level with weight loss instead of with BMI, as is the case with some screening equations (8,18). However, these tools require further validity assessment, even for general hospitalized adults.
The study did enable the development of simple PEM screening tools valid for the elderly in health-care facilities. If albumin level is available upon a patient's admission to an acute care facility, it should be used in combination with BMI for screening, i.e., tool #2 should be used. However, there is also good validity when anthropometric risk factors, i.e., those measured with tool #1, are used to screen the elderly hospitalized for acute care. In long-term care facilities, a valid nutrition screening should be based on anthropometric risk factors (tool #1). Screening with tool #1 will compensate for the lack of albumin levels in many Canadian nursing homes and even in some hospitals.
Anthropometric data included in the simple screening tools were often based on reported measurements; if weight and height were measured, measurements often were less accurate than those used for the nutrition assessment. Despite this, the simple screening tools indicated good validity indices when compared with the dietitians' assessments.
Finally, as Beck and Ovesen have recommended, the BMI zones in these simple tools are adapted for the elderly and any patient with weight loss is considered for screening (49).
The study has certain limitations. Although the MAMA equation that corrects for bone was used, it is not recommended for the elderly, as the correction factors are based on young adults with higher bone density. Second, although the two dietitians received training and participated in a reliability study before data collection, no intraclass correlation was completed for them. Third, the two simple tools include risk factors that were used as nutritional indicators in the dietitians' assessments (BMI, weight loss, and albumin level). This last limitation, plus the fact that the simple screens were validated with the sample of subjects used to identify significant risk factors, might falsely increase the values of the validity indices. In future research, screening tools should be validated with a different sample of subjects, and the tools' convergent validity should be assessed to confirm their usefulness. The tools' reliability should also be assessed. This second step of the study has already begun.
RELEVANCE TO PRACTICE
This study contributed to the development of simple and valid nutrition screening tools that will help dietitians prioritize their involvement in patient care. Given that it is not always possible for dietitians to screen all patients, the nursing staff can gather screening data upon patients' admission. Information about height, actual (measured) weight, usual weight, weight loss, and time frame can be included on the nursing admission assessment form. The data can be sent to the nutrition service by computer (if available), and be used to determine the total score.
The high prevalence of malnutrition in health-care facilities means that dietitians' availability has to be examined if appropriate nutritional assessments and treatments are to be provided. To ensure a positive impact, nutrition screening and care programs also need to be implemented in the community. These will help prevent malnutrition before admission, ensure continuity of care after hospital discharge, improve nutrition outcomes among the elderly, and perhaps even reduce the length of stay. The result could be reduced health-care costs.
Acknowledgments
This study was funded by the Medical Research Fund of New Brunswick.
[Reference]
References
[Reference]
1. McWhirter JP, Pennington CR Incidence and recognition of malnutrition in hospital. Br Med J 1994;308:945-8.
2. Coats KG, Morgan SL, Bartolucci AA, et al. Hospital-associated malnutrition: a reevaluation 12 years later. J Am Diet Assoc 1993;93:27-33.
3. Naber HJT, Schermer T, De Bree A, et al. Prevalence of malnutrition in nonsurgical hospitalized patients and its association with disease
complications. Am J Clin Nutr 1997;66:1232-39.
4. Mowe M, Bohmer T. The prevalence of undiagnosed protein-calorie undemutrition in a population of hospitalized elderly patients. J Am Geriatr Soc 1991;39:1089-92.
[Reference]
5. Rudman D, Feller AG. Protein-calorie undernutrition in the nursing home. J Am Geriatr Soc 1989;37:173-83.
6. McC(ave SA, Mitoraj TE, Thielmeier KA, et al. Differentiating subtypes (hypoalbuminemic vs marasmic) of protein-calorie malnutrition: incidence and clinical significance in a university hospital setting. J Parent Ent Nutr 1992;16:337-42.
7. Sullivan D, Lipschitz D. Evaluating and treating nutritional problems in older patients. Clin Geriatr Med 1997;13:753-68.
8. The Veterans Affairs Total Parenteral Nutrition Cooperative Study Groups. Perioperative total parenteral nutrition in surgical patients. N Engl J Med 1991;325:525-32.
[Reference]
9. Larsson J, Unosson M, Nilsson S, et al. Effect of dietary supplement on nutritional status and clinical outcome in 501 geriatric patients a randomised study. Clin Nutr 1990;9:179-84.
10. Keele AM, Bray MJ, Emery PW, et al. Two phase randomised controlled
[Reference]
clinical trial of postoperative oral dietary supplements in surgical patients. Gut 1997;40:393-9.
11. American Society for Parenteral and Enteral Nutrition (ASPEN) Board of Directors. Standards for nutrition support: hospitalized patients. Nutr Clin Pract 1995;10:208-18.
12. Reuben DB, Greendale GA, Harrison GG. Nutrition screening in older persons. J Am Geriatr Soc 1995;43:415-25.
13. Gibson RS. Principles of nutritional assessment. New York, NY: Oxford University Press, 1990:11-12.
[Reference]
14. Dempsey DT, Mullen JL Pronostic value of nutritional indices. J Parent Ent Nutr 1987;11:217-22.
15. Nagel MR Nutrition screening: identifying patients at risk for malnutrition. Nutr Clin Pract 1993;8:171-5.
16. Morley JE, Miller DK. Malnutrition in the elderly. Hosp Pract 1992 Jul: 95, 98, 101, 105-6, 108-9, 113-4, 116.
17. Naber HJ T, De Bree A, Schermer T, et al. Specificity of indexes of malnutrition when applied to apparently healthy people: the effect of age. Am J Clin Nutr 1997;65:1721-5.
[Reference]
18. Elmore MF, Wagner DR, Knoll DM, et al. Developing an effective adult nutrition screening tool for a community hospital. J Am Diet Assoc 1994;94:1113-21.
19. Kovacevich DS, Boney AR, Braunschweig CL, et al. Nutrition risk classification: a reproducible and valid tool for nurses. Nutr Clin Pract 1997;12:20-5.
20. Reilly HM, Martineau JK, Moran A, et al. Nutritional screening evaluation and implementation of a simple Nutrition Risk Score. Clin Nutr 1995;14:269-73.
21. The Nutrition Screening Initiative (The American Academy of Family Physicians, The American Dietetic Association, The National Council of Aging Inc). Nutrition screening manual for professionals caring for older Americans. Washington, DC: Nutrition Screening Initiative, 1991:1-24.
[Reference]
22. Cohendy R, Gros T, Arnaud-Battandier F, et al. Preoperative nutritional evaluation of elderly patients: the Mini Nutritional Assessment as a practical tool. Clin Nutr 1999;18:345-8.
23. Blaum, CS, O'Neill EF, Clements KM, et al. Validity of the Minimum Data Set for assessing nutritional status in nursing home residents. Am J Clin Nutr 1997; 66:787-94.
24. Rush D. Evaluating the Nutrition Screening Initiative. Am J Public Health 1993;83:944-5.
25. The American Dietetic Association. Identifying patients at risk: ADA's definitions for nutrition screening and nutrition assessment. J Am Diet Assoc 1994;94:838-9.
[Reference]
26. Torun B, Chew F. Protein-energy malnutrition. In: Shils ME, Olson JA Shike M, eds. Modern nutrition in health and disease. Philadelphia, PA: Lea and Febiger, 1994:960.
27. Blackburn GL, Bristan BR, Maini BS, et al. Nutritional and metabolic assessment of the hospitalized patient. J Parent Ent Nutr 1977;1:11-22.
28. Travenol Laboratories, Inc, Hospital Division. Fundamentals of nutritional support. Deerfield, Illinois: Travenol Laboratories, Inc, 1985:15.
29. Fogt EJ, Bell SJ, Blackburn GL. Nutrition assessment of the elderly. In: Money JE, GLick Z, Rubenstein LZ, eds. Geriatric nutrition.
A comprehensive review. New York, NY: Raven Press, 1995:55.
[Reference]
30. Hopkins B. Assessment of nutritional status. In: Goltschlich MM, Matarese LE, Shronts EP, eds. Nutrition support dietetics. Core curriculum. Silver Spring, MD: ASPEN, 1993:15-70.
31. Baden A, Karkeck JM, Chernoff R Geriatrics. In: Goltschlich MM, Matarese LE, Shronts EP, eds. Nutrition support dietetics. Core curriculum. Silver Spring, MD: ASPEN, 1993:202.
[Reference]
32. Harris T, Cook EF, Garrison R, et al. Body mass index and mortality among nonsmoking older persons. The Framingham Heart Study. JAMA 1988;259:1520-4.
33. Committee on Diet and Health. Diet and health: implications for reducing chronic disease risk. Washington, DC: National Academy Press, 1989:564.
34. Cornoni-Huntley JC, Harris TB, Everett DF, et al. An overview of body weight of older persons, including the impact on mortality. The National Health and Nutrition Examination Survey 1 - epidemiologic follow-up study. j Clin Epidemiol 1991;44:743-53.
[Reference]
35. Kubena KS, McIntosh WA, Georghiades MB, et al. Anthropometry and health in the elderly. J Am Diet Assoc 1991;91:1402-7.
36. Gelinas MD, Dubost-Belair M, Bernier P, et al. Evaluation de l'etat nutritionnel. In: Manuel de nutrition clinique, 2' ed. Montreal, QC: Corporation professionnelle des dietetistes du Quebec, 1991: ch 1.2, annexes.
37. Simko MD, Cowell C, Gilbride JA. Nutrition assessment. A comprehensive guide for planning intervention. Gaithersburg, MD: ASPEN Publishers, Inc, 1995:215.
[Reference]
38. Gibson RS. Nutritional assessment, a laboratory manual. New York, NY: Oxford University Press, 1993:21, 62-90.
39. Health and Welfare Canada. Recommandations sur la nutrition. Rapport du Comite scientifique de revision. Ottawa, ON: Health and Welfare Canada, 1992.
[Reference]
40. Osterweil D, Wendt PF, Ferrell BA. Pressure ulcers and nutrition. In: Morley JE, Glick Z, Rubenstein LZ, eds. Geriatric nutrition.
A comprehensive review. New York, NY: Raven Press, 1995:337.
41. McDonald JE. A pocket guide to physical examination and nutritional assessment. Toronto, ON: W.B. Saunders of Canada, 1994.
42. Chumlea WC, Roche AF, Steenbauch ML Estimating stature from knee height for persons 60 to 90 years of age. J Am Geriatr Soc 1985;33:116-20.
43. Frisancho AR Anthropometric standards for the assessment of growth and nutritional status. Ann Arbor, MI: University of Michigan Press, 1990.
[Reference]
44. Frisancho AR. Anthropometric standards for the assessment of growth and nutritional status. Ann Arbor, MI: University of Michigan Press, 1993:50.
45. McNeil BJ, Hanley JA. Statistical approaches to the analysis of receiver operating characteristic (ROC) curves. Med Decis Making 1984;4:137-50.
46. Hirsch S, de Obaldia N, Petermann M, et al. Subjective global assessment of nutritional status: further validation. Nutrition 1991;7:35-7.
47. Friedmann JM, et al. Predicting early nonelective hospital readmission in nutritionally compromised older adults. Am J Clin Nutr 1997;65:1714-20.
[Reference]
48. Sullivan DH, Patch GA, Walls RC, et al. Impact of nutrition status on morbidity and mortality in a select population of geriatric rehabilitation patients. Am J Clin Nutr 1990;51:749-58.
49. Beck AM, Ovesen L. At which body mass index and degree of weight loss should hospitalized elderly patients be considered at nutritional risk? Clin Nutr 1998; 17:195-8.
[Author Affiliation]
MANON LAPORTE, MSc Nutrition, RD, CNSD, Campbellton Regional Hospital, NB;
UTA VILLALON, PhD Nutrition, RD, Ecole de nutrition et d'etudes familiales, Universite de Moncton, NB; HElENE PAYETTE, PhD, Centre de recherche de l'Institut universitaire de geriatrie de Sherbrooke, QC
[Author Affiliation]
NOTES: 1. A detailed French copy of the manuscript will be available by contacting the corresponding author. L'article est disponible en frangais. Ils'agit d'adresser votre demande 5 Manon Laporte, Restigouche Health Services Corporation: Campbellton Regional Hospital, Nutrition Service, Box 910, 189 Lily Lake Road, Campbellton, New Brunswick, Canada E3N 31-13
Phone: 506-789-5096 Fax: 506-789-5107 E-mail: laportem@nbnet.nb.ca
2. The tools described in this paper may not be used or reproduced without permission from the author.
