You are on page 1of 7

Original Communication

Journal of Parenteral and Enteral


Nutrition
Nutrition Screening Tools: An Analysis of the Evidence Volume 36 Number 3
May 2012 292-298
© 2012 American Society
for Parenteral and Enteral Nutrition
DOI: 10.1177/0148607111414023
Annalynn Skipper, PhD, RD, FADA1; Maree Ferguson, PhD, MBA, RD2; http://jpen.sagepub.com
hosted at
Kyle Thompson, MS, RD3; Victoria H. Castellanos, PhD, RD4; and http://online.sagepub.com
Judy Porcari, MBA, MS, RD5

Abstract
In response to questions about tools for nutrition screening, an evidence analysis project was developed to identify the most valid and
reliable nutrition screening tools for use in acute care and hospital-based ambulatory care settings. An oversight group defined nutrition
screening and literature search criteria. A trained analyst conducted structured searches of the literature for studies of nutrition screening
tools according to predetermined criteria. Eleven nutrition screening tools designed to detect undernutrition in patients in acute care and
hospital-based ambulatory care were identified. Trained analysts evaluated articles for quality using criteria specified by the American
Dietetic Association’s Evidence Analysis Library. Members of the oversight group assigned quality grades to the tools based on the quality
of the supporting evidence, including reliability and validity data. One tool, the NRS-2002, received a grade I, and 4 tools—the Simple
Two-Part Tool, the Mini-Nutritional Assessment–Short Form (MNA-SF), the Malnutrition Screening Tool (MST), and Malnutrition
Universal Screening Tool (MUST)—received a grade II. The MST was the only tool shown to be both valid and reliable for identifying
undernutrition in the settings studied. Thus, validated nutrition screening tools that are simple and easy to use are available for application
in acute care and hospital-based ambulatory care settings. (JPEN J Parenter Enteral Nutr. 2012;36:292-298)

Keywords
administration; adult; geriatrics; home nutrition support; nutrition assessment; outcomes research/quality

Clinical Relevancy Statement the process we now term nutrition screening. Potosnak and
Chudnow4 added laboratory data and information about appetite
The finding that valid and reliable nutrition screening tools are and weight change over time to the first screening tool by Seltzer
available is expected to result in accurate identification of et al. The Nutrition Screening Initiative developed tools that
nutrition problems. If these tools are quick and easy to use, it is have not been independently validated.5,6 In 1994, the American
possible that they will reduce the amount of time required for Dietetic Association (ADA) recommended more than 60 nutri-
nutrition screening. If accurate nutrition screening can be com- tion screening criteria that were derived from selected litera-
pleted with fewer criteria and in less time, resources saved may ture.7 The ADA recommendations did not specify how these
be reallocated toward higher priority nutrition activities. criteria were to be used or whether they were valid indicators of
nutrition problems. In many U.S. hospitals, dietitians recom-
mend lengthy nutrition screening tools, which have not been
Introduction tested for reliability, validity, or accuracy.8 However, outside the
Screening for nutrition problems is a Medicare Condition of United States, accurate, streamlined nutrition screening tools
Participation and therefore a requirement for accreditation or that have been validated against a complete nutrition assessment
certification of U.S. healthcare facilities.1 Nutrition screening are widely used.9 During a nutrition screening workshop in
is also a primary mechanism for patients to be referred to a 2006, RDs from the United States, Europe, Australia, and the
registered dietitian (RD) for further nutrition assessment, diag-
nosis, and intervention.2 Thus, valid and reliable nutrition From 1Annalynn Skipper and Associates, Oak Park, Illinois; 2Princess
screening tools are needed to ensure that referrals are appropri- Alexandra Hospital, Brisbane, Australia; 3Appalachian State University,
Boone, North Carolina; 4Florida International University, Miami, Florida;
ate. RDs are responsible for developing nutrition screening
and 5North Shore University Hospital, Manhasset, NY.
criteria and supervising the nutrition screening process. How-
ever, nutrition screening is typically conducted by others. Financial disclosure: none declared.
Nutrition screening has become increasingly complex since Received for publication December 31, 2010; accepted for publication
it was first described by Seltzer and colleagues in 1979.3 These March 23, 2011.
authors describe the process of using albumin and total lympho- Corresponding Author: Annalynn Skipper, PhD, RD, FADA, Annalynn
cyte count to identify patients on hospital admission who require Skipper and Associates, PO Box 45, Oak Park, IL 60303; e-mail:
nutrition repletion as instant nutrition assessment; it resembles Annalynn_Skipper@comcast.net.
Nutrition Screening Tools / Skipper et al 293

Table 1.  Definition and Key Considerations in Nutrition Venrooij et al.12 Thus, acceptable reference standards for vali-
Screening dating a nutrition screening tool were the Subjective Global
Assessment (SGA), the Patient-Generated Subjective Global
Definition:
Assessment (PG-SGA), the Mini-Nutritional Assessment
Nutrition screening is the process of identifying patients,
(MNA), and dietitian assessments that included a measure of
clients, or groups who may have a nutrition diagnosis and
benefit from nutrition assessment and intervention by a body composition and change in that measure over time.13-15
registered dietitian (RD).
Key considerations:
Literature Search
May be conducted in any practice setting as appropriate
Tools should be quick, easy to use, valid, and reliable for the A trained and experienced lead analyst searched the English lit-
patient population or setting erature using the PubMed search engine as specified by the
Tools and parameters are established by RDs, but the screening workgroup. The dates of the literature search for nutrition
process may be carried out by registered dietetic technicians screening tools were 1997 to 2007. The search included acute
and other trained personnel care, community-based care, community dwelling, and the
Nutrition screening and rescreening should occur within an names of specific nutrition screening tools: MNA, MUST,
appropriate time frame for the setting DETERMINE checklist, Short Nutritional Assessment Ques-
tionnaire (SNAQ), meals on wheels, nutrition screening tools/
instruments, SGA, prognostic nutrition index (PNI), nutritional
risk index (NRI), prognostic inflammatory and nutritional index
Middle East began to discuss how to identify and implement (PINI), nutrition screening questionnaire, nursing nutrition
efficient, valid, and reliable nutrition screening tools.10 This screen, mass nutrition screening, nutrition scoring system, nurs-
group proposed an evidence analysis project to identify the most ing nutritional screening tool, nutrition assessment tool,
reliable and valid tools for nutrition screening in acute care and SCREEN II, sensitivity and specificity, diet surveys, and nutri-
hospital-based ambulatory care settings. tion surveys. The reference lists of review articles obtained were
searched by hand for additional articles, and workgroup mem-
bers also searched their personal files. A total of 168 articles
Methods describing 11 nutrition screening tools were identified.
A workgroup of RDs with expertise in research, clinical nutri- Articles were reviewed by the workgroup, and those that
tion management, and clinical practice provided oversight to met the inclusion criteria were graded for quality (positive,
the project. The workgroup began by establishing definitions negative, neutral) by trained analysts using criteria specified
to be used in the project. Workgroup members viewed nutrition by the ADA’s Evidence Analysis Library.16 The work of ana-
screening as separate and distinct from nutrition assessment, lysts was reviewed by the lead analyst, and agreement was
although the terms are often used interchangeably in the litera- negotiated in situations where the analyst and lead analyst dis-
ture and in practice. The workgroup also wanted to focus spe- agreed on quality rankings. The workgroup graded the evi-
cifically on nutrition screening and developed the definition of dence supporting the tools on a scale of I to IV, where I = good,
nutrition screening in Table 1, which was subsequently II = fair, III = limited, IV = expert opinion, and V = grade not
approved for use by the ADA’s Quality Management Commit- assignable. Articles describing studies of tools that received
tee.11 Because screening tools must identify a nutrition prob- grades of I and II were then reviewed by the workgroup for
lem to be valid, the workgroup also defined undernutrition as validity and reliability data.18 Validity was determined by mea-
the nutrition problem. suring sensitivity and specificity. Sensitivity is the percentage
Articles describing specific screening tools were included if of undernourished individuals correctly identified by the
the tools were (1) simple and reports indicated a completion screening tool as being at risk (true positive/true positive +
time of less than 10 minutes, (2) administered by a healthcare false negative). Specificity is the percentage of well-nourished
professional, (3) tested against an acceptable reference stan- individuals correctly identified by the screening tool as being
dard for validity, (4) tested on adults older than age 18, (5) not at risk (true negative/true negative + false positive).
tested in an acute or hospital-based ambulatory care setting, Reliability was measured by the κ value, a measure of interra-
and (6) if the authors reported the validity and reliability of the ter agreement, with a value of 0 indicating no agreement and 1
nutrition screening tool or sufficient data for reliability and indicating perfect agreement.
validity to be calculated. Articles were excluded if they
reported the results of trials with ≤20 participants.
The definition of the nutrition problem, undernutrition, was
Results and Discussion
defined as a measure of body composition at a given time and Eleven nutrition screening tools were evaluated for their valid-
change in body composition over time as described by van ity and reliability to identify nutrition problems of patients in
294
Table 2.  Results of Evidence Analysis of Nutrition Screening Tools

Positive Negative
Screening Reference Sample Sensitivity, Specificity, Predictive Predictive
Tool Grade Population Mean Age, y Standard Size % % Reliability Value, % Value, % Agreement
18
NRS-2002 I Medical-surgi- LOS 1–10 days: 50.5 ± SGA 995 62 93 NA 85* 79* κ = 0.48 (agree-
cal hospital- 21.9; >11 days: 65.4 ± ment among
ized 18.7; LOS unknown: methods)
44.4 ± 17.0
NRS-200219 Acute hospital- 80.2 ± 7.7 MNA 80 39 83 NA 85 37 84.6%
ized
NRS-200219 Acute hospital- 80.2 ± 7.7 SGA 120 70 85 NA 79 78 79.2%
ized
MNA-SF20 II Ambulatory 72 (range, 60–98) MNA 408 100 69.5 NA 92.8 78.5 NA
MNA-SF21 Subacute and France: 78.3 MNA 904 97.9 100 NA 98.7 NA NA
ambulatory Spain: 75.8
New Mexico: 76.8
MST22 II Acute hospital- 54.3 ± 14.8 SGA 2211 74 76 NA 27.9 96 NA
ized
MST23 Acute hospital- 57.7 ± 16.5 (range, SGA 408 93 93 κ = 0.84–0.93 98 73 NA
ized 19–74) (P < .01)
MST24 Oncology out- 59.1 ± 13.8 PG-SGA 50 100 92 κ = 0.83– 80 100 NA
patients 0.88;
N = 20
MST9 Oncology out- 59.9 ± 13.5 (range, SGA 106 100 81 NA 40 100 NA
patients 15–89)
MUST25 II Medical 45 ± 13.9 SGA 50 NA NA NA NA NA 92% (κ = 0.783)
MUST18 Medical-surgi- LOS 1–10 days: 50.5 ± SGA 995 61 76 NA 65 76 κ = 0.26 (agree-
cal hospital- 21.9; >11 days: 65.4 ± ment between
ized 18.7; LOS unknown: methods)
44.4 ± 17.0
NST/BAPEN26 II Acute hospital- 59.6 ± 19.4 (study 1); Dietitian as- 166 86* 95* κ = 0.66 80* 96* Pilot phase:
ized 59.6 ± 16 (study 2); sessment κ = 0.72 (CI,
59.6 ± 16.9 (study 3) 0.59–0.84)
Simple Two- II Acute hospital- 54.3 ± 14.8 SGA 2211 63 97 NA 70.2 95.5 NA
Part Tool22 ized
NRS27 III Acute care, 55.9 (range, 8.5 mo to Dietitian as- 153 NA NA 74% NA NA 93%
medical- 92 y) sessment
surgical

(continued)
Table 2.  (Continued)

Positive Negative
Screening Reference Sample Sensitivity, Specificity, Predictive Predictive
Tool Grade Population Mean Age, y Standard Size % % Reliability Value, % Value, % Agreement
SCREEN-II III Seniors in com- Elderly; 98% >64 y Dietitian as- 193 84 58 Study 2, (N = 83 59 NA
AB28 munity and sessment 149); test-
geriatrician’s retest reli-
clinic ability ICC
= 0.84,
95% CI
Study 3 (N
= 97) inter-
rater ICC =
0.79; intra-
rater ICC =
0.85
Rapid Screen29 III Subacute care >65 Standardized 65 78.6 97.3 NA 95.7* 85.7* 35.4% malnour-
(rehabilitation nutrition ished; 64.6%
center) assessment nourished
Tool #130 III ACE; LTCE ACE: 78.9 ± 7.1; LTCE: Dietitian as- ACE: 72; ACE: 75; ACE: 75; IRR diet NA NA NA
79.5 ± 7.1 sessment LTCE: 70; LTCE: LTCE: tech and
total: 142 63; total 63; total RN: 85%
sample: 70 sample: 70 ACE; 80%
LTCE; all
sample κ
= 0.60 ±
0.07; test-
retest 78%
ACE; 82%
LTCE;
overall κ
= 0.60 ±
0.07

ACE, elderly in acute care; CI, confidence interval; ICC, intraclass correlation coefficient; IRR, interrater reliability; LTCE, elderly in long-term care; LOS, length of stay; MNA,
Mini-Nutritional Assessment; MNA-SF, Mini-Nutritional Assessment–Short Form; MST, Malnutrition Screening Tool; MUST, Malnutrition Universal Screening Tool; NA, not
available; NRS, Nutritional Risk Screening; NST/BAPEN4, Nutrition Screening Tool/British Association of Parenteral and Enteral Nutrition; PG-SGA, Patient-Generated
Subjective Global Assessment; RN, registered nurse; SCREEN-II AB, Seniors in the Community: Risk Evaluation for Eating and Nutrition, Version II, abbreviated version; SOA,
subjective global assessment.
*Values calculated by a work group member.

295
296
Table 3.  Data Required to Complete Nutrition Screening Tools Evaluated in the Nutrition Screening Evidence Analysis Project

Tools

NST/ Simple Two- SCREEN-II Rapid


Criteria NRS-200219 MNA-SF20 MST22 BAPEN426 MUST24 Part Tool22 NRS27 AB29 Screen29 Tool #130

Recent unintentional weight loss x x x x x x x x x


Appetite x x x x x
Body mass index x x x x x
Disease severity x x x x
Age >70 y x
Weight x
Height x
Weight gain or loss x
Subcutaneous fat loss x
Impaired general condition x
Housebound x
Meal preparation habits and eating alone x
Dementia or depression x
Food intake or eating problem; skipping meals x x x x
Ability to eat and retain food x
Intake of fluid/fruits and vegetables x

MNA-SF, Mini-Nutritional Assessment–Short Form; MST, Malnutrition Screening Tool; MUST, Malnutrition Universal Screening Tool; NRS, Nutritional Risk Screening; NST/
BAPEN4, Nutrition Screening Tool/British Association of Parenteral and Enteral Nutrition; SCREEN-II AB, Seniors in the Community: Risk Evaluation for Eating and Nutrition,
Version II, abbreviated version.
Nutrition Screening Tools / Skipper et al 297

acute care and hospital-based ambulatory care settings. Results Clinicians selecting screening tools may wish to consider
of the analysis of the evidence supporting these 11 tools are the availability and quality of evidence supporting a particular
summarized in Table 2. A single tool, the NRS-2002, received screening tool as well as the availability of data required to use
a grade I, 4 tools received a grade II, and the remainder received the tool. Table 3 contains the data needed for each of the tools
a grade III. The SNAQ tool was not evaluated against an evaluated. Tools that are simple, quick, and easily completed
acceptable reference standard to determine validity and was by nonprofessionally trained staff such as administration staff,
given a grade V. Tools with grade I and II evidence in the high- patients, or family members are preferred over tools requiring
est quartile for sensitivity (>83%) and specificity (>90%) were calculations such as body mass index and percentage weight
the MNA–Short Form (SF), which had a sensitivity and speci- loss. Also, tools selected for clinical use should have been
ficity greater than >90% in 1 of 2 studies, and the MST, with a tested in patient populations similar to the ones where they will
sensitivity >90% in 3 of 4 studies and a specificity >90% in 2 be applied (Table 2). Reliable and valid nutrition screening
of 4 studies.20,23 The MST had a κ score of 0.83–0.88. No data tools have been developed and should be further tested in a
were available to evaluate the reliability of the MNA-SF. Thus, broader range of patients. When the results of further studies
the MST was the only tool shown to be both valid and reliable are made available, the grades of evidence supporting each of
for identifying nutrition problems in the settings studied. the tools evaluated may change. In the interim, selection of
simple tools that are reliable and valid may improve compli-
ance in completing nutrition screening and may also free up
Further Research time that may be reallocated toward nutrition intervention.
Some studies reviewed and excluded were limited by small sam-
ple size and the use of weak research methodologies. A major References
issue in validating nutrition screening tools is clear identification
  1. Medicare and Medicaid Programs. Hospital conditions of participation: quality
of an acceptable reference standard for nutrition problems. Stan- assessment and performance improvement. Fed Regist. 2003;68:3435-3455.
dard definitions of nutrition problems are available in the Inter-   2. Writing Group of the Nutrition Care Process/Standardized Language Com-
national Dietetics and Nutrition Terminology, and an mittee. Nutrition care process and model part I: the 2008 update. J Am Diet
international workgroup is working to further refine the specific Assoc. 2008;108:1113-1118.
criteria for malnutrition.31,32 Future studies validating nutrition   3. Seltzer MH, Bastidas JA, Cooper DM, Engler P, Slocum B, Fletcher HS.
Instant nutritional assessment. JPEN J Parenter Enteral Nutr. 1979;3:157-159.
screening tools should incorporate these standard definitions.   4. Potosnak L, Chudnow LP. S. A simple tool for identifying patients at nutri-
The ideal screening tool would be 100% sensitive and specific. tional risk. Quality Rev Bull. 1983;9:81-83.
As this is generally not achievable, and there is almost always   5. White JV, Ham RJ, Lipschitz DA, Dwyer JT, Wellman NS. Consensus of
some misclassification, the need to correctly classify all patients the Nutrition Screening Initiative: risk factors and indicators of poor nutri-
who are malnourished (sensitivity) takes precedence over mis- tional status in older Americans. J Am Diet Assoc. 1991;91:783-787.
  6. Posner BM, Jette AM, Smith KW, Miller DR. Nutrition and health risks
classifying well-nourished patients (specificity). Reliability in the elderly: the Nutrition Screening Initiative. Am J Public Health.
should also be determined to ensure there is consistency in the 1993;83:972-978.
use of the tool between raters. Future studies should adhere to   7. Council on Practice Quality Management Committee. ADA’s definitions
international guidelines for reporting research to facilitate com- for nutrition screening and nutrition assessment. J Am Diet Assoc. 1994;94:
parison of findings from this study with new studies.33 838-839.
  8. Chima CS, Dietz-Seher C, Kushner-Benson S. Nutrition risk screening in
acute care: a survey of practice. Nutr Clin Pract. 2008;23:417-423.
Application and Conclusions   9. Ferguson ML, Capra S, Bauer J, Banks M. Development of a valid and
reliable malnutrition screening tool for adult acute hospital patients. Nutri-
Clearly, simple nutrition screening tools are valid identifiers of tion. 1999;15:458-464.
nutrition problems of patients in acute care and hospital-based 10. Annonymous. The Mini Nutritional Assessment. Proceedings of an IANA
(International Academy on Nutrition and Aging) workshop. J Nutr Health
ambulatory care. Reliability testing for nutrition screening tools Aging. 2006;10:455-560.
is needed as only the MST and British Association for Parenteral 11. Definitions and criteria. Nutrition Screening Evidence Analysis Project
and Enteral Nutrition (BAPEN)–II were tested for reliability. American Dietetic Association. https://www.adaevidencelibrary.com/
Also, there are opportunities for researchers to test nutrition topic.cfm?cat=3958. Accessed November 11, 2010.
screening tools in settings such as rehabilitation facilities, long- 12. van Venrooij LMW, de Vos R, Borgmeijer-Hoelen MMMJ, Haaring C,
deMol BAJM. Preoperative unintended weight loss and low body mass
term acute care, and long-term care facilities. The MNA-SF was index in relation to complications and length of stay after cardiac surgery.
developed in elderly populations, and further testing for reliabil- Am J Clin Nutr. 2008;87:1656-1661.
ity and validity in younger populations is recommended. The 13. Detsky AS, McLaughlin JR, Baker JP, et al. What is subjective global
most reliable and valid tools could be tested in the pediatric assessment of nutritional status? JPEN J Parenter Enteral Nutr. 1987;11:8-13.
population where valid and reliable nutrition screening tools 14. Ottery FD. Rethinking nutritional support of the cancer patient: the new field
of nutritional oncology. Semin Oncol. 1994;21:770-778.
have not been developed. Given the number of nutrition screen- 15. Guigoz Y, Vellas B, Garry PJ. Mini nutritional assessment: a practical
ing tools described in the literature, further research directed assessment tool for grading the nutritional state of elderly patients. Facts
toward developing new screening tools is not recommended. Res Gerontol. 1994;2(suppl):15-59.
298 Journal of Parenteral and Enteral Nutrition 36(3)

16. American Dietetic Association. ADA Evidence Analysis Manual. Ed IIIc. 2005. 25. Stratton RJ, Hackston A, Longmore D, et al. Malnutrition in hospital out-
http://www.adaevidencelibrary.com/topic.cfm?cat=1155&library=EAL. patients and inpatients: prevalance, concurrent validity and ease of use in
Accessed November 11, 2010. the ‘malnutrition universal screening tool’ (‘MUST’) for adults. Br J Nutr.
17. Altman DG, Bland JM. Statistics notes: diagnostic tests 1: sensitivity and 2004;92:799-808.
specificity. BMJ. 1994;308:1552. 26. Weekes CE, Elia M, Emery PW. The development, validation and reli-
18. Kyle UG, Kossovsky MP, Karsegard VL, Pichard C. Comparison of tools ability of a nutrition screening tool based on the recommendations of the
for nutritional assessment and screening at hospital admission: a popula- British Association for Parenteral and Enteral Nutrition (BAPEN). Clin
tion study. Clin Nutr. 2006;25:409-417. Nutr. 2004;23:1104-1112.
19. Bauer JM, Vogl T, Wicklein S, Trögner J, Mühlberg W, Sieber CC. Com- 27. Reilly HM, Martineau JK, Moran A, Kennedy H. Nutritional screening:
parison of the Mini Nutritional Assessment, Subjective Global Assess- evaluation and implementation of a simple Nutrition Risk Score. Clin Nutr.
ment, and Nutritional Risk Screening (NRS 2002) for nutritional screening 1995;14:269-273.
and assessment in geriatric hospital patients. Z Gerontol Geriatr. 2005;38: 28. Keller HH, Goy R, Kane SL. Validity and reliability of SCREEN II
322-327. (Seniors in the Community: Risk Evaluation for Eating and Nutrition, Ver-
20. Cohendy R, Rubenstein LZ, Eledjam JJ. The Mini Nutritional Assessment– sion II). Eur J Clin Nutr. 2005;59:1149-1157.
Short Form for preoperative nutritional evaluation of elderly patients. 29. Visvanathan R, Penhall R, Chapman I. Nutritional screening of older peo-
Aging Clin Exp Res. 2001;13:293-297. ple in a sub-acute care facility in Australia and its relation to discharge
21. Rubenstein LZ, Harker JO, Salvà A, Guigoz Y, Vellas B. Screening for outcomes. Age Ageing. 2004;33:260-265.
undernutrition in geriatric practice: developing the Short-Form Mini- 30. Laporte M, Villalon L, Payette H. Simple nutrition screening tools for
Nutritional Assessment (MNA-SF). J Gerontol A Biol Sci Med Sci. healthcare facilities: development and validity assessment. Can J Diet
2001;56:M366-M372. Pract Res. 2001;62:26-34.
22. Nursal TZ, Noyan T, Atalay BG, Köz N, Karakayali H. Simple two-part tool 31. International Dietetics and Nutrition Terminology (INDT) Reference Man-
for screening of malnutrition. Nutrition. 2005;21:659-665. ual. 3rd ed. Chicago: American Dietetic Association; 2011.
23. Ferguson M, Bauer J, Gallagher B, Capra S, Christie DR, Mason BR. Vali- 32. Jensen GL, Mirtallo J, Compher C, et al. Adult starvation and disease-
dation of a malnutrition screening tool for patients receiving radiotherapy. related malnutrition: a proposal for etiology-based diagnosis in the clinical
Australas Radiol. 1999;43:325-327. practice setting from the International Consensus Guideline Committee.
24. Isenring EA, Bauer JD, Capra S. Nutrition support using the American JPEN J Parenter Enteral Nutr. 2010;43:156-159.
Dietetic Association medical nutrition therapy protocol for radiation oncol- 33. Altman DG, Schulz KF, Moher D, et al. The revised CONSORT statement
ogy patients improves dietary intake compared with standard practice. for reporting randomized trials: explanation and elaboration. Ann Intern
J Am Diet Assoc. 2007;107:404-412. Med. 2001;134:663-694.

You might also like