MalJNutr13(1): 29-44, 2007
      Validation of Nutritional Screening Tools Against
      Anthropometric and Functional Assessments Among
      Elderly People in Selangor
      SuzanaShaharandSitiSaifaHussain
      Department of Nutrition and Dietetics, Faculty of Allied Health Sciences
      Universiti Kebangsaan Malaysia, 50300 Kuala Lumpur Malaysia
                       ABSTRACT
        This cross sectional study was conducted to determine the validity of three
        screening tools, Mini Nutritional Assessment Short Form (MNA-SF), Malnu-
        trition Risk Screening Tool for Community (MRST-C) and Malnutrition Risk
        Screening Tool for Hospital (MRST-H) among elderly people at health clinics.
        The screening tools were validated against anthropometric and functional
        assessments. The anthropometric assessments that were carried out included
        body weight, height, arm span, body mass index (BMI), calf circumference
        (CC) and mid upper arm circumference (MUAC). A set of questionnaire on
        manual dexterity, muscular strength, instrumental activities daily living
        (IADL) and cognitive status was used to assess functional abilities. A total of
        156 subjects were recruited from rural (38 subjects) and urban (118 subjects)
        health clinics at Sabak Bernam and Cheras respectively. Subjects’ age ranged
        from 60 to 83 years old, with 44.2% were men and 55.8% women. The preva-
        lence of muscle wasting among the subjects assessed from MUAC and CC
        were both 7.0%. MNA-SF had the highest correlation with BMI (r = 0.497,
        p<0.001), followed by MUAC(r = 0.398, p<0.001), CC (r = 0.473, p<0.001), cog-
        nitive assessment (r = 0.229, p<0.001) and handgrip strength (r = 0.209,
        p<0.001). Whilst MRST-C had the highest correlation with IADL score (r =
        -0.320, p<0.001) and MRST-Hhadthehighestcorrelationwiththelockandkey
        test (r = -0.325, p<0.01). Sensitivity was the highest for MNA-SF (93.2%), fol-
        lowed by MRST-H (52.5%) and MRST-C (25.8%). Specificity was the highest
        for MRST-H (97.3%), followed by MRST-C (90.8%) and MNA-SF (79.4%).
        Positive predictive value (PPV) for MRST-H, MNA-SF and MRST-C was
        55.5%,18.2%and14.1%,respectively.Inconclusion,amongthescreeningtools
        beingvalidated,MNA-SFisconsideredthemostappropriatetooltobeusedin
        health clinics for identification of elderly individuals who are at high risk of
        malnutrition.
      ____________________
      Correspondence author: Assoc Prof Dr Suzana Shahar, Email: suzanas@medic.ukm.my
              30                                  Suzana S & Siji Saifa H
              INTRODUCTION                                       al screening tools involve estimation of
                                                                 food intake and anthropometric assess-
                    Mortality rates have declined in vir-        ment with several biochemical indicators
              tually all countries due to progress in pre-       that are influenced by nutritional status
              venting infectious diseases and improving          (Omran & Morley, 2000). However, some
              hygiene, sanitation and overall social             of the nutritional assessment tools, for
              development and living standards. As a             example anthropometry, dietary intake
              result,   the   average    life   expectancy       and biochemical tests may not be appro-
              throughout the world is projected to reach         priate to detect elderly who are at risk of
              72 years in 2020 (Fahey et al., 2003). This        malnutrition. This is because the methods
              aging population phenomenon exists                 are expensive, time consuming and
              worldwide, both in developing and devel-           require a trained personnel to conduct the
              oped countries (Nourhashemi et al., 2001;          assessments (Mohs, 1994).
              WHO 2002). In Malaysia, based on the                     There has been a lot of work done to
              year 2005 statistics, it is estimated that the     develop nutritional screening tools that
              percentage of elderly people age ≥ 65 was          are simple, rapid, cheap, effective and
              4.6% compared to 3.9% in 2000. It is esti-         comprehensive for elderly people. The
              mated that by the year 2050, the propor-           value of a screening tool depends on its
              tion will be increased by four-fold to 21%         sensitivity, specificity, predictive value
              (7.9  million people) (Department of               and also acceptability to both the targeted
              Statistics, 2005).                                 subjects and healthcare workers (Elia,
                    Afewstudieshaveshownthatelder-               2003). The tools should also consider
              ly people are at high risk of malnutrition         current weight status (e.g. underweight or
              (Gambert & Kassur, 1994; Jensen et al.,            obesity), as well as past and likely future
              2001). In Malaysia, several studies among          changes in weight, both of which are
              older people in the community have                 linked to food intake or appetite and
              shownthatmalnutrition still exists among           diseaseseverity(Elia,Zellipour&Stratton,
              this group of population. The prevalence           2005).
              varies from more than 2% (Sumaiyah,                      Amongthedifferent kinds of screen-
              Muhammad Tauffik & Samiah Yasmin,                  ing tools that have been used include
              2003; Zaitun et al., 2004) to almost 38%           Nutrition Screening Index (NSI) (Dwyer,
              (Suzana, Dixon &Earland,1999;Sherinaet             1994; Grinder & Costello, 1996), Malnutri-
              al., 2004; Tan, 2006).                             tion Universal Screening Tool (MUST)
                    Ahealthy diet is an important factor         (Elia,  2003), Malnutrition Risk Scale
              to ensure optimum health and functional            (SCALES) (Morley, 1989) and Mini
              capability and has a major influence               Nutritional    Assessment      Short    Form
              towards physical and functional demands            (MNA-SF) (Guigoz, Vellas, & Garry 1994
              of the well being of the elderly (Pirlich &        &1996;Vellasetal., 2000; Rubenstein et al.,
              Lochs, 2001). Nutritional problems often           2001). They have been used in community
              go unrecognised and untreated (Reilly et           populationtoscreenforelderlywhoareat
              al., 1995). It is therefore important to have      risk of malnutrition.
              screening methodsthatareabletoidentify                   The Mini Nutritional Assessment
              those who are malnourished and those               Short Form (MNA-SF) consists of 6 of the
              who are at risk of developing malnutri-            18itemsofthefullMNA(Rubensteinetal.,
              tion. Nutritional screening tools can rapid-       2001). MNA-SF has been developed in
              ly identify individuals who are at risk of         Switzerland to identify elderly who are at
              malnutrition for the purpose of further            high risk of malnutrition either in the
              nutritional assessment and intervention            hospital or the community. Although the
              (Green&Watson,2005).Specificnutrition-             MNAwasdeveloped specifically for frail
                                            Validation of nutritional screening tools among elderly                                       31
                  older people, it has been validated in a                        nutritional status of the elderly. As ade-
                  healthy older population and has been                           quate nutrition is essential towards the
                  widely used and validated in many                               well being of the elderly, it is necessary to
                  European countries. It was reported that                        have an assessment method that could
                  this screening tool has sensitivity, speci-                     identify elderly individuals who are prone
                  ficity and positive predictive value (PPV)                      to malnutrition. Screening tools provide
                  of 96%, 98% and 97% respectively in iden-                       aneconomicalandrapidmethodofidenti-
                  tifying those elderly who are at risk of                        fying those elderly who are at high risk of
                  malnutrition (Guigoz, Vellas & Garry 1994                       malnutrition. However, the screening
                  &1996;Vellasetal., 2000; Rubenstein et al.,                     tools used have to be validated to assess
                  2001). It has also been evaluated in a group                    their appropriateness on the intended
                  of Japanese frail elderly and found to have                     targeted population. This cross-sectional
                  a sensitivity and specificity of 85.9% and                      study was conducted to determine the
                  84%inidentifying under-nutrition respec-                        validity of three screening tools, namely
                  tively (Kuzuya et al., 2005).                                   the Mini Nutritional Assessment Short
                         In Malaysia, a Malnutrition Risk                         Form (MNA-SF), Malnutrition Risk
                  Screening Tool for Community (MRST-C)                           Screening Tool for Community (MRST-C)
                  specifically for the community elderly has                      and Malnutrition Risk Screening Tool for
                  been developed based on local studies                           Hospital (MRST-H) among elderly people
                  (Suzana, Dixon & Earland, 1999). The                            who attended the health clinics in an
                  MRST-C has been validated among rural                           urban and a rural area (Appendix). The
                  elderly Malays and Chinese and also insti-                      screening tools were validated against
                  tutionalised Chinese elderly at several                         anthropometric and functional assess-
                  places in Malaysia, including Kedah,                            ments, as shown in Figure 1.
                  Kelantan and Negeri Sembilan (Suzana et
                  al.,  2007). However, its usefulness in
                  detecting malnourished individuals in an                        METHOD
                  urbansettingorhealthclinicshasnotbeen
                  tested. Another screening tool, Malnutri-                               This cross sectional study was con-
                  tion Risk Screening Tool for Hospital                           ducted among elderly people who visited
                  (MRST-H) recently has been developed                            the outpatient health clinics of Klinik
                  locally and validated to identify elderly                       Kesihatan Cheras Baru (an urban area),
                  hospitalised patients who are at risk of                        Klinik Kesihatan Bagan Terap at Sabak
                  malnutrition. It is a screening tool that                       Bernam (a rural area) and at the Rumah
                  includes physical, clinical and anthropo-                       Sejahtera Day Care Centre, Cheras (an
                  metric examinations (Sakinah, 2006). As                         urban area), which is a day care centre for
                  this tool has been developed andvalidated                       non-institutionalised elderly people. Sub-
                  basedonhospitalisedelderly,italsoneeds                          jects recruited in this study included those
                  to be validated in a community setting.                         who were aged 60 years and above, were
                         In Malaysia, government health clin-                     free from physical deformation that could
                  ics provide inexpensive access to medical                       have affected the anthropometric assess-
                  services for most of the elderly. Currently,                    ments, were able to communicate and had
                  functional assessments have been includ-                        given consent. The subjects were recruited
                  ed as part of the medical assessment for                        from July to September 2006.
                  elderlypatientswhoattendselectedhealth                                  Subjects were asked to provide infor-
                  clinics which conduct the “Elderly Health                       mationonsocio-demographicandperson-
                  Program”. Besides measurements of body                          al profile such as marital status, source of
                  weightandheight,therearenonutritional                           income, level of education and job status,
                  risk assessments carried out to assess the                      through an interview. All subjects were
             32                               Suzana S & Siji Saifa H
                “NewTests”fornutritional screening            “Established Tests” of nutritional
                                                                     andfunctional status
                               MNA-SF                                  Anthropometry:
                                                                             BMI
                                                                              CC
                                                                            MUAC
                                                                          Armspan
                               MRST-C                                  Functional status:
                                                                            IADL
                                                                      Quadriceps muscle
                                                                         Lock and key
                               MRSC-H                                Cognitive assessment
                                                                      Handgrip strength
             Figure 1. Tests used in the study
             screened for malnutrition risk using three    especially in the lower limb (Sakinah,
             screeningtools,namelyMNA-SF,MRST-C            2006).
             and MRST-H(Appendix). Then they were               Functional assessment involved a
             assessed for anthropometric and function-     self reported functional disability using
             al status, as reference standards or estab-   Instrumental Activity Daily Living (IADL)
             lished tests for nutritional status assess-   (Fillenbaum et al., 1988) and cognitive
             ments. The anthropometric assessments         assessmentbasedontheElderlyCognitive
             that were carried out included body           Assessment Questionnaire (ECAQ) (Kua
             weight and height (Fidanza & Keller,          & Ko, 1992). Manual dexterity was also
             1991), arm span (Kwok&Whitelaw,1991),         assessed using Lock and Key test (Manan-
             calf circumference (CC) (Chumlea, Guo &       dhar, 1995). Climbing stairs to assess
             Vellas, 1994) and mid upper arm circum-       quadriceps muscle strength (Bennet, 1999)
             ference (MUAC) (Ferro-Luzzi & James,          and handgrip strength using Hand Dyna-
             1996). Body mass index (BMI) was calcu-       nometer (Hillman et al., 2005) were also
             lated from measured height and estimated      conducted as functional assessment. The
             height from arm span for those with           whole process of data collection took
             kyphosis(Suzana&Ng,2003).MUACand              about 40 minutes to complete.
             CCare parameters used for measurement              Data analysis was analysed using
             of muscle mass and subcutaneous adipose       “Statistical Package for the Social Sciences
             tissue (Woods & Moshang, 2005) and a          12.0” (SPSS version 12.0). Unpaired t test
             low MUAC among the elderly has been           wasusedtodifferentiate between sex, age
             showntoincreaseriskofmortality(Tajima         group and locations (urban and rural) for
             et al., 2004). A MUAC value of less than      numerical data. Chi squared test was used
             23.0 cm for men and 22.0 cm for women         to assess the differences between sex, for
             indicates loss of peripheral muscle mass      factors on demography and psychosocial
             (Ferro-Luzzi & James, 1996). As for CC, a     factors, and functional status that are cate-
             valueoflessthan30.1cmformenand27.3            gorical data. Correlation test was used to
             cm for women will indicate muscle loss,       assess the nutritional screening tools