Application of Rasch Analysis to Examine Psychometric Aspects of the Activities-Specific Balance Confidence Scale When Used in a New Cultural Context

Application of Rasch Analysis to Examine Psychometric Aspects of the Activities-Specific Balance Confidence Scale When Used in a New Cultural Context
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institutionand sharing with colleagues.Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third partywebsites are prohibited.In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further informationregarding Elsevier’s archiving and manuscript policies areencouraged to visit:  Author's personal copy ORIGINAL ARTICLE  Application of Rasch Analysis to Examine PsychometricAspects of the Activities-Specific Balance Confidence ScaleWhen Used in a New Cultural Context Solveig A. Arnadottir, MSc, Lillemor Lundin-Olsson, PhD, Elin D. Gunnarsdottir, PhD, Anne G. Fisher, ScD ABSTRACT. Arnadottir SA, Lundin-Olsson L, Gunnarsdot-tir ED, Fisher AG. Application of Rasch analysis to examinepsychometric aspects of the Activities-Specific Balance Con-fidence Scale when used in a new cultural context. Arch PhysMed Rehabil 2010;91:156-63. Objective:  To investigate by using Rasch analysis the psy-chometric properties of the Activities-Specific Balance Confi-dence (ABC) Scale when applied in a new Icelandic context. Design: Cross-sectional, population-based, random selectionfrom the Icelandic National Registry. Setting:  Community-based. Participants:  Icelanders (N  183), 65 to 88 years old, and48% women. Interventions:  Not applicable. Main Outcome Measure:  ABC, an instrument used toevaluate how confident older people are in maintaining balanceand remaining steady when moving through the environment.An Icelandic translation of the ABC (ABC-ICE) scale wasevaluated by implementing Rasch rating scale analysis to trans-form ordinal ABC-ICE scores into interval measures and eval-uating aspects of validity and reliability of the scale. Results:  Participants were not able to differentiate reliablybetween the 11 rating scale categories of the ABC-ICE. Addi-tionally, 3 items failed to show acceptable goodness of fit to theABC-ICE rating scale model. By collapsing categories andcreating a new 5-category scale, only 1 item misfit. Removingthat item resulted in a modified version of ABC-ICE with 5categories and 15 items. Both item goodness-of-fit statisticsand principal components analysis supported unidimensional-ity of the modified ABC-ICE. The ABC-ICE measures reliablyseparated the sample into at least 4 statistically distinct strata of balance confidence. Finally, the hierarchical order of itemdifficulties was consistent with theoretic expectations, and theitems were reasonably well targeted to the balance confidenceof the persons tested. Conclusions:  Rasch analysis indicated a need to modify theABC-ICE to improve its psychometric properties. Further stud-ies are needed to determine if similar analyses of other versionsof the ABC, including the srcinal one, will yield similarresults. Key Words:  Aging; Outcome assessment (health care);Postural balance; Psychometrics; Rehabilitation; Self-efficacy.©  2010 by the American Congress of Rehabilitation Medicine F ALL-RELATED PSYCHOLOGIC factors are well-knownbarriers to older persons’ mobility and participation invarious activities. 1-3 Among these factors are fear of falling, 4 low self-efficacy, 5 and lack of confidence in maintaining bal-ance. 6 Fortunately, an increasing body of evidence indicatesthat fall-related psychologic factors are modifiable and thatinterventions provided by rehabilitation professionals play animportant role. 7-10 However, there remains a need for improve-ments in quantifying these important fall-related phenomena. 3 Moreover, in rehabilitation, as in other human sciences, there isa call for uniform and improved evaluation strategies 2 andquality linear measures rather than ordinal-level scales. 11 Anadditional requirement is that these measures are suitable foruse across a range of languages and different cultural con-texts. 12 The ABC Scale 6,13 is an instrument that has the potential tofulfill the requirements for use with different languages and/orworld regions, provided people need to deal with comparableenvironmental challenges when maintaining postural balancein their daily lives. ABC is a standardized 16-item question-naire in which respondents are asked to rate their level of confidence in maintaining balance and remaining steady whileperforming specific activities in the home and outside it. 13 These activities vary in difficulty and are performed duringposition changes, standing, or walking. The ABC is used inNorth America, and published translations of the scale includeUnited Kingdom English, 14 French, 15 Chinese, 16,17 and Ger-man. 18 The popularity of the ABC is likely based on factorssuch as its theoretic background, the involvement of olderadults in the scale development, and evidence of validity andreliability. 3,19 Additionally, based on Bandura’s theory of self-efficacy, 20 the highly situation-specific items of ABC are con-sidered to be one of its major strengths because it facilitates From the School of Health Sciences (Arnadottir) and School of Humanities andSocial Sciences (Gunnarsdottir), University of Akureyri, Akureyri, Iceland (Arnado-ttir); and Division of Physiotherapy, Department of Community Medicine and Reha-bilitation (Arnadottir, Lundin-Olsson) and Division of Occupational Therapy, Depart-ment of Community Medicine and Rehabilitation (Fisher), Umeå University, Umeå,Sweden.Presented to the Nordic Congress of Gerontology, May 25–28, 2008, Oslo, Nor-way; the World Congress in Physical Therapy, June 5–12, 2007, Vancouver, Canada.Supported by RANNÍS (grant no. 050410031), KEA University Fund, and Uni-versity of Akureyri Research Fund.The sponsors reviewed annual reports of study process but had no role in thedesign, methods, subject recruitment, data collections, analysis, or preparation of themanuscript.No commercial party having a direct financial interest in the results of the researchsupporting this article has or will confer a benefit on the authors or on any organi-zation with which the authors are associated.Correspondence to Solveig A. Arnadottir, MSc, School of Health Sciences, Uni-versity of Akureyri, Solborg v/Nordurslod, 600 Akureyri, Iceland, e-mail:  Reprints are not available from the author.0003-9993/10/9101-00120$36.00/0doi:10.1016/j.apmr.2009.09.010 List of Abbreviations ABC Activities-Specific Balance Confidence ScaleABC-ICE Icelandic version of the Activities-SpecificBalance Confidence ScaleMMSE Mini-Mental State ExaminationMnSq mean square 156 Arch Phys Med Rehabil Vol 91, January 2010  Author's personal copy consistent individual interpretation of the questions and reliableexpression of perceived balance confidence. 6 However, thisattribute of the ABC may limit its applicability in areas of theworld where life tasks are profoundly different from thoseincluded in the test items.Previous research has supported high internal consistency(Cronbach    0.96) of the ABC items and test-retest reliability( r   0.92), concurrent validity ( r   0.63), 6,21 and discriminantvalidity of ABC scores. 6 ABC scores have been shown todifferentiate between fearful and nonfearful older people, be-tween those who avoided activity because of a fear of fallingand those who did not avoid activity, 22 and between groups of older people physically functioning at different levels. 21 Someevidence of reliability and validity of the ABC scores whenused with groups with specific medical conditions, such asvestibular disfunction, 23 lower-limb amputations, 24 stroke, 15,25 and Parkinson’s disease, 26 also exist. Finally, when developingthe ABC, scalogram analysis was used to establish the hierar-chical order of the items. 6 The results revealed that the easiestitem was “reach at eye level,” and the most difficult item was“walk on icy sidewalks.”Despite its strengths, some aspects of the psychometric prop-erties of the ABC can been criticized, warranting further eval-uation of the scale. First, although the use of rating scales withtoo many categories has been criticized, 27,28 each of the 16activities (items) on the ABC is rated by using an 11-categoryordinal scale (with 10 percentage point increments betweencategories), ranging from 0% (no confidence) to 100% (com-pletely confident). 13 Second, calculating the mean of summedordinal data to obtain total scores is inappropriate 11 becauseequal interval (linear) data are a fundamental requirement foreven the most basic mathematic manipulations. 29 Yet, with theABC, ordinal ratings are summed and divided by 16 to providean overall mean balance confidence score that can range from0 to 100. Additionally, each time a test is translated and appliedto a new cultural context, the validity and reliability of thetranslated version need to be examined. 30 Rasch measurement methods, a modern test theory, 31 are in-creasingly being used within rehabilitation sciences 11 to developnew tests, 32-34 to modify and improve older ones, 35,36 and forcross-cultural validity testing. 37,38 In contrast to methods basedin classic test theory, 31 Rasch measurement methods offer theadvantage of enabling analysis of an instrument at the level of each item and each person rather than focusing almost entirelyon the total score. 11 Yet, existing evidence for validity andreliability of the srcinal ABC is based primarily on classic testtheory. Modern test theory (ie, item response theory, Raschmeasurement) has only been used to study the properties of asimplified version of the ABC with a 4-category rating scaleand the item “walk on icy sidewalks” removed. 28 Thus, incontrast to the present study, modern test methods have notbeen used to analyze or guide modification of either the src-inal or any earlier translated versions of the ABC.Therefore, the overall purpose of this study was to translatethe ABC to Icelandic and then to investigate some psychomet-ric aspects of the ABC-ICE when applied to a new culturalcontext, urban and rural Iceland, by using Rasch analysis. If any psychometric features were found not to be satisfactory,we also planned to use Rasch analysis methods to guide mod-ification the ABC-ICE scale. More specifically, we proposedthe following research questions: (1) Is the 11-category ratingscale of ABC-ICE psychometrically sound? (2) Do the itemson ABC-ICE define a single unidimensional construct of bal-ance confidence? (3) Do items reliably separate older commu-nity-living people into statistically distinct strata related todifferent levels of balance confidence? and (4) Is the hierar-chical ordering of test items logical, and are they targeted interms of balance confidence to older community-living people? METHODSStudy Design This Rasch rating scale analysis of the ABC-ICE was a partof a cross-sectional population-based study among older com-munity-living Icelanders. 39 Data for the study were collectedfrom June 2004 through September 2004. Translation of the Activities-Specific BalanceConfidence Scale A translation and back translation method 40 was used toproduce an Icelandic version of the ABC scale and its instruc-tions for administration. 13 Two bilingual physical therapistsindependently translated the ABC from English to Icelandic.They were instructed to aim for conceptual equivalence be-tween the srcinal and the translated versions. Inconsistenciesin these translations were resolved by discussion among thetranslators. A specialist in the Icelandic language then re-viewed the draft and made some recommendations related tolanguage usage. A third bilingual physical therapist, who hadno knowledge of the srcinal ABC, translated the Icelandicdraft back into English. All 3 translators, in cooperation withthe test developer, compared the srcinal and back-translatedABC versions and negotiated a final version of the ABC-ICE.Some of the test items required an extra discussion becausethere are no escalators or malls in rural areas of Iceland.However, in the standardized protocol for ABC administra-tion, 13 respondents are asked to imagine how confident they arewhen doing the activity, even though they do not do it cur-rently. Therefore, the escalator questions were retained; mostIcelanders have experienced negotiating escalators when theytraveled to urban areas or abroad. The word “mall” was trans-lated to Icelandic as a word with a broader meaning (ie, variousclusters of shops: malls, strip malls, shopping centers). TheABC-ICE was piloted on 3 urban and 2 rural volunteers beforedata collection to evaluate the relevance and ease of compre-hension of the items, and no need for further modification wasidentified. Participants Inclusion criteria for participation were as follows: (1) had tobe at least 65 years of age, (2) had to be community-living, and(3) had to be able to communicate verbally on a telephone andset up a meeting time with a research assistant. An intendedsample of 251 was randomly selected, based on the first 2inclusion criteria, from the national registry for 1 urban and 2adjacent rural areas in Iceland. A total of 15 persons was thenexcluded from participation because they did not fulfill thethird criteria (5 of 15), had recently moved into an institution (7of 15), were inaccessible (2 of 15), or had passed away (1 of 15). Of the remaining 236 persons who fulfilled the inclusioncriteria, 50 declined to participate, and an additional 3 did notcomplete the testing battery. Therefore, the study sample com-prised 183 participants (78% participation rate).The 50 people who declined to participate did not differsignificantly from the study sample with regard to age, sex, orresidence. The reasons for not participating included not hav-ing time, being opposed to research, being too young andhealthy, and being too old and sick. The additional 3 personswho did not complete the test battery were 80 to 84 years old;2 were women, and 1 lived in a rural area. 157 RASCH ANALYSIS OF THE ACTIVITIES-SPECIFIC BALANCE CONFIDENCE SCALE, ArnadottirArch Phys Med Rehabil Vol 91, January 2010  Author's personal copy Procedures Before data collection, 3 research assistants were trained inthe administration of a test battery that included the ABC-ICEas well as other standardized instruments and questionnairesused for purposes of sample description. An introductory letterwas mailed to the intended sample of 251 people. A follow-uptelephone call was then made to each potential participant.During that call, each person was given an offer to participatein the study and a basic screening for eligibility. If he/she metall inclusion criteria, a time to meet was also established. Thosewho declined to participate were offered the opportunity tostate why they declined to participate.Data for age, sex, and residency were obtained from theNational Registry of Iceland. With the exception of 1 perfor-mance assessment of basic mobility, the test battery was ad-ministered in a face-to-face interview format. The participantswere also shown, in an enlarged font, the response options foreach question. Participants self-reported on home nursing andsocial domestic service from the community, height, weight,health problems, use of walking aids, and how often they hadfallen during the last year. Body mass index (kg/m 2 ) wascalculated from the information on height and weight. Finally,all standardized instruments, which included the ABC-ICE, 13 Timed Up & Go, 41 MMSE, 42 and Geriatric Depression Scale, 43 were administered by using established protocols. The Icelan-dic National Bioethics Committee approved the study before itsinitiation (04-037-V2), and all participants gave written in-formed consent for participation in the study. Data Analysis The Statistical Package for the Social Sciences (SPSS, ver-sion 14.0) a was used to generate descriptive statistics, and theWinsteps Rasch-model computer software program b was usedfor Rasch rating scale analyses. A detailed description of Raschanalysis methods has been provided elsewhere. 11 We first evaluated the psychometric properties of the 11-cate-gory rating scale by using Linacre’s 27 criteria to determine if (1)each category contained at least 10 observations, (2) distribu-tion of category ratings was reasonably uniform, (3) averagecategory measures advanced progressively with higher ratings,(4) rating scale outfit MnSq goodness-of-fit statistics (fit sta-tistics) were less than 2.0, and (5) rating scale threshold cali-brations advanced monotonically. The evaluation of advancingthreshold calibrations was also evaluated by using visual in-spection of category probability curves. More specifically, vi-sual inspection of a well-functioning rating scale will revealcategory probability curves with distinct peaks. 11 Where indi-cated, our plan was to modify the rating scale by collapsingcategories, one by one, until the best combination of reason-able category ratings, category measure advancement, ratingscale fit statistics, threshold calibration advancement (withdistinct probability curves), separation indexes, and reliabilitycoefficients was attained. 11 To maximize sensitivity of therating scale, our plan was to retain as many categories aspossible, 44 provided the rating scale showed good psychomet-ric qualities.The second step was to evaluate if the ABC-ICE itemsdefined a unidimensional construct by examining fit statisticsfor each item. Both outfit and infit MnSq and standardized (z)fit statistics were considered. 45 An expected MnSq value is 1.0and can range from 0 to infinity. 45,46 Lower MnSq values implythat ratings varied less than expected, and higher values indi-cate that scores varied more than expected. The former canreflect some form of redundancy in the data but, overall, is notconsidered to be a major threat to measurement 11,45 (ie, al-though items with low MnSq are less productive for measure-ment, they are not considered to degrade it). 46 However, highMnSq values can suggest that the item belongs to anotherconstruct and/or that the item is not scored in a reliable or validmanner (eg, because of a lack of clarity). Because a rule of thumb for reasonable fit statistics is 0.6 to 1.4 for ratingscales, 46 we set our criteria for item removal based on outfit orinfit MnSq greater than 1.4 accompanied by z greater than orequal to 2. Unidimensionality was also evaluated by usingprincipal components analysis of the residuals. In accordancewith previous suggestions, 47 unidimensionality is confirmedwhen greater than 50% of the variance is explained by theRasch dimension and less than 5% of the unexplained varianceis accounted for by the first contrast. Additionally, we evalu-ated fit statistics for each participant to explore if he/she re-sponded to the ABC items in a valid manner. Misfitting personswere detected by using the same criteria as misfitting items. Weused person fit as one of the indicators of the effects of modifying the rating scale and/or item removal.Our third step was to examine Rasch indices of reliability toexplore if items reliably separated participants into statisticallydistinct strata of balance confidence. Although the Rasch itemseparation reliability coefficient is interpreted similar to Cron-bach   , there is no equivalent to the person separation reliabil-ity coefficient. The closely related person separation indexprovides an estimate for the spread or separation of personsinto strata along the scale. 11 A person separation index of 1.5accompanied by a person reliability coefficient of 0.70 is theabsolute minimum required to divide a sample of persons intoat least 2 significantly distinct strata separated by at least 3 SEunits. 48 The preferable minimum is a separation index of 2.0with a separation reliability of 0.80. 49 Our final step was to examine the ABC-ICE item difficultyand person confidence hierarchies. The center of the logit scaleis commonly set at 0 logits and represents the average difficultyof the items. 11 Because the person balance confidence mea-sures and item difficulty measures are placed along the samelinear continuum, it is possible to compare the mean and rangefor the person balance confidence measures and determine if the difficulty level of the ABC-ICE items is targeted to thebalance confidence level of the sample. Examination of theitem hierarchy in relation to theoretic expectations and priorresearch can also be used to evaluate if the hierarchical order-ing of test items is logical. RESULTS The characteristics of the study sample are presented in table 1.Fifteen participants received MMSE scores below 24, whichindicated that they may have had cognitive impairments. 50 Toconfirm that persons with lower cognitive levels did not impactour results, we also implemented an analysis in which the 15persons with low MMSE scores were omitted. When theseparticipants were removed, we obtained essentially identicalresults. Therefore, because these 15 persons met our inclusioncriteria and do represent a proportion of community-livingolder persons, we chose to report only the results for the entiresample.The observed counts for each of the 11 categories of theABC-ICE rating scale exceeded 10 observations. The lowestfrequency of ratings was 29 in the 10% confidence category,and the highest was 1291 in the 100% confidence category; the0% category had 117 ratings. The average category measuresadvanced monotonically along the rating scale, with outfitMnSq values ranging from 0.57 to 7.34. Threshold calibrationswere disordered, and, consequently, some of the probabilitycurves did not show distinct peaks (fig 1A). 158  RASCH ANALYSIS OF THE ACTIVITIES-SPECIFIC BALANCE CONFIDENCE SCALE, ArnadottirArch Phys Med Rehabil Vol 91, January 2010  Author's personal copy Analysis of goodness of fit of the 16 ABC items to the 11category Rasch rating scale model revealed that 3 items(“sweep the floor”: infit MnSq  1.45, z  1.9, outfit MnSq  1.96,z  2.4; “pick up a slipper”: infit MnSq  1.81, z  4.5, outfitMnSq  1.77, z  3.0; “on/off escalator while holding parcels”:infit MnSq  1.40, z  3.0, outfit MnSq  1.24, z  1.4) failed toshow acceptable levels of fit. Analysis of goodness of fit of the157 participants who did not receive maximum ABC scoresrevealed that 17 participants (10.8%) failed to show acceptablelevels of fit.Because our initial analysis revealed a need to modify theABC-ICE rating scale, we followed the simple guideline thatcollapsed categories must make sense 11 and started to system-atically (step by step) collapse adjacent categories having lowobservation counts. At each step, we confirmed improved psy-chometrics of the rating scale. We stopped the process whenthe best solution was achieved; the best result was a 5-categoryscale. More specifically, the extreme categories from the src-inal ABC-ICE were retained, and the intermediate categorieswere collapsed (such that 0  0%, 1  10% to 40%, 2  50% to70%, 3  80% to 90%, and 4  100%). This modification re-sulted in reasonably equal distribution of ratings, advancingcategory measures, acceptable rating scale outfit statistics, andordered threshold calibrations displayed graphically as cate-gory probability curves with distinct peaks (fig 1B). The num-ber of misfitting items decreased from 3 to 1 (“pick up aslipper”: infit MnSq  1.53, z  3.6, outfit MnSq  1.58, z  2.6),and the number of misfitting persons decreased from 17(10.8%) to 12 (7.6%). Revision of the rating scale also resultedin slight improvements in person reliability (0.88–0.91) andperson separation (2.65–3.12). The person separation index of 3.12 indicated that the items on the modified ABC-ICE couldreliably separate the 157 participants who did not receivemaximum scores into at least 4 statistically distinct strata of person balance confidence (number of strata  [4Gp  1]/3  4.5). 48 Subsequent deletion of “pick up a slipper” resulted in aunidimensional modified ABC-ICE scale, with the remaining15 items showing acceptable levels of fit to the Rasch ratingscale model. Principal component analysis of this modifiedABC-ICE, with 5 categories and 15 items, supported unidi-mensionality because 72.5% of the total variance was ex-plained by the Rasch dimension and only 4.6% of the unex-plained variance was accounted for by the first contrast.The modified ABC-ICE item measures and their fit statisticsare presented in table 2. The item difficulty measures rangedfrom   1.91 to 2.53 logits. People were most confident inmaintaining balance while sweeping the floor (  1.91 logits)and walking outside the house to a car parked in the driveway(  1.15 logits) and least confident in stepping onto or off anescalator while holding onto parcels (2.53 logits) and walkingoutside on icy sidewalks (2.43 logits).A person-item map (fig 2) shows graphically the hierarchyand spread of participants and items along the common linearlogits scale. The person balance confidence measures rangedfrom  2.29 to 5.42 logits. The spread of the participants wasskewed toward the lower end of the scale and exceeded therange of the items at the higher end, resulting in no items beingtargeted to the most confident persons and 26 persons receivingmaximum scores (ceilings effects). The mean balance confi-dence (2.25 logits) for the sample was higher than the meandifficulty (0 logits) for the items, also indicating less than idealtargeting. DISCUSSION Rasch analysis indicated a need to modify the ABC-ICE toimprove its psychometric properties. A modified ABC-ICEwith a 5-category rating scale and 15 items showed goodseparation reliability and internal scale validity. These findingssupport the application of the ABC to an Icelandic urban andrural context. However, they do not ensure generalizability of the tool when applied in other cultural contexts. Rather, ourresults reinforce the need to analyze the srcinal version as wellas any existing translations of the ABC to determine if similarmodifications are indicated. In this study, we have shown howRasch analysis methods can be used to guide such revision of an existing tool.A major problem with the 11-category ABC-ICE rating scalewas that a large proportion of the available categories wasunderused. This situation can occur when participants are un-able to distinguish among similar rating categories. For exam-ple, it may have been difficult for a person who has very littleconfidence to distinguish between rating categories 10%, 20%,and 30%. Item response theory modeling has revealed that a4-category response format of a simplified version of the ABCimproved its psychometric properties. 28 We obtained similarresults that indicated that a modified 5-category version of ABC-ICE may make it possible for persons to reliably differ-entiate between available categories. The potential result couldbe a simplified evaluation process for users. However, it isimportant to note that our participants responded to an 11-category scale, which we recoded. Further research would beneeded to determine if comparable results are obtained byhaving persons respond directly to a 5-category scale as areobtained when an 11-category scale is used, but data are thenrecoded.The ABC-ICE items represent activities performed duringposition changes, standing, or walking. 6 However, the “pick upa slipper” item failed to show acceptable fit, indicating it likelyrepresents a different construct and supporting its removal fromthe ABC-ICE. The reason for the misfit may be related to thefact that this item differs from the other ABC-ICE items Table 1: Characteristics of the Participants Characteristics N  183 Age (y) 73.8  6.2 (65–88)Sex: women 87 (47.5)Residency: rural 67 (36.6)Use of a walking aid 25 (13.7)Service from the communityHome nursing 13 (7.1)Social domestic service 61 (33.3)History of a fallOne fall in the past year 37 (20.2)Two or more falls in the past year 21 (11.4)Health problems (number) 4.1  2.1 (0–10)Body mass index (kg/m 2 ) 26.8  3.8 (17–39)ABC-ICE score* 83.3  18.3 (21–100)TUG (s) † 10.7  3.6 (5–24)MMSE ‡ 27.2  2.5 (16–30)GDS § 6.5  4.3 (1–20)NOTE. Values are mean  SD (range) or n (%).Abbreviations: GDS, Geriatric Depression Scale; TUG, Timed Up &Go Test.*Theaverageofrawordinalscores,maximumrange0to100;higherscore indicates more balance confidence. † No maximum range, higher score indicates more impaired mobil-ity. ‡ Maximum range 0 to 30; higher score indicates better cognitivefunction. § Maximum range 0 to 30; higher score indicates more depressivesymptoms. 159 RASCH ANALYSIS OF THE ACTIVITIES-SPECIFIC BALANCE CONFIDENCE SCALE, ArnadottirArch Phys Med Rehabil Vol 91, January 2010
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!