Development of a Clinical Multiple-Lunge Test to Predict Falls in Older Adults

Transcription

1 458 ORIGINAL ARTICLE Development of a Clinical Multiple-Lunge Test to Predict Falls in Older Adults Ruth Wagenaar, MHSc, Justin W. Keogh, PhD, Denise Taylor, PhD ABSTRACT. Wagenaar R, Keogh JW, Taylor D. Development of a clinical Multiple-Lunge Test to predict falls in older adults. Arch Phys Med Rehabil 2012;93: Objective: To develop a new Multiple-Lunge Test to distinguish between fallers and nonfallers in community-dwelling older adults. Design: A cross-sectional design was used to establish the sensitivity and specificity of the test to predict faller status based on retrospective self-reported fall history. Setting: Local retirement villages. Participants: Community-dwelling older adults (N 130; mean age SD, 77 7y) with (n 40) and without (n 90) a history of falls. Interventions: The Multiple-Lunge Test required individuals to lunge forward to a step length determined as 60% of their leg length, and return to start, for 5 consecutive repetitions. Interday and intraday test-retest reliability of the Multiple-Lunge Test was established across 2 testing occasions. Main Outcome Measures: Number of steps performed correctly, total time to complete 5 steps. Results: The Multiple-Lunge Test was found to be reliable across trials (Intraday: intraclass correlation coefficient [ICC] for steps, ICC for time; Interday: ICC.77 for steps; ICC.84 for time). Sensitivity and specificity values were calculated as 73% and 63%, respectively, for predicting multiple fallers using the measure of all 5 steps done correctly. Conclusions: The test is easily administered and because of its challenging nature, it may be well suited to detect subtle differences in abilities of higher functioning, communitydwelling older adults. A practitioner can be confident in 7 of 10 cases that an older adult who cannot complete all 5 steps of the Multiple-Lunge Test is at high risk of falls. The results suggest that there is potential for the Multiple-Lunge Test to be used in clinical practice; however, additional research on how to further increase its validity appears warranted. Key Words: Accidental falls; Aging; Rehabilitation by the American Congress of Rehabilitation Medicine THE NUMBER OF OLDER adults is increasing worldwide, so an understanding of the potentially negative issues associated with aging is essential. 1 One of the most prevalent and often injurious consequences related to aging is falls. The incidence of falls in people 65 years and older is 1 fall per year in approximately 30% of the population, 2 with up to 30% of falls leading to serious injuries that hinder these individuals ability to function independently 3 and contribute to an increased fear of falling, a reduced quality of life, and even death. 4 Although falls are prevalent in older adults, there is considerable evidence that targeted approaches addressing specific risk factors for falls may reduce fall rates in higher risk individuals or groups. 5 Prospective and observational studies 6-8 have shown high validity and reliability of fall risk, but quickly and easily administered assessment tools that give a predictor value immediately may be more useful to clinicians. Clinical tests are therefore often used in practice to identify those who are at a higher risk of falls, so that specific fall prevention programs can be implemented for that individual or group of older adults. Examples of these tests include the Berg Balance Scale (BBS), Timed Up and Go (TUG), Functional Reach Test (FRT), Five Times Sit to Stand (5-STS), and the Fullerton Advanced Balance (FAB) scale. Although most of these tests have been reported to have excellent test-retest reliability (intraclass correlation coefficient [ICC].80) and high sensitivity ( 80%) and specificity ( 80%) in predicting falls in frailer older adults, 9-13 the ability of many of these tests to predict falls in better functioning, community-dwelling older adults may not be as strong. For example, when used with community-dwelling elders, the BBS exhibits ceiling effects and has low sensitivity (25% 45%) in predicting falls. 15 Trueblood et al 17 examined the predictive ability of the TUG in assessing falls in community-dwelling elders and reported a sensitivity of 10% and specificity of 95%. Dite and Temple 18 reported that a score of 25cm on the FRT gave the best combination of sensitivity (63%) and specificity (59%) in predicting fall status in community-dwelling elders. When assessing the predictive ability of the 5-STS in community-dwelling older adults, Tiedemann et al 19 also observed relatively low levels of sensitivity (66%) and specificity (55%). More recently, Hernandez and Rose 12 found that the FAB with a cutoff score of 25 out of 40 gave a sensitivity of 75% and a specificity of 53% in predicting falls List of Abbreviations From the Centre for Physical Activity and Nutrition Research, School of Sport and Recreation (Wagenaar, Keogh), and the Health and Rehabilitation Research Institute, School of Physiotherapy (Taylor), Auckland University of Technology, Auckland, New Zealand, and Faculty of Health Sciences and Medicine, Bond University, Australia (Keogh). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Ruth Wagenaar, MHSc, School of Sport and Recreation, Auckland University of Technology, Private Bag 92006, Auckland 1142, New Zealand, ruth.wagenaar@aut.ac.nz /12/ $36.00/0 doi: /j.apmr BBS FAB FRT FSST ICC PARQ RAPA 5-STS TUG Berg Balance Scale Fullerton Advanced Balance Functional Reach Test Four Square Step Test intraclass correlation coefficient Modified Physical Activity Readiness Questionnaire Rapid Assessment of Physical Activity Questionnaire Five Times Sit to Stand Timed Up and Go

2 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar 459 in community-dwelling elders. The relative inability of most of these clinical tests to predict falls in well-functioning, community-dwelling elders therefore suggests that additional tests still need to be developed to assess fall risk in this population. Such tests would need to exhibit high reliability and validity, and be relatively quick and easy to use by health professionals with minimal specific training. Because most falls involving older adults are characterized by a sudden loss of balance and an inability to quickly produce an appropriate corrective response, 5 many studies have investigated how older adults respond to unexpected perturbations, 5,20 initiate stepping responses, or both. These studies have shown that fall risk is associated with the use of more steps to regain balance after perturbation, 24 decreased step length, 12,21,22 and longer response and transfer (step) times. 23 This would suggest that stepping tasks in which the base of support continually changes position may be ideal tasks to assess older adults fall risk in clinical situations. Because tripping is the most common cause of falls for older adults, accounting for up to 53% of all falls, 5,25 it would also appear useful to develop a stepping clinical test that assesses balance in a manner similar to that encountered when trying to arrest a trip-type fall. The movements required for trip recovery would appear to resemble a lunge-type movement, with a relatively large, fast step in the forward direction and on landing, the production of large forces required to counter the forward momentum of the center of mass and to regain balance. 5 While some lunge-type stepping tasks have been proposed to be useful fall risk assessment tools for older adults, they typically have been laboratory-based measures and involved only 1 anterior (out) or anterior-posterior (out and back) step. As such, these single-step tasks may not be able to be used in clinical practice because of the inability to reliably and accurately measure parameters such as the duration of a step via a stopwatch. The development and validation of a new, more challenging lunge-type stepping assessment tool for predicting fall risk in the increasingly active community-dwelling older adult population would therefore appear justified. The purpose of the following study was to determine the reliability and validity of 1 such test, the Multiple-Lunge Test, which involves 5 consecutive anteriorposterior steps, in distinguishing fallers and nonfallers. METHODS Experimental Design A cross-sectional design was used to examine the ability of the Multiple-Lunge Test to predict faller status in a group of community-dwelling older adults. Consistent with other recent studies, 26,27 fall history was obtained by retrospective selfreport by participants and constituted the criterion variable. The participants were categorized as nonfallers (no falls), single fallers (1 fall), or multiple fallers ( 1 fall) based on their report of the falls they experienced in the previous 12 months. Two outcomes measures from the Multiple-Lunge Test were calculated and used as potential predictors of falls: (1) the number of steps done correctly, and (2) the time taken to complete the test. Participants Participants were recruited for the study through several organizations. Various retirement villages in the Auckland region were approached, and advertising was done at the villages by way of newsletters, posters, and/or presentations by the researcher. A number of community groups, such as the New Zealand 60s Up Movement and Returned Services Associations, were visited at meeting times where the details of the study were presented by the researcher, and volunteers requested. To establish eligibility for the study, the Modified Physical Activity Readiness Questionnaire (PARQ) was used to determine the general level of ability to participate in the physical activity required for this study, and whether the participants could safely perform the lunge tasks that comprised the experimental procedures of this study. Normative cognition was determined by the participants ability to complete all forms and questionnaires with minimal or no assistance. Individuals had to be 65 years or older, independently living in the community, with normal vision or vision corrected by eyewear. Individuals who had any self-reported neurologic, musculoskeletal, sensory, or cognitive impairments that would affect their strength and balance, particularly of the lower extremity, were excluded from the study. All participants signed an informed consent form. The study was approved by Auckland University of Technology Ethics Committee. The Falls History Questionnaire was administered to determine fallers from nonfallers. The testers helped to explain the requirements of the questionnaire and encouraged the participants to take their time to answer in order to obtain as accurate results as possible. The questions included were those recommended by Lord et al 5 to establish a detailed history of the events surrounding a fall. A fall was defined as an event resulting in an individual unintentionally coming to rest on the floor, or other lower level, not as the result of a major intrinsic 15,26,28,29 (p 1702) event or overwhelming hazard. Fourteen individuals who participated in the full study and were available for a second testing session also gave informed consent to complete another session in order for the researchers to gain some insight into the reliability of the Multiple-Lunge Test measures. The second testing session was conducted 7 days after the initial session, to allow sufficient recovery from the first session and to control for weekly differences in activity. Physical Activity The Rapid Assessment of Physical Activity Questionnaire (RAPA) was used to determine the amount of different types of physical activity (light, moderate, or vigorous) the participants engaged in during a normal week. 30 Questions relating to strength and flexibility activities were also included. The questionnaire was scored to categorize each individual s physical activity levels as sedentary, underactive, or active. The RAPA has been found to have moderate to high sensitivity (81%) and specificity (69%) to predict activity participation in older adults. 30 Procedures After completing the 3 questionnaires (PARQ, Falls History, and RAPA), the participants were given a demonstration and explanation of the Multiple-Lunge Test. The goal was to step forward with the dominant leg a predetermined distance and return to the starting position as quickly as possible for 5 repetitions. After considerable pilot testing, the required step distance was determined to be equal to 60% of leg length (as measured from the midpoint of the anterior superior iliac spine of the pelvis to the midpoint of the lateral malleolus of the ankle). The 60% of leg length distance was marked with tape on the floor, to make it clear what distance the participants were required to step. While the maximal return length of older adults is approximately 58% of their body height, 22 Hsiao-Wecksler and Robinovitch 21 suggested that a step equal in distance to 35% of standing height (quite similar to

3 460 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar Fig 1. Schematic representation of the movements involved in the Multiple-Lunge Test. (A) Correct step. (B) Incorrect step. Unshaded foot represents foot of stance leg, and shaded foot represents foot of swing leg. our 60% of leg length) may be optimal to regain balance. Our results during pilot testing also indicated that the use of step lengths greater than 60% of leg length was perhaps too difficult to consistently achieve across 5 repetitions for many participants. The instructions given to the participants before starting the Multiple-Lunge Test were to start with the toes of both feet immediately behind the first line and step forward so that the heel of the stepping foot landed just past the second line, while keeping the stance foot on the floor (figs 1 3). Either leg could be chosen as the step leg, but the same leg had to be used as the step leg for all 5 steps in both trials, and it was suggested that the dominant leg be the step leg. After the step foot landed, the participants were to step back with the swing leg so that it was repositioned just behind the start line and next to the stance leg. They were told to attempt the lunge step in a single movement, without holding onto the tester or wall for support unless imperative to avoid a fall. A spotter was standing nearby to assist participants if they lost their balance. Consistent with other common clinical tests such as the TUG, 10,31 5-STS, 32 and Four Square Step Test (FSST), 18 a stopwatch was used to determine the total time taken for the Multiple-Lunge Test. The total time was recorded from the go command of the first repetition to when the participant returned to the starting position after 5 repetitions. The participants performed a practice trial of between 2 and 3 steps followed by 2 timed trials, with a 1-minute rest between trials. Data Analysis The outcome measures used in this study were the time to complete all 5 repetitions (total time) and the number of correct steps out of 5. Total time was calculated as the time from the Fig 2. Photograph of Multiple-Lunge Test start and end position.

4 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar 461 initiation of the first step to the step foot being completely flat on the ground behind the start line after completing the fifth step. In order for each step to be classified as correct, the older adult must have received no assistance from the tester, walking aid, or by holding onto a wall; passed the mark on the floor with the heel of the stepping foot; and taken only 1 step to reach the starting and finish lines. The trial with the most correct steps was used for further analysis. If both trials for a participant had the same number of correct steps, the fastest trial was used for data analysis. Statistical Analysis ICCs were used to evaluate intraday and interday test-retest reliability of the Multiple-Lunge Test. Analysis of all trials was done using the reliability spreadsheet developed by Hopkins. 33 ICC categories were based on standards recommended by Fleiss, 34 where ICC values of 0.4 or below represent poor reliability, 0.5 to 0.6 represent moderate reliability, and values of.75 and above represent good to excellent reliability. To establish the accuracy of the Multiple-Lunge Test in distinguishing between nonfallers and fallers, and thus the ability to predict falls in older adults, the sensitivity and specificity of the test were calculated. Sensitivity refers to the proportion of people who are fallers who test positive. If sensitivity is high, a negative test will rule out high fall risk. Specificity refers to the proportion of people who are not fallers who test negative. If specificity is high, a positive test will rule in high fall risk. 35 Diagnostic test characteristics are defined and calculations shown in table 1. These characteristics were determined using the truth table, adapted from Davidson 35 and displayed in table 2. RESULTS A total of 130 independent, community-dwelling adults aged 65 years or older (mean age SD, 77 7y; range, 65 93y) met Table 1: Definitions and Calculations of Test Characteristics Test Characteristic Definition Calculation Sensitivity Specificity Positive predictive Negative predictive Accuracy Prevalence have the disorder who test positive do not have the disorder who test negative test positive who have the disorder test negative who do not have the disorder were correctly identified as either having or not having a history of falls The proportion of people in the sample who were fallers NOTE. Adapted from Davidson. 35 a/(a c) d/(b d) a/(a b) d/(c d) (a d)/(a b c d) (a c)/(a b c d) the inclusion criteria and participated in this study. The distribution of participants by sex was 88 women (68%) and 42 men (32%). One individual was excluded because of being younger than the age criteria, and one was excluded because of very poor eyesight caused by glaucoma. Ninety participants (69% of sample) reported no falls in the last year. Of the 40 participants (31% of sample) reporting at least 1 fall in the previous 12 months, there were 25 single fallers (19%) and 15 multiple fallers (12%). These 40 fallers reported 64 falls in the previous year. Within the multiple-faller group, 9 participants reported 2 falls (60%), 3 participants reported 3 falls (20%), and 3 participants reported 4 falls (20%). The most common cause of a fall was a trip (61%), followed by turning (9%) and a slip (8%). Participant characteristics are outlined in table 3. Participant physical activity categories by group are displayed in table 4. There were no significant between-group differences in subject physical activity categories. Both the nonfaller and faller group had a low percentage of sedentary participants (3.3% and 2.5%, respectively). Most participants from all groups were scored a physical activity level of underactive. The 14 individuals (mean age SD, 79 7y) who participated in the reliability study consisted of 43% (n 6) fallers and Table 2: A 2 by 2, or Truth Table to Calculate Predictive Properties of an Assessment Tool Predictor Present (Faller) Absent (Nonfaller) Totals Positive True positive a False positive b a b predictor Predicts faller Actual faller Predicts faller Not actual faller Negative False negative c True negative d c d predictor Predicts nonfaller Actual faller Predicts nonfaller Not actual faller Totals a c b d a b c d Fig 3. Photograph of Multiple-Lunge Test midway position. NOTE. Adapted from Davidson. 35

5 462 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar Table 3: Participant Characteristics of 130 Older Adults Characteristic Nonfallers (n 90) Single Fallers (n 25) Multiple Fallers (n 15) Age (y) (65 91) (65 93) (67 89) Height (m) ( ) ( ) ( ) Mass (kg) (35 119) (56 94) (62 119) Leg length (cm) (75 101) (80 97) (84 95) Falls history* (2 4) NOTE. Values are mean SD (range) or as otherwise indicated. *Number of falls reported within the last year. 57% (n 8) nonfallers. Of the 6 fallers, 2 (14%) were multiple fallers ( 1 fall in the previous year). The proportion of fallers in the sample was similar to the proportion generally reported in the literature for this age group 2 and that of our overall sample of 130 participants. Good to excellent levels of intraday reliability were reported for the number of steps correctly performed (ICC.79.81) and the total time (ICC.86.88). Similar levels of interday reliability were observed for the number of steps correctly performed (ICC.77) and the total time (ICC.84). The number of steps performed correctly and the total time to complete all 5 repetitions by each group are presented in figures 4 and 5, respectively. Five of 5 steps were performed correctly by 63% of the nonfaller group and 48% of the faller group. This difference in completion rate of successful steps was greater when the faller group was divided into single and multiple fallers, with only 27% of the multiple-faller group able to correctly perform all 5 steps. The ability of the test to distinguish between nonfallers and fallers (sensitivity and specificity, and predictive values) was determined by using the truth table format shown in table 2. Comparison 1 evaluated the percentage of participants with 0 of 5 steps correct, and involved comparisons of fallers with nonfallers, as well as multiple fallers with nonfallers. Comparison 2 evaluated the percentage of participants with 5 of 5 steps correct, using the same comparisons. The predictive properties of the Multiple-Lunge Test were then calculated, and are presented in table 5. The highest specificity (87%) was achieved using comparison 1, with 0 steps correct out of 5 steps. However, this comparison resulted in low sensitivity: 25% for fallers versus nonfallers, and 40% for multiple fallers versus nonfallers. Comparison 2, with all 5 steps correct, resulted in moderate sensitivity and specificity, but these values were greater (sensitivity, 73%; specificity, 63%) when multiple fallers and nonfallers were compared. The highest positive predictor variable (75%) was for 0 correct steps (multiple fallers vs nonfallers), but the negative predictor (59%) was only moderate. The highest negative predictor (70%), with a positive predictor of 67%, was for 5 correct steps, multiple fallers versus nonfallers. Accuracy of the test was higher (0 correct steps, 63%; 5 correct steps, 68%) in the multiple-faller and nonfaller comparisons, for both conditions, than for the faller, nonfaller comparisons (0 correct steps, 56%; 5 correct steps, 58%). DISCUSSION The aim of the study was to develop a new step-based falls assessment tool (Multiple-Lunge Test) for community-dwelling older adults and determine its reliability and validity in distinguishing between older fallers and nonfallers. Results would suggest that both of these aims were relatively well achieved, indicating some potential for the use of this test in clinical practice, research, or both. The intraday test-retest reliability for the number of steps correctly performed (ICC.79.81) and the total time to complete all 5 steps (ICC.86.88) of the Multiple-Lunge Test was good to excellent. Although little research appears to have assessed the intraday reliability of similar tests in communitydwelling older adults, the intraday reliability of the Multiple- Lunge Test appears to be slightly lower than that of the FSST (ICC.94.99). 36 Similar to that recommended for the FSST, it would appear that the performance of 2 trials of the Multiple- Lunge Test per testing occasion is sufficient to reliably determine performance. The interday test-retest reliability for the number of steps correctly performed (ICC.77) and the total time to complete all 5 steps (ICC.84) was also good to excellent. Such interday reliability values appear to be within the range of values reported in the literature for similar tests involving community-dwelling elders. Specifically, the interday reliability of the Multiple-Lunge Test appears to be substantially greater than that reported for the TUG (ICC.56), 37 quite similar to that of the 5-STS (ICC.82.89), 19,32 but somewhat lower than that reported for the FSST (ICC.93.98) 18 and FAB (Spearman rank correlation coefficient,.96). 9 These results suggest that the Multiple-Lunge Test exhibits sufficient interday test-retest reliability to be a useful tool in evaluating the success of fall risk intervention programs in this population. Regardless of its reliability, the Multiple-Lunge Test will not be commonly used in clinical practice or research if it is not a valid predictor of falls in older adults. Results of the present study indicated small to moderate differences between fallers and nonfallers in terms of Multiple-Lunge Test performance. The comparison of mean total time to complete all 5 steps showed a significant difference between nonfallers (12.50s) and fallers (15.40s), and between the multiple- (16.79s) and single-faller (14.57s) groups. However, comparisons of the results for mean total time are limited in that they do not take into account the number of steps completed correctly. For example, an individual may have completed the test in a fast time at the expense of performing all steps correctly. When comparing the number of correct steps performed, a high specificity (87%) but low sensitivity (25% for faller vs nonfaller; 40% for multiple faller vs nonfaller) was found for 0 of 5 steps performed correctly. Therefore, participants who are unable to complete any steps correctly are at increased risk of falls, but participants who perform at least 1 correct step cannot Table 4: Subject Physical Activity Categories by Group Group Sedentary Underactive Active n % n % n % Nonfaller NA Faller NA Total participants 4 NA 84 NA 42 NA 130 Single faller NA Multiple faller NA Total fallers 1 NA 29 NA 10 NA 40 N

6 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar 463 Fig 4. Percentage of nonfallers, fallers, single fallers, and multiple fallers and the number of steps performed correctly from 5 steps during the Multiple-Lunge Test. necessarily be classified as nonfallers. Another simple comparison involved the number of participants who could do all 5 steps correctly. Five of 5 steps were done correctly by 63% of the nonfaller group and 60% of the single fallers, but only by 27% of the multiple fallers. This very basic analysis shows a difference in the abilities of multiple but not single fallers, in relation to nonfallers, to correctly perform all 5 steps. This comparison also resulted in the best combined predictive ability, with a sensitivity of 73% and specificity of 63% for multiple fallers and nonfallers. This implies that an individual who completed all 5 steps correctly had a 73% chance of not being a multiple faller, and that an individual who is unable to perform all 5 steps correctly has a 63% chance of being a multiple faller. These results regarding the ability to perform all 5 steps correctly are therefore somewhat better than chance, but how well does this compare with the fall prediction literature for community-dwelling older adults? The sensitivity of 73% and specificity of 63% for 5 correct steps as observed in the Multiple-Lunge Test in the current study appear to be better in predicting falls in communitydwelling elders than many other common tests. Examples of this include the BBS (sensitivity, 25% 45%), 15 TUG (sensitivity, 10%; specificity, 95%), 17 FRT (sensitivity, 63%; specificity, 59%), 18 5-STS (sensitivity, 62%; specificity, 55%), 19 and perhaps the FAB (sensitivity, 75%; specificity, 53%). However, the Multiple-Lunge Test did not have as high a sensitivity (89%) and specificity (85%) as the FSST. 18 Since the predictive ability of the FSST in distinguishing fallers and nonfallers in community-dwelling older adults therefore appears to be much greater than that of other fall assessment tools, it is interesting to determine what may account for this. Inspection of the participant demographics and fall definitions used by Dite and Temple 18 indicates some differences between their study compared with the current study and the literature for community-dwelling elders. Of the 81 participants in their study, 18 all of the multiple fallers (n 27) and single fallers (n 27) were currently a part of a community rehabilitation program. Further, the definitions they used for multiple and single fallers were 2 or more falls or 1 fall, respectively, in the previous 6 months, 18 rather than 12 months as used in the present study and much of the literature. 19,38,39 The relative frailty of Dite and Temple s 18 multiple-faller group was also

7 464 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar demonstrated by the fact that the subjects in the group had scores on the Step Test, FRT, and TUG that were very similar to those of 300 older adults who were assessed 12 to 33 days after admission to an emergency department after a fall-related injury. It is therefore likely that even though the participants in their study 18 were all community dwelling, the fallers (especially the multiple fallers) were frailer than what would be expected of other community-dwelling elders. The high levels of validity reported in other studies may also be a result of the relative frailty of the sample. 31 The similar physical activity categories of participants in the present study further suggest that the Multiple-Lunge Test may be a useful clinical assessment tool for this population. So while the reliability and validity of the Multiple-Lunge Test both appear to be quite good, greater sensitivity and specificity might still be achieved with some modifications. For example, future studies assessing the psychometric properties of the Multiple-Lunge Test may strive to recruit a greater proportion of multiple fallers, as recent studies are suggesting that single fallers may not be truly representative of fallers, and that the definition of fallers should be more than 1 fall per year. 12,40 Additional studies should also examine the ability of the Multiple-Lunge Test in predicting trip-related falls. This may be done by comparing trip-type fallers with non trip-type fallers, nonfallers, or both. Such studies may be a world first, as the authors are unaware of any other studies that have assessed the potential of any of the common fall risk tests to predict a certain type of fall. There was also a tendency of some of the better functioning participants to perform the Multiple-Lunge Test at such a fast speed that they did not complete all 5 steps correctly. For example, 3 individuals completed the Multiple- Lunge Test in under 10 seconds, but only performed 3 of 5 steps correctly. Thus, some subtle changes in the instructions given to the older adults regarding the primary goal of the test being the performance of correct steps, as opposed to the fastest time, might be appropriate and lead to improved predictive ability. Study Limitations The results of this study need to be viewed in light of some of the study limitations. One limitation was the relatively small number of multiple fallers (n 15) among the 130 total participants. The multiple fallers comprised 38% of the faller group, Fig 5. Time taken to complete Multiple-Lunge Test. Table 5: Sensitivity, Specificity, Predictive Values, Accuracy, and Prevalence Values for Multiple-Lunge Test by Number of Correct Steps Comparison 1: (0/5 Steps) Comparison 1: (0/5 Steps) Comparison 2: (5/5 Steps) Comparison 2: (5/5 Steps) F vs NF MF vs NF F vs NF MF vs NF Sensitivity (%) Specificity (%) Positive predictor (%) Negative predictor (%) Accuracy (%) Prevalence (%) NOTE. Sensitivity a/(a c); Specificity d/(b d); Positive predictor a/(a b); Negative predictor d/(c d); Accuracy (a d)/ (a b c d); Prevalence (a c)/(a b c d). Abbreviations: F, faller; MF, multiple faller; NF, nonfaller. which was slightly less than the approximately 50% that is commonly reported in the literature. 41,42 A larger number of multiple fallers may have produced better results, but recruitment difficulties did not allow this. Such recruitment challenges are not unique, with other studies 14,16 in this area obtaining a similar number (n 16 20) of multiple fallers to that used in the current study. Furthermore, a similar study 27 involving only 5 multiple fallers has also recently been published. Secondly, the use of a retrospective research design, while common in this area of research, 12,14,27 may have led to some inaccuracies in the reporting of falls. Specifically, it is quite possible that this design led to an underestimation of the true number of multiple fallers, since recalling the number of falls sustained in the previous 12 months would require a good memory. The use of a prospective design in which fall incidence is monitored over time may therefore have allowed a more accurate evaluation of the Multiple-Lunge Test s predictive properties. CONCLUSIONS The Multiple-Lunge Test was found to have good to excellent intraday and interday test-retest reliability and to exhibit a good ability for identifying multiple fallers, with a sensitivity of 73% and a specificity of 63% when using 5 of 5 correct steps as the cutoff. When compared with the literature for healthy, communitydwelling older adults, these reliability and validity results appear to be at least as good, if not better than those found for the BBS, TUG, FRT, 5-STS, and FAB scale. It is therefore apparent that the Multiple-Lunge Test has the potential to be used in fall risk assessments of older adults, especially those who are communitydwelling older adults, because of its challenging nature and lack of ceiling effects. However, certain modifications, such as the development of other cutoff scores and the use of split times per step, as well as step accuracy scores, may improve validity while providing more information on reaction time abilities. Assessment of the maximum number of correct lunges that can be performed within a certain time may also be useful. References 1. Statistics New Zealand. The impact of structural population change (structural change and the 65 population articles). Wellington: Statistics New Zealand Campbell AJ, Borrie MJ, Spears GF, Jackson SL, Brown JS, Fitzgerald JL. Circumstances and consequences of falls experi-

8 MULTIPLE-LUNGE TEST AND FALL PREDICTION, Wagenaar 465 enced by a community population 70 years and over during a prospective study. Age Ageing 1990;19: Hausdorff JM, Rios DA, Edelberg HK. Gait variability and fall risk in community-living older adults: a 1-year prospective study. Arch Phys Med Rehabil 2001;82: Pereira C, Vogelaere P, Baptista F. Role of physical activity in the prevention of falls and their consequences in the elderly. Eur Rev Aging Phys Act 2008;5: Lord SR, Sherrington C, Menz H, Close J. Falls in older people: risk factors and strategies for prevention. Cambridge: Cambridge University Pr; Halfon P, Eggli Y, Van Melle G, Vagnair A. Risk of falls for hospitalized patients: a predictive model based on routinely available data. J Clin Epidemiol 2001;54: Stalenhoef PA, Diederiks JPM, Knottnerus JA. A risk model for the prediction of recurrent falls in community-dwelling elderly: a prospective cohort study. J Clin Epidemiol 2002;55: Tromp AM, Pluijm SMF, Smit JH, Deeg DJH, Bouter LM, Lips P. Fall-risk screening test: a prospective study on predictors for falls in community-dwelling elderly. J Clin Epidemiol 2001;54: Rose DJ, Lucchese N, Wiersma LD. Development of a multidimensional balance scale for use with functionally independent older adults. Arch Phys Med Rehabil 2006;87: Steffen TM, Hacker TA, Mollinger L. Age- and gender-related test performance in community-dwelling elderly people: Six- Minute Walk Test, Berg Balance Scale, Timed Up & Go Test, and gait speeds. Phys Ther 2002;82: Shumway-Cook A, Baldwin M, Polissar NL. Predicting the probability for falls in community-dwelling older adults. Phys Ther 1997;77: Hernandez D, Rose DJ. Predicting which older adults will or will not fall using the Fullerton Advanced Balance Scale. Arch Phys Med Rehabil 2008;89: Berg KO, Wood-Dauphinee SL, Williams JI. Measuring balance in the elderly: validation of an instrument. Can J Public Health 1992;83:S Brauer SG, Burns YR, Galley P. A prospective study of laboratory and clinical measures of postural stability to predict communitydwelling fallers. J Gerontol A Biol Sci Med Sci 2000;55:M Muir SW, Berg K, Chesworth B, Speechley M. Use of the Berg Balance Scale for predicting multiple falls in community-dwelling elderly people: a prospective study. Phys Ther 2008;88: Boulgarides LK, McGinty SM, Willett JA, Barnes CW. Use of clinical and impairment-based tests to predict falls by communitydwelling older adults. Phys Ther 2003;83: Trueblood PR, Hodson-Chennault N, McCubbin A, Youngclarke D. Performance and impairment-based assessments among community dwelling elderly: sensitivity and specificity. J Geriatr Phys Ther 2001;24: Dite W, Temple VA. A clinical test of stepping and change of direction to identify multiple falling older adults. Arch Phys Med Rehabil 2002;83: Tiedemann A, Shimada H, Sherrington C, Murray S, Lord S. The comparative ability of eight functional mobility tests for predicting falls in community-dwelling older people. Age Ageing 2008;37: Zettel JL, McIlroy WE, Maki BE. Effect of competing attentional demands on perturbation-evoked stepping reactions and associated gaze behavior in young and older adults. J Gerontol A Biol Sci Med Sci 2008;63: Hsiao-Wecksler ET, Robinovitch SN. The effect of step length on young and elderly women s ability to recover balance. Clin Biomech 2007;22: Schulz BW, Ashton-Miller JA, Alexander NB. Maximum step length: relationships to age and knee and hip extensor capacities. Clin Biomech 2007;22: St George RJ, Fitzpatrick RC, Rogers MW, Lord SR. Choice stepping response and transfer times: effects of age, fall risk, and secondary tasks. J Gerontol A Biol Sci Med Sci 2007;62: Maki BE, Edmondstone MA, McIlroy WE. Age-related differences in laterally directed compensatory stepping behavior. J Gerontol A Biol Sci Med Sci 2000;55:M Berg WP, Alessio HM, Mills EM, Tong C. Circumstances and consequences of falls in independent community-dwelling older adults. Age Ageing 1997;26: Hernandez D, Rose DJ. Predicting which older adults will or will not fall using the Fullerton Advanced Balance Scale. Arch Phys Med Rehabil 2008;89: Tucker MG, Kavanagh JJ, Morrison S, Barrett RS. What are the relations between voluntary postural sway measures and fallshistory status in community-dwelling older adults? Arch Phys Med Rehabil 2010;91: Accident Compensation Corporation. Preventing injuries from falls: the national prevention strategy Wellington: ACC; Tinetti ME, Speechley M, Ginter SF. Risk-factors for falls among elderly persons living in the community. N Engl J Med 1988;319: Topolski TD, LoGerfo J, Patrick DL, Williams B, Walwick J, Patrick MB. The Rapid Assessment of Physical Activity (RAPA) among older adults. Prev Chronic Dis 2006;3: Shumway-Cook A, Brauer S, Woollacott M. Predicting the probability for falls in community-dwelling older adults using the Timed Up & Go Test. Phys Ther 2000;80: Schaubert KL, Bohannon RW. Reliability of the Sit-to-Stand test over dispersed test sessions. Isokinet Exerc Sci 2005;13: Hopkins WG. A new view of statistics Available at: Accessed November 23, Fleiss J. The design and analysis of clinical experiments. New York: Wiley; Davidson M. The interpretation of diagnostic tests: a primer for physiotherapists. Aust J Physiother 2002;48: Blennerhassett JM, Jayalath VM. The Four Square Step Test is a feasible and valid clinical test of dynamic standing balance for use in ambulant people poststroke. Arch Phys Med Rehabil 2008;89: Rockwood K, Awalt E, Carver D, MacKnight C. Feasibility and measurement properties of the Functional Reach and the Timed Up and Go tests in the Canadian study of health and aging. J Gerontol A Biol Sci Med Sci 2000;55:M Lord SR, Ward JA, Williams P, Anstey KJ. Physiological factors associated with falls in older community-dwelling women. J Am Geriatr Soc 1994;42: Muir SW, Berg K, Chesworth B, Speechley M. Use of the Berg Balance Scale for predicting multiple falls in community-dwelling elderly people: a prospective study. Phys Ther 2008;88: Buchner DM, Hornbrook MC, Kutner NG, et al. Development of the common data base for the FICSIT trials. J Am Geriatr Soc 1993;41: O Loughlin JL, Robitaille Y, Boivin JF, Suissa S. Incidence of and risk factors for falls and injurious falls among the communitydwelling elderly. Am J Epidemiol 1993;137: Hanlon JT, Landerman LR, Fillenbaum GG, Studenski S. Falls in African American and white community-dwelling elderly residents. J Gerontol A Biol Sci Med Sci 2002;57:M473-8.