Journal of Exercise Physiologyonline 104 December 2015 Volume 18 Number 6 Editor-in-Chief Official Research Journal of Tommy the American Boone, PhD, Society MBA of Review Exercise Board Physiologists Todd Astorino, PhD Julien Baker, ISSN 1097-9751 PhD Steve Brock, PhD Lance Dalleck, PhD Eric Goulet, PhD Robert Gotshall, PhD Alexander Hutchison, PhD M. Knight-Maloney, PhD Len Kravitz, PhD James Laskin, PhD Yit Aun Lim, PhD Lonnie Lowery, PhD Derek Marks, PhD Cristine Mermier, PhD Robert Robergs, PhD Chantal Vella, PhD Dale Wagner, PhD Frank Wyatt, PhD Ben Zhou, PhD Official Research Journal of the American Society of Exercise Physiologists ISSN 1097-9751 JEPonline Functional Fitness Test for Screening the Risk of Falls in the Elderly: Using Decision Tree Technique Tachapon Tongterm 1, Areerat Suputtitada 2, Chaipat Lawsirirat 1, Prawit Janwantanakul 3 1 Faculty of Sports Science, Chulalongkorn University, Bangkok, Thailand, 2 Department of Rehabilitation Medicine, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand 3 Department of Physical Therapy, Faculty of Allied Health Science, Chulalongkorn University, Bangkok, Thailand ABSTRACT Tongterm T, Suputtitada A, Lawsirirat C, Janwantanakul P Functional Fitness Test for Screening the Risk of Falls in the Elderly: Using Decision Tree Technique. JEPonline 2015;18(6):104-111. The aim of this study was to make the Functional Fitness Test for screening the risk of falls in the elderly by using the Decision Tree Technique. Selected by Multistage Random Sampling, we analyzed elderly subjects in 64 districts in 16 provinces in Thailand. The crosssectional study consisted of 5,632 elderly (of which 1,869 were men and 3,763 were women). The subjects physical fitness parameters (muscle strength, aerobic endurance, flexibility, and agility/balance) were studied using six different fitness tests: Chair Stand, Arm Curl, Step in Place, Chair Sit-and-Reach, Back Scratch, and the 8 Foot Up and Go. The fall risk parameter assessed was Berg Balance Scale (BBS). The analysis of Functional Fitness Test for screening the risk of falls in the elderly used the Decision Tree Technique algorithm C4.5 (J48). The 8 Foot Up and Go, Chair Stand, and Step in Place predicted the risk of falls in the elderly with accuracy in prediction (Correctly Classified Instances) equal to 95.76%. When using the Berg Balance Scale test as the dependent variable, the BBS was 45 points (meaning, low risk of fall) and when it was <45 points (a high risk of fall) for classifying the risk of falls in the elderly, with Decision Tree Technique model gave information that functional fitness in the 8 Foot Up and Go, Chair Stand, and Step in Place classified the risk of falls in the elderly with level of Correctly Classified Instances equal to 95.76%. Key Words: Functional fitness, Falls, Elderly, Decision tree
105 INTRODUCTION It is common knowledge that falls are a frequent occurrence in the elderly. Fortunately, many of the risk factors have been identified. For example, there are balance considerations, activity levels, the effects of medication, and the loss of muscle strength. Early detection is important along with exercise program to strengthen the musculoskeletal system. Often, the result is a significant reduction in the risk of falling among elderly persons. Early acknowledgment of impairments and functional limitations associated with reduced mobility, deconditioning, and risk of injury is critical to decreasing falls. Also, physical examination of the neurological system relative to clinical gait analysis and joint motions is necessary to help prevent future falls. Hence, there is clear evidence that screening of diseases and fall risk factors in the elderly in conjunction with properly prescribed exercise programs will help to improve overall quality of life by reducing the frequency and serverity of falls. Functional fitness consists of muscular strength, aerobic endurance, flexibility, agility, and balance. With the right exercise program to build functional fitness, the risk of falling in the elderly can be decreased (3,9,12). In fact, Toraman and Yildirim (11) reported that the risk of falls is highly correlation with balance and muscular strength of arms and legs. In agreement, Shin and colleagues (10) reported that the elderly who fall less often than the elderly with fall problems have a more efficient muscular system with better balancing skills. Although Rogers et al. (9) reported that muscle strength test and performance measures that consist of balance test, functional reach test, limits of stability, the 8 Foot Up and Go, and walking around two cones test are helpful in assessing the risk of falls in the elderly, there is still inconsistency in the knowledge and application of physical fitness test in determining the risk of falls in the elderly population. Hence, the purpose of this study was to analyze the functional fitness test for screening the risk of falls in the elderly by using the Decision Tree Technique. METHODS Subjects Multistage random sampling was used to identify 5,632 Thai elderly males and females who were between 60 and 99 yrs of age. The subjects consisted of 1,869 males and 3,763 females from 64 districts in 1 6 provinces throughout Thailand (selected by Multistage Random Sampling). The subjects gave consent to participate in the study and finished the test battery within 1 day. The ethics review committee for research involving human research subjects of Chulalongkorn University approved the study protocol. All subjects signed an informed consent form prior to the study getting underway. Procedures Six different fitness tests (Chair Stand, Arm Curl, Step in Place, Chair Sit-and-Reach, Back Scratch, and the 8 Foot Up and Go) were used to determine the subjects muscle strength, aerobic endurance, flexibility, and agility/balance. The Berg Balance Scale (BBS) was used to assess the subjects fall risk parameters. The analysis of functional fitness test for screening the risk of falls in the elderly was carried out using the Decision Tree Technique algorithm C4.5 (J48).
Functional Fitness Test Researchers selected lists of six functional fitness test from the Senior Fitness Test (5,7,8) as follows Chair Stand, Arm Curl, Step in Place, Chair Sit-and-Reach, Back Scratch, and the 8 Foot Up and Go. Assessment of Fall Risk The Berg Balance Scale (BBS) was used to determine the risk of falls during a series of predetermined tasks. It is a 14 item list with each item consisting of a five-point ordinal scale ranging from 0 to 4, with 0 indicating the lowest level of function and 4 the highest level of function and takes approximately 20 minutes to complete. 1. Sitting to Standing 2. Standing Unsupported 3. Sitting Unsupported 4. Standing to Sitting 5. Transfers 6. Standing with Eyes Closed 7. Standing with Feet Together 8. Reaching Forward with Outstretched Arm 9. Retrieving Object from Floor 10. Turning to Look Behind 11. Turning 360 12. Placing Alternate Foot on Stool 13. Standing with One Foot in Front 14. Standing on One Foot A score of 45 has been shown to be an appropriate cut-off for safe independent ambulation and the need for assistive devices or supervision (1,2,4,6). A score of 56 indicates functional balance. A score of <45 indicates individuals may be at greater risk of falling. 106 Statistical Analysis The raw data of 5,632 collections were analyzed in Weka 3.6.12 program (open source software for data mining) using the Decision Tree Technique algorithm C4.5 (J48), which is a cross validation technique to verify by random sampling with the data divided into group. In case of k-fold cross validation was halve the data to k part and calculated k time, each time used the data of k as a testing and used the data k-1 for learning data. RESULTS The cross-sectional study consisted of 5,632 elderly (of which 1,869 were men and 3,763 were women). The subjects age range was 60 to 99 yrs of age. Their mean BMI was 23.70 ± 4.46 kg m -2 with 458 elderly (8.13%) with a BMI of <18.5 kg m -2, 1,854 elderly (32.92%) with a BMI of 18.5 to 22.9 kg m -2, 3,320 elderly (58.95%) with a BMI of 23 kg m -2. The functional fitness tests for the elderly classified by age range are presented in Table 1. The BBS scores for the elderly classified by age range are presented in Table 2. The analysis of functional fitness test for screening risk of falls for the elderly using the Decision Tree Technique algorithm C4.5 (J48) is presented in Figure 1.
107 Table 1. Functional Fitness Tests for the Elderly Classified by Age Range. Functional Fitness Test 60-64 yrs )n = 1,633( 65-69 yrs )n = 1,449( 70-74 yrs )n = 1,105( Back Scratch (cm) Mean ± SD Mean ± SD Mean ± SD - left hand on top -13.7 ± 12.2-14.8 ± 12.7-18.0 ± 12.7 - right hand on top -8.1 ± 11.9-9.0 ± 12.3-12.2 ± 12.8 Chair Sit-and-Reach (cm) - stretch the left leg 11.2 ± 7.4 11.5 ± 7.8 10.1 ± 8.1 - stretch the right leg 11.1 ± 7.6 11.6 ± 7.7 10.2 ± 7.8 8 Foot Up and Go (sec) 7.5 ± 2.2 7.8 ± 2.3 8.8 ± 3.4 Arm Curl (number of curls) - left arm 19.9 ± 4.7 19.9 ± 4.7 18.2 ± 3.8 - right arm 19.5 ± 4.7 19.5 ± 4.7 17.8 ± 3.7 Chair Stand (number of stands) 15.4 ± 4.1 15.3 ± 4.1 13.9 ± 3.6 Step in Place (number of steps) 90.9 ± 19.6 89.3 ± 19.2 82.4 ± 19.3 Functional Fitness Test 75-79 yrs )n = 848( 80-84 yrs )n = 394( >85 yrs )n = 203( Back Scratch (cm) Mean ± SD Mean ± SD Mean ± SD - left hand on top -19.4 ± 13.0-22.3 ± 12.6-25.9 ± 13.7 - right hand on top -13.2 ± 13.5-15.6 ± 14.1-19.6 ± 14.5 Chair Sit-and-Reach (cm) - stretch the left leg 9.9 ± 8.2 8.9 ± 7.1 5.5 ± 9.2 - stretch the right leg 10.0 ± 7.7 9.5 ± 6.8 6.4 ± 8.4 8 Foot Up and Go (sec) 9.7 ± 4.2 11.9 ± 5.2 14.6 ± 7.6 Arm Curl (number of curls) - left arm 17.5 ± 4.3 15.9 ± 4.1 14.5 ± 4.1 - right arm 17.0 ± 4.3 15.6 ± 3.7 14.1 ± 3.6 Chair Stand (number of stands) 13.1 ± 3.9 12.1 ± 5.0 10.6 ± 3.7 Step in Place (number of steps) 78.3 ± 21.9 71.0 ± 23.5 63.9 ± 26.0 Table 2. BBS Score for the Elderly Classified by Age Range. BBS Score 60-64 yrs )n = 1,633( 65-69 yrs )n = 1,449( 70-74 yrs )n = 1,105( Normal (BBS = 45-56) 19(1.16) 18(1.31) 46(4.16) High Risk (BBS <45) 1,614(98.84) 1,431(98.69) 1,059(95.84) BBS Score 75-79 yrs )n = 848( 80-84 yrs )n = 394( >85 yrs )n = 203( Normal (BBS = 45-56) 65(7.67) 71(18.02) 62(30.54) High Risk (BBS <45) 783(92.33) 323(81.98) 141(69.46) *BBS Scores = Berg Balance Scale, number (%)
108 Figure 1. The Analysis of Functional Fitness Test for Screening Risk of Falls in the Elderly using the Decision Tree Technique Algorithm C4.5 (J48). The analysis of the Functional Fitness Test for screening the risk of falls in the elderly found that: (a) balance and agility tested by the 8 Foot Up and Go test; (b) strength of lower extremities tested by the Chair Stand; and (c) cardiorespiratory endurance tested by the Step in Place can be used to determine the risk of falls (Figure 1). The Berg Balance Scale produced two important findings. First, when the 8 Foot Up and Go test was >11 sec and the Chair Stand was >9 times, the elderly subjects were classified as a regular group with a low of risk of falls (with a BBS score >45 points). When the 8 Foot Up and Go was >11 sec and the Chair Stand was <9 times, and the Step in Place was >55 times, the elderly subjects were classified as a regular group with a low of risk of falls (with a BBS score >45 points). When the 8 Foot Up and Go was >11 sec and the Chair Stand was >9 times, and the Step in Place was <55 times, the elderly subjects were classified as a risk group with a high of risk of falls (with a BBS score <45 points). Second, when the 8 Foot Up and Go was <11 sec, the elderly subjects were classified as a regular group with a low of risk of falls (with a BBS score >45 points).
109 Table 3. Confusion Matrix of Model for Functional Fitness Criteria in Screening the Risk of Falls for the Elderly. A B <-- Classified as 5,205 146 a = A 102 179 b = B Class A referred to Berg Balance Scale ( 45 points). Class B referred to Berg Balance Scale (<45 points). Table 4. Detailed Accuracy by Class of Model for Functional Fitness Criteria in Screening the Risk of Falls for the Elderly. TP Rate FP Rate Precision Recall F-Measure ROC Area Class 0.97 0.36 0.98 0.97 0.98 0.91 A 0.64 0.03 0.55 0.64 0.59 0.91 B 0.96 0.35 0.96 0.96 0.96 0.91 Weighted Avg. Class A referred to Berg Balance Scale ( 45 points). Class B referred to Berg Balance Scale (<45 points). DISCUSSION The 8 Foot Up and Go, Chair Stand, and Step in Place predicted the risk of falls for the elderly with the accuracy in prediction (Correctly Classified Instances) equal to 95.76%. The 8 Foot Up and Go was used to assess the subjects agility and dynamic balance. Both agility and dynamic are important in tasks that require quick maneuvering, such as getting off a bus in time or getting up to attend to something in the kitchen, to go to the bathroom or to answer the phone (5,7,8). The Chair Stand was used to assess lower body strength, which is important for numerous tasks such as climbing stairs, walking, and getting out of a chair, tub or car (5). The Step in Place is the alternate aerobic endurance test. This is important when space limitations or weather prohibits taking the 6-Minute Walk Test (7,8). The Arm Curl was used to assess the subjects upper body strength, which is needed to perform household and other activities involving the lifting and carrying of groceries, suitcases, and grandchildren (7,8). The Chair Sit-and-Reach was used to assess lower body flexibility that is important for good posture, for normal gait patterns, and for various mobility tasks, such as getting in and out of a bathtub or car (5,7,8). The Back Scratch was used to assess subjects upper body (shoulder) flexibility, which is important when combing hair, putting on overhead garments, and reaching for a seat belt (5,7,8). The Berg Balance Scale (BBS) was developed to measure balance among older people with impairment in balance function by assessing the performance of functional tasks. It is a valid instrument used for evaluation of the effectiveness of interventions and for quantitative descriptions of function in clinical practice and research. The BBS has been evaluated in several reliability studies. A recent study of the BBS, which was completed in Finland, indicates that a change of eight (8) BBS points is required to reveal a genuine change in function between two assessments among older subjects who are dependent in activities of daily living (ADL) and living in residential care facilities (1,2,4,6).
110 CONCLUSIONS This model was taken from the Berg Balance Scale as the independent variable. The findings indicate that lower body muscle strength, cardiorespiratory endurance, balance, and agility in the elderly are associated with a high risk of fall. The 8 Foot Up and Go, Chair Stand, and Step in Place predicted the risk of falls for the elderly with the accuracy in prediction (Correctly Classified Instances) of 95.76%. If the elderly has a risk of fall by all 3 tests, it is imperative that exercises are recommended to improve lower body muscle strength, cardiorespiratory endurance, balance, and agility. ACKNOWLEDGMENTS This research was supported via scholarship from the Office of the Higher Education Commission of Thailand, the National Research Council of Thailand, and the 90th Anniversary of Chulalongkorn University (Ratchadaphisek somphot Endowment Fund). Address for correspondence: Professor Areerat Suputtitada, MD, Department of Rehabilitation Medicine, Faculty of Medicine, Chulalongkorn University, King Chulalongkorn Memorial Hospital, Rama 4 Road, Pathumwan, Bangkok 10330, Thailand, Email: sareerat1111@gmail.com REFERENCES 1. Berg K, Wood-Dauphinee S, Williums JI. Measuring balance in the elderly: Validation of an instrument. Can J Public Health. 1992;83(2):S7-S11. 2. Berg K, Wood-Dauphinee S, Williums JI. Measuring balance in the elderly: Preliminary development of an instrument. Physiother Can. 1989;41(6):304. 3. Costello E, Edelstein JE. Update on falls prevention for community-dwelling older adults: Review of single and multifactorial intervention programs. J Rehabil Res Dev. 2008;45(8):1135-1152. 4. Donoghue D. How much change is true change? The minimum detectable change of the Berg Balance Scale in elderly people. J Rehabil Med. 2009;41(5):343-346. 5. Jones CJ, Rikli RE. Measuring functional fitness in older adults. J Active Aging. 2002;March April:25-30. 6. Kornetti DL, Fritz SL, Chiu YP, Light KE, Velozo CA. Rating scale analysis of the Berg Balance Scale. Arch Phys Med Rehabil. 2004;85(7):1128-1135. 7. Rikli RE, Jones CJ. Development and validation of functional fitness test for community-residing older adults. J Aging Phys Act. 1999a;7:129-161.
8. Rikli RE, Jones CJ. Functional fitness normative scores for community-residing older adults, ages 60-94. J Aging Phys Act. 1999b;7:162-181. 9. Rogers ME, Rogers NL, Takeshima N, Islam MM. Methods to assess and improve the physical parameters associated with fall risk in older adults. Prev Med. 2003;36(3): 255-264. 10. Shin KR, Kang Y, Jung D, Kim M, Lee E, A comparative study on physical function test between faller group and nonfaller group among community-dwelling elderly. Asian Nurs Res. 2012;6(1):42-48. 11. Toraman A, Yıldırım N. The falling risk and physical fitness in older people. Arch Gerontol Geriatr. 2010;51(2):222-226. 12. Tuunainen E, Rasku J, Jäntti P, Pyykkö I. Risk factors of falls in community dwelling active elderly. Auris Nasus Larynx. 2014;41(1):10-16. 111 Disclaimer The opinions expressed in JEPonline are those of the authors and are not attributable to JEPonline, the editorial staff or the ASEP organization.