Fall in older adults is a fast-growing global public health concern. The socioeconomic impact of falls includes increased health care costs and burden on the family or caregivers.1 Falls are the most common cause of injuries at home resulting in loss of function.2
It has been predicted that, globally, 30% to 40% of older adults aged 65 years and older experience a fall annually.3,4 In the case of Malaysia, by 2035, 15% of its population will be aged 65 years and older.5 Already, between 27% and 34% of Malaysian community-dwelling older adults aged 60 years and older experienced a fall in the past year.6,7 The incidence and risk of falls was higher among older adults with chronic illnesses (47%) and who live in institutions (98%).8 These figures will increase in the near future.
Information regarding the determinants and correlates of falls in older adults can assist in the development of prevention and management strategies. So far, the known sociodemographic determinants of falls in this population are age,6 gender, and a previous history of falls.9 In addition, low levels of education, polypharmacy, visual impairments, urinary incontinence, and the use of such medications as laxatives and antipsychotics have been identified as risk factors for falls in hospitalized older adults.10
Knowing which physical performance measures are correlates of falls in older adults is useful as these measures can be used as fall screening tools. The ability to balance has been identified as a determinant of falls in Malaysian older adults.6 Gait and balance impairments have been reported to be risk factors for falls among older patients.10 Instrumental Activities of Daily Living (IADL),11 Timed Up and Go (TUG) Test,12 Gait Speed Test, Walking While Talking (WWT),13,14 and Physiological Profile Assessment (PPA)6 are some of the other physical performance measures that may predict the risk of falls among community-dwelling other adults.
It is noteworthy that the source of most epidemiological data regarding falls was cross-sectional studies or falls history. Moreover, information regarding combined sociodemographic data and physical performance measures as determinants and correlates of falls among community-dwelling older adults is limited. Thus, the aim of our present study was to identify whether sociodemographic (age, gender, previous history of falls, the number of medications, and education level) factors and physical performances measures (TUG, GS, WWT, and PPA) were significant determinants and correlates for subsequent falls through prospective study (with up to 6 months' follow-up) of Malaysian community-dwelling older adults.
Sample size calculations showed that 326 participants were required to demonstrate significant changes according to power analysis, set at 80%. The sample size calculation was performed manually on the basis of the formula of single-proportion sample size with 25% estimated prevalence for health problems, a margin of error of 0.05% and confidence interval of 95%.
Three hundred twenty-five community-dwelling older adults (145 men and 180 women), aged 60 to 89 (M = 67.64, SD = 5.56) years, were recruited through a multistage random sampling for this prospective study. This study is part of a larger study, the LRGS TUA, a prospective study of aging focusing on a wide range of neuroprotective factors among a population-based sample of Malaysian older adults.15 The participants for this study were from the state of Selangor, which is located on the west of Peninsular Malaysia with a total area of 7930 km2. The participants were randomly recruited from 9 districts by using a multistage sampling method.16 The districts were chosen as they had the highest percentage of older adults in the state. Participants aged 60 years and older, living in the community and ambulating independently with or without assistive device, were recruited on the basis of the sampling frame taken from the National Population and Housing Census 2010 data, provided by Department of Statistics Malaysia.
Individuals not living in a community and had scores of Mini-Mental State Examination (MMSE) lower than 20 or Geriatric Depression Scale (GDS) of 5 and more were excluded. GDS was used to determine depressive symptoms. Its score ranges are 0-4, 5-8, 9-11, 12-15, indicating no presence of depressive symptoms, mild, moderate, and severe depression, respectively. Individuals who were on prescribed medications that could potentially affect physical function and balance, had impaired physical function due to such conditions as musculoskeletal and neurological disorders or malignancy, or were unable to follow instructions in Malay, English, Mandarin, Cantonese, or Tamil were also excluded.
Ethical approval was obtained from the Secretariat for Research and Ethics, Universiti Kebangsaan Malaysia (UKM 220.127.116.11/244/NN-060-2013). Verbal and written information regarding the study was provided to the participants and informed written consent was obtained prior to participation in the study. After the participants were recruited, a structured questionnaire was used to obtain their demographic data, socioeconomic conditions, and a detailed past medical history. The IADL questionnaire,17 which assesses an individual's ability to use a telephone, shop, prepare food, do housekeeping and laundry, use transportation, take responsibility for medications and handle finances, was administered to determine functional status. Lower total scores indicate a higher level of dependency.
Participants were required to complete a falls diary for a period of 6 months. This diary was prestamped and addressed to the Assessment Center in order to facilitate its return by post. Monthly calls were made to remind participants to record any falls they had experienced. The definition of a fall was “an event which resulted in a person coming to rest inadvertently on the ground or floor or other lower level.”1(p1) Participants were then assessed, using physical performance measures, by trained physiotherapists and field workers (who were trained by and under the supervision of a physiotherapist). The measurements included TUG, Gait Speed (GS), WWT Tests, and PPA. The participants were allowed to rest between all assessments and whenever required.
Timed Up and Go18
The TUG Test measures the time in seconds for participants to rise from a standard chair, walk 3 m, turn, walk back, and sit. Participants were instructed to perform this test using their regular footwear and walking aids. No physical assistance was allowed during the test and the type of walking aid used was noted. TUG Test has an excellent test-retest reliability (ICC = 0.98) among older adults.12
When instructed to “go,” the participants walked at their normal pace along a flat, marked surface, 10 m in length. Gait speed was recorded in seconds using a stop watch. Recording began when the participants' feet touched the 2 m line and stopped when the feet touched the 8 m line, to account for acceleration and deceleration of walking speed. The participants were allowed to use their assistive device (if they used one regularly) and the type of assistive device used was recorded. Gait speed was calculated by dividing 6 m with the time taken to complete the test. The reliability of GS is high (ICCs ≥ .90) among older adults.20
Walk While Talking21
In the WWT test, participants were required to count backwards, starting from 30, while walking at a normal pace on a 10-m line flat surface, without prioritizing either task. The time taken to complete the test was recorded in seconds using a stop watch. Participants were allowed to count either in Malay, English, Chinese (Cantonese, Mandarin, Hokkien), or Tamil language. The score for the Walk while Talking Test was calculated by dividing the distance of 10 m by the time taken to complete the test. Any counting mistakes were not corrected. WWT test with varied levels of complexity have been reported to be a valid and reliable test to identify fallers, with a sensitivity and specificity between 39%-46% and 89%-96%, respectively.
Physiological Profile Assessment22
The PPA, Short Version was used to quantify the risk of falls. The PPA consists of 5 tests, namely, visual contrast sensitivity, peripheral sensation, joint position, muscle strength, hand reaction time, and postural sways. PPA has a reliability of ≥0.50 and a 75% accuracy in determining falls among older adults. Fall risk was calculated using the PPA Web-based program (FallScreen). Falls risk scores were categorized on the basis of PPA scores that range between −2 and 3, representing a very low, low, small, moderate, marked, and very marked risk of falls.
Out of 385 older adults, 325 (84.4%) met the inclusion criteria. Fifty-seven individuals (16.6%) were excluded as their MMSE scores were less than 20 or their GDS score were more than 5. After 6 months of follow-up, (55.8%; n = 170) participants returned their falls diary. Participants who did not return their falls diary after 6 months' follow-up were contacted via phone calls and their falls diary information was recorded. In total, 94% (n = 305) of the participants' falls status were obtained. Results showed that 26.6% (n = 81/305) of the participants had falls during the 6-month period. As shown in Table 1, there were significant differences in the education level (P = .01), TUG (P = .03), and PPA (P = .001) scores between fallers and nonfallers.
Prior assumption tests were conducted and outliers were removed. Multicollinearity and logit linearity tests showed that there was no violation for the tests. Further, binary logistic regression analysis using the enter method was performed on all the sociodemographic and physical performance measures to determine the predictors of falls (Table 2). The omnibus model for the logistic regression analysis was statistically significant, χ2(df = 11, N = 31) = 63.81, P < .01, Cox and Snell R2 = 0.21, Nagelkerke R2 = 0.31. The model was 76.6% accurate in its predictions of falls. Hosmer and Lemeshow test result confirmed that the model was a good fit for the data χ2(df = 8, N =305) = 10.80, P = .21. PPA was the only measure that significantly improved the model's predictive ability. The result indicates that one-unit increase in PPA score increased the risk of falls by 3.2 times.
Table 3 shows a further binary logistic regression test without PPA in the model. The omnibus model for the logistic regression analysis was statistically significant, χ2 (df = 10, N = 305) = 18.30, P < .05, Cox and Snell R2 = 0.07, Nagelkerke R2 = 0.95. The model was 74.1% accurate in its predictions of falls. Hosmer and Lemeshow test result confirmed that the model was a good fit for the data χ2 (df = 8, N =305) = 4.77, P = .78.
The level of education (2) and TUG were factors that improved the model's predictive ability. The odd ratio for education level (2) indicated that if the older adults' education level increased by 1 unit, there was a predicted 55% reduction in the probability of having a fall. Older adults who took longer to perform TUG were found to be at risk of fall by 1.1 times (OR: 1.070, 95% CI: 1.017-1.125).
Identifying determinants and correlates of falls among community-dwelling older adults can assist with the prediction of early falls and the development of management strategies. Our study results showed that the PPA is the strongest correlate measure of falls among community-dwelling older Malaysians. Although less robust, the TUG Test appears to be a correlate of falls in this population with the absence of PPA in the model. The PPA has been previously shown to be a robust fall risk measure among community-dwelling older adults.22 In addition, there is a significant correlation between PPA and TUG test.23 Similarly, in our earlier study, we demonstrated that PPA is associated with objectively measured static balance test.6 Although simple to use, PPA consist of a few tests and has to be administered by a trained personnel. It is also noteworthy that participants in our present study were multicultural Asian older population compared to Western older adults in previous studies.
TUG and education levels were found to be significantly different (P < .05) in fallers and nonfallers. However, both TUG test and education levels were only found to be able to predict falls among community-dwelling older without PPA in the model. Our study results are consistent with recent reports,24–28 suggesting that TUG has limited to moderate predictive ability to detect future falls in older adults. Older adults with higher levels of education have a lower risk of falls in our second model. This is supported by previous reports.10,29 Higher education levels among older adults may be associated with better cognitive ability, health, physical and socioeconomic status that then affect the level of knowledge regarding falls and the older adults' engagement in the prevention of falls. In our recent studies, we have shown that a combination of TUG and other measures of falls (compared to using TUG alone) was a better predictor for the prediction.30,31
The other measures used in this study—IADL, WWT, and GS—did not appear to be correlated with falls among community-dwelling older adults. This may be due to the fact that community-dwelling older adults have higher physical functioning levels. Thus, IADL, WWT, and GS may not be sensitive measures of falls in the case of community-dwelling older adults. Moreover, most physical performance measures such as IADL, WWT, and GS were designed to detect certain levels of impairments in older adults.
Sociodemographic characteristics such as age, gender, past history of falls, and medications have been associated with falls in previous studies on other populations.9,10,32 However, in our present study, with PPA combined with other physical performance measures and sociodemographic characteristics, PPA dominates the model. This may be because the 5 physiological tests comprised in PPA accounts for all the other fall risk factors in the model. It is also noteworthy that participants in our present study included older adults aged 60 years and older. Participants in previous studies included older adults aged 65 years and above.
A limitation of this study is that it was limited to 4 states in Peninsular Malaysia. However, we used multistage random sampling that allows us to generalize our results to other older Malaysian adults. Another limitation may be the limited duration of 6 months' follow-up. This may have resulted in some falls not being detected. Nevertheless, we obtained prospective falls records from diaries that were recorded daily by participants with regular follow-up reminders by phone calls.
Approximately 27% of the participants experienced falls in the 6-month follow-up period. The PPA was the best physical performance correlate of falls among Malaysian community-dwelling older adults. However, the TUG as a simple physical performance tool may be used to predict falls among community-dwelling older adults when the PPA is not available. In addition, the TUG, combined with such sociodemographic data as education levels, is more cost-effective and simpler to use compared with the PPA when performing large-scale community falls screening.
1. World Health Organization. WHO global report on falls
prevention in older age. Geneva, Switzerland: World Health Organization; 2007.
2. Lim K, Jasvindar K, Normala I, et al Risk factors of home injury among elderly people in Malaysia. Asian J Gerontol Geriatr. 2014;9(1):16–20.
4. Rubenstein LZ. Falls
in older people: epidemiology, risk factors and strategies for prevention. Age Ageing. 2006;35(2):37–41. doi:10.1093/ageing/afl084.
6. Singh DKA, Pillai SGK, Tan ST, Tai CC, Shahar S. Association between physiological fall risk and physical performance
tests among community-dwelling older adults
. Clin Interv Aging. 2015;10:1319–1326. doi:10.2147/CIA.S79398.
7. Rizawati M, Mas Ayu S. Home environment and fall at home among the elderly in Masjid Tanah Province. J Univ Malaya Med Cent. 2008;11(2):72–82.
8. Singh DK, Manaf ZA, Yusoff NAM, Muhammad NA, Phan MF, Shahar S. Correlation between nutritional status and comprehensive physical performance
measures among older adults with undernourishment in residential institutions. Clin Interv Aging. 2014;9:1415–1423. doi:10.2147/CIA.S64997.
9. Rafiq M, McGovern A, Jones S, et al Falls
in the elderly were predicted opportunistically using a decision tree and systematically using a database-driven screening tool. J Clin Epidemiol. 2014;67(8):877–886. doi:10.1016/j.jclinepi.2014.03.008.
10. Abreu HC, Reiners AA, Azevedo RC, Abreu DR, Silva AM, Oliveira AD. Incidence and predicting factors of falls
of older inpatients. Public Health Pract. 2015;49:37. doi:10.1590/S0034-8910.2015049005549.
11. Brown J, Kurichi JE, Xie D, Pan Q, Stineman MG. Instrumental activities of daily living staging as a possible clinical tool for fall risk assessment in PM&R. Natl Inst Health. 2014;6(4):316–323. doi:10.1016/j.pmrj.2013.10.007.
12. 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(9):896–903.
13. Abellan Van Kan G, Rolland Y, Andrieu S, et al Gait Speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J Nutr Heal Aging. 2009;13(10):881–889. http://www.ncbi.nlm.nih.gov/pubmed/19924348
14. Ayers EI, Tow AC, Holtzer R, Verghese J. Walking While Talking and falls
in aging. Gerontology. 2014;60(2):108–113. doi:10.1038/nature13314.A.
15. Shahar S, Omar A, Vanoh D, et al Approaches in methodology for population-based longitudinal study on neuroprotective model for healthy longevity (TUA) among Malaysian older adults. Aging Clin Exp Res. 2016;28(6):1089–1104. doi:10.1007/s40520-015-0511-4.
17. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9(3):179–186. doi:10.1093/geront/9.3_Part_1.179.
18. Podsiadlo D, Richardson S. The Timed Up & Go: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142–148. doi:10.1111/j.1532-5415.1991.tb01616.x.
19. Bohannon RW. Comfortable and maximum walking speed of adults aged 20-79 years: reference values and determinants. Age Ageing. 1997;26:15–19. http://ageing.oxfordjournals.org/
20. 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(2):128–137.
21. Verghese J, Buschke H, Viola L, Katz M, Hall C, Kuslansky G, Lipton R. Validity of divided attention tasks in predicting falls
in older individuals: a preliminary study. J Am Geriatirc Soc. 2002;50(9):1572–1576. doi:10.1046/j.1532-5415.2002.50415.x.
22. Lord SR, Menz HB, Tiedemann A. A physiological profile approach to fall risk assessment and prevention. Phys Ther. 2003;83(3):237–252.
23. Whitney JC, Lord SR, Close JC. Streamlining assessment and intervention in a falls
clinic using the Timed Up and Go Test and Physiological Profile Assessments. Age Ageing. 2005;34(6):567–571.
24. Kojima G, Masud T, Kendrick D, Morris R, Gawler S, Treml J, Iliffe S. Does the Timed Up and Go Test predict future falls
among British community-dwelling older people? Prospective cohort study nested within a randomised controlled trial. BMC Geriatr. 2015;15:38. doi:10.1186/s12877-015-0039-7.
25. Barry E, Galvin R, Keogh C, Horgan F, Fahey T. Is the Timed Up and Go test a useful predictor of risk of falls
in community dwelling older adults: a systematic review and meta-analysis. BMC Geriatr. 2014;14(1):14. doi:10.1186/1471-2318-14-14.
26. Gates S, Smith LA, Fisher JD, Lamb SE. Systematic review of accuracy of screening instruments for predicting fall risk among independently living older adults. J Rehabil Res Dev. 2008;45(8):1105–1116. doi:10.1682/JRRD.2008.04.0057.
27. Schoene D, Wu SM, Mikolaizak AS, Menant JC, Smith ST, Delbaere K, Lord SR. Discriminative ability and predictive validity of the Timed Up and Go Test in identifying older people who fall: systematic and meta-analysis. Am Geriatr Soc. 2013;61(2):202–208. doi:10.1111/jgs.12106.
28. Beauchet O, Fantino B, Allali G, Muir SW, Montero-Odasso M, Annweiler C. Timed Up and Go Test and risk of falls
in older adults: a systematic review. J Nutr Heal Aging. 2011;15(10):933–938. doi:10.1007/s12603-011-0062-0.
30. Ibrahim A, Singh DKA, Shahar S. “Timed Up and Go” test: age, gender and cognitive impairment stratified normative values of older adults. PLoS One. 2017;12(10):e0185641. doi:10.1371/journal.pone.0185641
31. Ibrahim A, Singh DKA, Shahar S, Omar MA. Timed Up and Go test combined with self-rated multifactorial falls
risk questionnaire and sociodemographic factors predicts falls
among community-dwelling older adults
better than the Timed Up and Go test on its own. J Multidiscipl Healthc. 2017;10:409–416. doi:10.2147/JMDH.S142520
32. Slattum P, Ansello EF. Medications as a risk factor in falls
by older adults with and without intellectual disabilities. Age Action. 2013;28(1):1–6.