Aging Clin Exp Res DOI 10.1007/s40520-014-0273-4

ORIGINAL ARTICLE

Balance performance in older adults and its relationship with falling Mir Mohammad Jalali • Hooshang Gerami Abtin Heidarzadeh • Robabeh Soleimani



Received: 8 April 2014 / Accepted: 1 October 2014 Ó Springer International Publishing Switzerland 2014

Abstract Background and aims A normal consequence of aging is a general deterioration in a number of musculoskeletal and sensory systems that affect postural control and balance. The aim of this study was to evaluate history of falls among active older individuals in Iran, and estimate the risk factors for falls among this population. Methods A total of 448 active older subjects from rural region of Rasht city, Iran, were included. They were divided into three groups depending on their age: youngold (n = 266); middle-old (n = 154) and oldest-old (n = 28). We assessed balance performance by One-Leg Balance (OLB), Functional Reach (FR), Timed Up and Go (TUG) and Romberg tests. Results The fall rate ([2 in the last year) was 27.0 %. The cut-off point 13.75 s for TUG test showed 84.7 % sensitivity and 56 % specificity. Also the best cut-off point for OLB test was 12.7 s (63 % sensitivity and 83.5 % specificity). Logistic regression analysis revealed that age, BMI, diabetes, and failure in OLB, FR, and Romberg tests predicted fall risk. The decision tree classification of older individuals showed three categorical variables, which in

M. M. Jalali (&)  H. Gerami Department of Otolaryngology, Otolaryngology Research Center, Amiralmomenin Hospital, Guilan University of Medical Sciences, 41396-38459 Rasht, Iran e-mail: [email protected] A. Heidarzadeh Department of Community Medicine, Medical Faculty, Guilan University of Medical Sciences, Rasht, Iran R. Soleimani Department of Psychiatry, Medical Faculty, Guilan University of Medical Sciences, Rasht, Iran

their order of importance included diabetes, Romberg test, and OLB test. Conclusions This study revealed the value of history taking about diabetes as a predictor for existing falling. Decision tree technique showed that Romberg and OLB tests help in identifying older adults with balance problems. Given the incidence and consequences of falls among older adults, large-scale prospective studies on older individuals to identify those prone to falls are warranted. Keywords Postural balance  Accidental falls  Aged  Decision tree

Introduction With a general increase in life expectancy, the population of older adults is increasing throughout the developed world [1], with newly industrialized countries of Asia showing more rapid aging compared with other countries. The annual rate of increase of the older population (65 years or more) in these countries is reported to be approximately 3 %, compared with 1.0–1.3 % in the United Kingdom, Sweden, and the United States [2]. Individuals aged 65 years or more accounted for 6 % of the total population of the Islamic Republic of Iran in 2005, while estimates for 2030 placed the proportion of this age group at 19 % [3]. Injurious fall events requiring acute medical attention were estimated at 143.1, 336.7, and 848.3 per 100,000 person-years among males, and 190.2, 416.8, and 854.7 per 100,000 person-years among females of the age groups 60–69, 70–79, and 80?, respectively, in Iran in 2003 [4]. The biological aging process includes certain changes in the musculoskeletal and neuromuscular systems, which

123

Aging Clin Exp Res

affect complex motor performance and tend to increase the incidence of falls. Falls have detrimental effects on the well-being and quality of life of older adults and their family caregivers [5], and are a leading cause of injury and death among older adults; therefore, falls are considered a significant public health issue [6]. On an annual basis, falls affect one in three adults over the age of 65 [7] and 50 % of the adults over the age of 80 [8]; 20–30 % of these patients suffer moderate to severe injuries, which interfere with their ability to live independently in the community, necessitate hospitalization, or result in increased risk of death [9]. According to Tinnetti et al. [10], two-thirds of fall incidents in older adults are potentially preventable; thus, identification of significant risk factors is an important step towards fall prevention [11, 12]. Although several studies have reported various risk factors for falls, the results have been inconclusive. Previous studies have shown the association of old age [13], female gender [14], history of falls [15], visual problems [16], physical frailty [17], depressive disorder [18], polypharmacy [19], use of psychotropic medications [20], and certain long-term conditions [21] with increased risk of falls over a period of 90 days to 1 year. The Guideline for the Prevention of Falls in Older Persons [15] states that inquiries about two or more instances of falls in the past 12 months should be conducted during the initial screening to identify potential fallers in need of intervention. In addition, the Guideline suggests that risk assessment is not necessary for older individuals who reported only a single instance of falling in the absence of gait disturbances. In contrast, older individuals presenting with recurrent falls should be subject to an in-depth evaluation. Taking efficiency and cost-effectiveness into consideration, interventions should preferably be focused on older individuals at a higher risk of falling [22, 23]. Therefore, a strategy to identify the older individuals who face a higher risk of falling is required. Various balance tests are employed for assessing impairments in the components of postural control among older people. The assessment process requires clinical tests that are easy to administer, have demonstrated predictive validity for future falls, and have acceptable properties for use in clinical settings [5, 24, 25]. Therefore, health care professionals require a simple and practical clinical approach for the identification of older adults at an elevated risk of falls. Previous studies on the epidemiology and risk assessment of falls have been predominantly based on Caucasian populations from developed countries, with few studies focusing on the older population in Iran. The aim of the present work is to: (a) evaluate history of falls among active older individuals in Iran; (b) estimate the risk factors for falls among the older adults [including age, gender,

123

body mass index (BMI), and certain chronic medical diseases]; (c) compare the results of balance performance (BP) tests among older individuals, with or without a history of falls; and (d) compare the frequency of impaired balance tests according to different host factors. To this end, we have developed a decision tree for predicting the probability of falls among older individuals.

Methods Study design This retrospective observational study was carried out from March 2012 to March 2013. The participants of the study included older individuals residing in Rasht, a large city in the Guilan Province of north Iran and located 300 km from Tehran (the capital of Iran). Data from the last census in the year 2000 was used for estimating that approximately 18,770 individuals residing in Rasht are aged 65 years or more. Sample size was calculated on the basis of a previous report by Melzer [26], who estimated that 30 % of the aging population suffered from some form of disability. By using the prevalence of fall 30 %, taking the precision of 0.05 and a two-sided 5 % significance level, the calculated sample size was 350. The random samples were stratified by regional council area. Inclusion criteria were (a) age C65 years; (b) capability to perform functional balance tests; (c) absence of medical conditions, which would interfere with participation in the present study; and (d) being ambulatory without assistance (i.e., capable of walking without a cane or walker). Exclusion criteria were: (a) an established definitive diagnosis of neurological, psychiatric, or cognitive illnesses; (b) communication difficulties; (c) presence of any visual problems; and (d) incidence of an acute disease. All older individuals were examined at Rural Health Centers by final-year medical students from the Guilan University of Medical Sciences, who were trained to employ standardized instructions for administering the tests. The procedures were supported and approved by the Institutional Review Board (IRB) of Guilan University of Medical Sciences. Informed consent for participating in the study was obtained in writing from all the subjects. Data collection All participants were examined at Rural Health Centers. Demographic data were collected in order to describe the characteristics of the study sample (age, height, and weight). Information pertaining to the history of falls over the past year was also recorded. In this study, older individuals who reported at least 2 falls during the preceding

Aging Clin Exp Res

12-month period were considered recurrent fallers. A fall has been defined as any event that led to an unplanned, unexpected contact with a supporting surface [27]. Data pertaining to demographic characteristics, known diseases, and medication history were collected. A comprehensive physical examination was carried out following the interview. In addition, the risk of falls and static and dynamic body balance were assessed by the One-Leg Balance (OLB), Functional Reach (FR), Timed Up and Go (TUG), and Romberg tests. Balance assessment tools The One-Leg Balance (OLB) test The OLB test, employed for assessing balance, is a simple clinical test of static balance in which the subject is asked to stand unassisted on one leg. Impaired OLB, defined as the lack of ability to stand on one leg for 5 s or more, has been identified as a predictor of injurious falls among community-dwelling older individuals [28]. The participants were instructed to start in a position with a comfortable base of support, and then to stand unassisted on one leg (the dominant leg) and with open eyes. The OLB test was timed in seconds from the moment one foot was lifted off the floor to the time it touched the ground, with longer times indicating better balance capability. The best cut-off point for the OLB test was calculated to allow analyses with the OLB test as a binary variable. The functional reach (FR) test The FR test represents the maximal distance, which can be reached forward beyond arm’s length by a subject while maintaining a fixed base of support in the standing position. The test has shown high intra-rater and retest reliability in a sample of healthy subjects [29]. In this test, the subjects were instructed to stand with their feet a comfortable distance apart and behind a line perpendicular and adjacent to a wall. The arm closest to the wall was then raised to shoulder height, and the position of the tip of the middle finger was measured. The subjects were then instructed to keep their feet flat on the floor and lean forward as far as possible without losing balance, touching the wall, or taking a step. The position of the tip of the middle finger was then recorded at the point of furthest reach, and the difference between the two points was recorded as the maximal distance. Three measures were recorded on the same side, and mean score for each was calculated [29, 30]. Test results were interpreted as normal if the subject could confidently reach forward more than 25 cm.

The timed up and go (TUG) test The TUG test measures speed during several functional maneuvers, which include standing up, walking, turning, and sitting down, and requires both static and dynamic balance. High inter-rater and intra-rater reliability have been reported for this test among the older individuals [31, 32]. Subjects were seated in a normal armchair (45 cm high) with their back against the chair. They were instructed to stand up, walk 3 m as quickly and safely as possible past a line on the floor, turn around, walk back to the chair, and sit down once again with their back against the chair. Two trials were conducted with a rest in between; the time taken for each was recorded, and the mean scores were calculated [30, 32]. A score of 13.5 s (or 14 s as per some reports) indicates that the individual is likely to be prone to falling [32]. The cut-off value of 14 s was employed for the decision tree to identify older individuals at an elevated risk of falls. Romberg with Jendrassik maneuver Romberg’s test forms a part of the standard neurologic assessment, and a positive result indicates presence of a neurological disease. While its clinical application still lacks a standard approach, Romberg test with Jendrassik maneuver was employed in the present study. Patients were instructed to remain still with both feet together and closed eyes, and perform abduction of the upper limbs for a period of 30 s. The test was initiated when the patients assumed correct position and stopped when they moved their feet, lost the position of their upper limbs, or opened their eyes. The test was considered negative if patients maintained the position for the entire 30-s duration; otherwise, it was considered positive [33]. Statistical analysis Statistical analysis was performed using the software SPSS (Statistical Package for the Social Sciences) version 21.0 (IBM Corp, Armonk, NY). Descriptive analyses were performed for demographic and predictor variables. The prevalence of falls was calculated. Univariate analysis of the predictive accuracy of continuous predictor (TUG and OLB tests) variables was performed using receiver operating characteristic (ROC) curve, with area under the curve (AUC) employed as a measure of overall diagnostic value. AUC values of 0.5 and 1.0 represented low and perfect discriminatory abilities of a test, respectively. Such analysis simultaneously provided different cut-off points, as a consequence of different sensitivities and specificities, for different diagnostic tests. The optimal cut-off points for a parameter corresponding to maximal Youden’s index

123

Aging Clin Exp Res Table 1 Descriptive statistics of the sample (mean ± SD) or (%)

Characteristics

Young-old (65–74 years)

Middle-old (75–84 years)

Oldest-old (C85 years)

Number

266

154

28



Age (year)

69.52 ± 2.99

78.57 ± 2.84

87.78 ± 2.43



BMI (kg/cm2)

26.52 ± 5.29

27.90 ± 5.36

26.67 ± 5.28

0.492a 0.004b

Gender, n (%) Male

124

96

18

Female

142

58

10 0.043b

Falls history, n (%) No a b

ANOVA Chi-square

p value

207 (77.83)

103 (66.9)

17 (60.7)

1–12 in the last year

40 (15.03)

31 (20.1)

8 (28.6)

[12 in the last year

19 (7.14)

20 (13.0)

3 (10.7)

(J = sensitivity ? specificity - 1) were selected. The cutoff points (and any other significant covariates) were then employed in logistic regression analysis to establish the risk of an individual being a faller or nonfaller. Univariate analyses were performed for predictors of falls using Chi-square statistics for categorical and Student’s t test for continuous data. Logistic regression analyses for predictors of falls were performed for variables, which were significant in univariate analyses. A value of p \ 0.05 (two-tailed) was considered statistically significant. A decision tree was employed as a data mining method. The analysis assigned equal misclassification costs for false positives and false negatives. The following predictor variables were considered for a history of falling: gender, age, BMI, osteoarthritis, diabetes, and balance assessment tools (OLB, TUG, FR, and Romberg tests). Of these, age and BMI were continuous scale variables, whereas the other variables (including balance performance tests) were binary variables. Decision trees are a rapid and effective method of classifying data set entries, and offer a powerful solution to classification and prediction problems [34].

The mean age of the participants was 73.78 ± 6.3 years, and the most prevalent ages ranged between 65 and 74 years (59.38 %). The subjects were categorized into three age groups [32]: young-old (n = 266; 65–74 years), middle-old (n = 154; 75–84 years) and oldest-old (n = 28; C85 years). Sensitivity and specificity Of the balance assessment tools, AUC was computed for TUG and OLB tests. For the TUG times, the cut-off value of 13.75 s yielded the best combination of sensitivity and specificity. The TUG results were impaired in 84.7 % of patients with a history of falling (sensitivity). TUG results were normal in 56 % of patients without a history of falling (specificity). The positive and negative predictive values of the TUG test were found to be 29.9 and 94.3 %, respectively. For the OLB test, the best cut-off point suggested with this sample set was 12.7 s (63 % sensitivity and 83.5 % specificity), with positive and negative predictive values of 90.4 and 47.6 %, respectively. Univariate analyses

Results Demographics A total 486 subjects that fulfilled the inclusion and exclusion criteria were approached during study period. However, only 448 subjects consented. Thus, the participation rate of the study was 92.2 %. The present study included 209 female and 239 male participants. All participants were physically active and free from musculoskeletal injury or psychiatric disorders at the time of testing. Characteristics of the participants are shown in Table 1.

123

One hundred and twenty-one participants (27.0 %) had a history of falls (C2 falls the preceding year). Univariate analyses revealed that several variables showed a significant association with the occurrence of falls (Table 2), including gender, age, BMI, osteoarthritis, and diabetes. The patients with a history of falls were more likely to be female, of the oldest-old age group, and to have osteoarthritis and diabetes. In addition, older persons with a history of falls were more likely to obtain impaired balance scores than those with no history of falls (Table 2). The correlations between the FR value and BMI and osteoarthritis were found to be 0.337 (p = 0.004) and

Aging Clin Exp Res Table 2 Main baseline characteristics of the study population and univariate ORs (n = 448) Potential predictor variables

Without falls history (n = 327)

With falls history (n = 121)

n

n

%

%

OR (CI 95 %)

p value

0.048a

Demographic characteristics Gender Male

183

76.9

55

23.1

1

Female

144

68.6

66

31.4

1.53 (1.00–2.32)

Age (year), mean (SD) 2

Body mass index (kg/cm ), mean (SD)

73.23 (6.12)

75.26 (6.49)

0.002b

26.09 (5.13)

29.35 (5.37)

0.043b

Life style factors Smoking No

201

70.8

83

29.2

1

Yes

64

73.6

23

26.4

0.87 (0.51–3.70)

No

218

72.9

81

27.1

1

Yes

48

65.8

25

34.2

1.41 (0.81–2.44)

No

262

71.4

105

28.6

Yes

3

0.614a

Addiction 0.248a

Alcohol consumption 100

0

0

0.561a

Chronic disease Hearing loss No

154

77.0

46

23.0

1

Yes

122

71.3

49

28.7

1.34 (0.84–2.15)

No

256

73.8

91

26.2

1

Yes

71

70.3

30

29.7

1.19 (0.73–1.92)

No

253

70.9

104

29.1

1

Yes

15

78.9

4

21.1

0.65 (0.21–2.00)

No

259

71.5

103

28.5

1

Yes

9

64.3

5

35.7

1.39 (0.46–4.35)

No

195

92.9

15

7.1

Yes

100

66.7

50

33.3

6.66 (1.61–25)

No

309

82.8

64

17.2

1

Yes

18

24.0

57

76.0

15.38 (8.40–27.78)

[12.70 s

189

90.4

20

9.6

B12.70 s

111

52.4

101

47.6

\13.75 s

182

94.3

11

5.7

C13.75 s

143

70.1

61

29.9

6.98 (1.42–34.29)

[25 cm

286

89.7

33

10.3

1

B25 cm

39

50.0

39

50.0

8.67 (2.26–33.29)

[20 s

243

85.0

43

15.0

1

B20 s

57

42.5

77

57.5

7.53 (4.58–12.38)

0.214a

Heart disease 0.488a

Lung disease 0.448a

Kidney disease 0.555a

Osteoarthritis 1

0.004a

Diabetes 0.001a

Balance assessments OLB test 1

0.001a

8.54 (4.86–14.99)

TUG test 1

0.008a

FR test 0.002a

Romberg 0.001a

ORs odds ratios, CI confidence interval, SD standard deviation, OLB test One-Leg Balance test, TUG test Timed Up and Go test, FR test Functional Reach test a

Chi-square

b

t test

123

Aging Clin Exp Res Table 3 Probability (P) of impaired One-Leg Balance (OLB), Timed Up and Go (TUG), Functional Reach (FR) and Romberg tests according to host factors

Variable Host factor

OLB

TUG

Impaired

p value

Male

45.4

0.04

Female

56.2

a

FR

Impaired

p value

81.5

0.25

a

Impaired

Romberg p value

a

Impaired

p valuea

33.3

0.51

Gender 77.1

9.7

0.07

26.8

30.2

Age group (year) B75

50.5

[75

50.0

0.92

74.1

0.002

88.3

17.2

0.18

73.5

29.7

0.32

34.5

Body mass index (kg/m2) B30

50.5

[30

57.1

0.60

45.7

0.35

56.8

5.7

0.004

32.4

31.5

0.37

42.9

Osteoarthritis No Yes

50.4 47.4

0.80

No

43.2

0.001

Yes

86.9

42.9 63.3

0.08

87.4

0.001

9.5 33.3

0.01

19.0

1

29.0 41.9

0.02

30.6

0.41

Diabetes

a

Chi-square

40.0

Table 4 Logistic regression by the forward method for the risk factor for falls (n = 447) Variable Age

Ba

Standard error

Odds

p value

0.09

0.02

-0.09

0.001

Body mass index

-0.24

0.08

0.27

0.003

Diabetes

-2.66

0.37

13.28

0.001

OLB test

-1.55

0.33

3.71

0.001

FR test

-1.55

0.73

3.71

0.03

Romberg test

-1.37

0.31

2.94

0.001

a

Unstandardized coefficients

-0.297 (p = 0.004), respectively, while the other variables failed to influence FR test results among the study participants (Table 3). A correlation of 0.438 (p \ 0.001) and 0.344 (p \ 0.002) were observed between the TUG test values and diabetes and age, respectively. Variables such as gender and diabetes were found to influence the OLB test results, with correlations of 0.107 (p = 0.04) and -0.322 (p \ 0.001), respectively. Statistically significant correlations of 0.337, -0.297, and -0.302 were observed between Romberg test results and BMI, osteoarthritis, and diabetes, respectively. The correlations of age and balance assessment tools (except TUG) were insignificant, as shown (Table 3). Logistic regression analysis The variables considered for this model were determined from the analysis of individual risk factors for falls, and include the variables shown in Table 2 with probability values of \0.05. The final model, shown in Table 4,

123

21.4

35.2

included age, BMI, diabetes, and failure in OLB, FR, and Romberg tests. Decision tree Data for 378 of the 448 older participants were implemented in the decision tree. The resulting classification tree graphically summarized the predictors, in the order of their discriminatory power relative to each other. The decision tree classification of these older individuals consisted of three categorical variables, which in their order of importance included diabetes, Romberg test, and OLB test. Each node in Fig. 1 shows the probability of a history of falls for the older participants. Based on the existent interactions between predictor variables, four rules were extracted from the decision tree classification corresponding to four leaf nodes. The risk estimate of 0.157 indicates that the correct category is predicted by the model and that the model accurately classifies approximately 84.3 % of the older participants.

Discussion The pragmatic goal of the present study was to compare the clinical characteristics and balance performance test results of older individuals and assess the predictive factors for falls. The strength of this work is the objective measurements of four balance performance tests, and a comparative study design. To the best of our knowledge, this is the first population-based study in Iran to determine the predictive factors for falls among older individuals.

Aging Clin Exp Res

Fig. 1 Decision tree classification of older patients

Falling is one of the major problems leading to reduced levels of daily activities and decreased quality of life among the older individuals [35]. A small percentage of the falls (10–15 %) results in major soft tissue injury, while 5 % results in fracture [36]. Thus, reduction in fall-related morbidity, mortality, and functional deterioration is an urgent problem [37]. In the present study, 27.0 % of the older subjects were found to have a history of falling, with lower values for males (23.1 %) compared to females (31.4 %). Zhang and Chen [38] reported significantly lower fall incidence for males (10.3–14.7 %) compared to females (20.2–23.8 %), and for males aged 65 years or more (21–23 %) compared to females of the same age group (43–44 %). Four tests were employed for objective measurement of balance performance. Moreover, ROC curves were constructed and their AUCs were calculated for comparing the diagnostic usefulness of TUG and OLB tests. The odds for a history of falling were seven times greater when the TUG test value was C13.75 s. On the other hand, the odds ratio of the OLB test was 4.26 (p \ 0.001). Logistic regression analysis revealed that OLB test with the employed cut-off level accurately identified older adults with a history of falling. The present study showed that compared to younger elderly adults, the older elderly subjects showed poorer BP on the basis of the four clinical tests, which is in line with previous studies [29, 30, 39, 40]. Correlation was not identifiable between age and balance assessment tools, suggesting that this variable may have high dispersion. This finding is consistent with previous studies on the characteristic heterogeneity of aging [41, 42]. The extensive applications and results of the FR test have been discussed by Lin and Liao, who failed to observe a clear association between the test and alterations due to the aging process [43]. A difference between genders was observed with respect to BP based on the OLB and FR tests in both groups (with and without history of falling), but not in the TUG and Romberg tests. Similarly, Takahashi et al. [27] reported the absence of differences based on gender in the FR and TUG values of subjects aged over 65 years. Similar

observations were made by Steffen et al. [40] for the TUG test among older individuals aged 61–89 years. On the other hand, several reports have described differences between the two genders in BP assessments among older adults using different tests [39, 44]. Cultural and sociodemographic factors are likely to be responsible for the different results obtained in balance tests in the present study compared to previous ones. For example, in the present study, osteoarthritis was found in 41.7 % of the older subjects; moreover, better results in the TUG, OLB, and Romberg tests were obtained for individuals without diabetes, while the incidence of diabetes could influence neuropathy-related issues, and therefore, the risk of falls. The results revealed a significant effect of diabetes on the rate of falls. The logistic regression analyses of this sample set revealed that advanced age, high BMI, and diabetes were clinical risk factors for falls. Old age and gender were the most commonly reported risk factors, and recurrent fallers tended to be older and predominantly women. Nevertheless, gender was not identified as a variable that significantly affected risk of falls in multivariate analysis. Moreover, osteoarthritis was found to be associated with the risk of falls in univariate but not multivariate models, which is attributable to the collinearity of osteoarthritis with factors bearing stronger effects such that its effect is attenuated. The small sample size is also likely responsible for rendering this variable insignificant in the adjusted model. Physical examination-based risk factors included failure in the OLB, FR, and Romberg tests. Failure in TUG test did not increase the risks of falls in a statistically significant manner in logistic regression analyses; this is in agreement with the observations of Bongue et al. [45] that the TUG test failed to accurately predict risk of falls among healthy community-dwelling older adults. In contrast, instability in OLB test was found to be a risk factor for falls in that study. The use of a decision tree model enabled the identification of risk factors for falls on three different levels. Diabetes was selected as the first partitioning variable. Older individuals with diabetes showed the highest risk of

123

Aging Clin Exp Res

falling (90.2 %), indicating that diabetes affects the physiological functioning of various systems and is an important contributor to the risk of falling. The second factor in the model was Romberg test results, which were found to be associated with fall risk in previous studies [46, 47]. Increased risk (39.5 %) associated with a failed Romberg test was observed among older individuals without diabetes. The final factor for categorizing individuals with high risk included the results of OLB test. Individuals who failed the OLB test showed a 14.6 % risk of falls, whereas those who did not showed 5.9 % risk. The model showed that older individuals without diabetes and normal Romberg and OLB test results were the least likely to fall. This group, constituting 28.6 % of the total participants, represented the healthiest subgroup of the older population, such that a dedicated healthcare intervention for the prevention of falls is unlikely to significantly alter risk of falls in this group. The decision tree showed that FR test is not useful for discriminating subjects at greater risk of falling, regardless of the predictive value of this test in logistic regression analyses. The FR test is known to be influenced more by trunk flexibility than by center-of-pressure displacement [48], and may therefore not be a true measure of balance. Of the four tests employed, TUG was the only one that included a gait component, which is functionally important because many falls occur during ambulation [49]. Despite observations in previous studies [50–52] that TUG test could discriminate between fallers and nonfallers, the current study failed to reveal any predictive value of TUG test in logistic regression analyses. The best cut-off point for the TUG test was found to be 13.75 s in the present study. The accuracy (R true positive ? R true negative/R total population) and precision (R true positive/R true positive ? false positive) of the TUG test were found to be 61.2 and 29.9 %, respectively. The low precision of the TUG test likely results in its insignificant predictive value. The OLB test assesses only one aspect of balance—that of maintaining equilibrium on a reduced base of support, and shows good accuracy and precision. The high specificity (83.5 %) of the OLB test allows for correct identification of ‘‘true negatives’’. In the decision tree model, the OLB test forms the last node and is best employed for the identification of nonfallers. This model therefore has implications for the design and implementation of fall prevention interventions. The classification tree also provides practical advantages. Notably, a limited number of predictors need to be measured. In the risk profiles developed using logistic regression, all predictors included in the model need to be measured for identifying the risk of falls. However, with the classification tree, not all predictors in the tree need to be measured, because specific combinations of predictors identify different end groups. Moreover, the classification

123

tree does not contain a complex algorithm, and no score need to be calculated, because the risk profile identifies an end group with matching risk for recurrent falls. The present study has certain limitations. Retrospective studies are susceptible to under-reporting of falls, particularly falls, which did not result in injury. In a prospective study of 304 ambulatory patients, Cummings et al. [53] found that 13–32 % of the patients denied having fallen, depending on how long after the event they were questioned; longer intervals were associated with lower recall. Thus, the predictive validity of clinical examination results for identifying future fallers, which is the clinically relevant question, cannot be evaluated with certainty. Moreover, causation is unlikely to be revealed in retrospective studies, which is an important limitation. Significant association was observed in the present study between diabetes, results of Romberg and OLB tests, and a history of falls. However, retrospective data collection could result in confounding results, as seen with the research design of Tinetti and Williams [54]. Therefore, there is a requirement for further studies that employ a prospective approach for testing the predictors of falls.

Conclusions This study aimed to assess multiple aspects that contributed to increased risk of falls among older adults. Balance assessment tools such as OLB, TUG, FR, and Romberg tests were found to be appropriate for predicting balance problems in older adults. Moreover, this study also revealed that a history of diabetes could be employed as a predictor for falls with a high level of confidence. Decision Tree technique showed that Romberg and OLB tests help in identifying older adults with balance problems. However, the importance of a careful examination of patient history and comprehensive physical examination cannot be overemphasized in the clinical setting, unless the results of the present study are reproduced in larger randomized clinical trials. Large-scale prospective studies in the future on this subject are therefore warranted. Acknowledgments The authors wish to thank Saeidi Nora, Nickpour Mina, and Karimi Zahra for carrying out the measurements. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.

References 1. World Health Organization (1998) Resolution of the Executive Board of the WHO on health promotion. Health Promot Int 13:266

Aging Clin Exp Res 2. Ng TP, Niti M, Chiam PC, Kua EH (2006) Prevalence and correlates of functional disability in multiethnic elderly Singaporeans. J Am Geriatr Soc 54:21–29 3. Amir-Sadri A, Soleimani H (2005) Elderly phenomena and its outcomes in Iran. Int J Hyg Health 1:19–35 4. Abolhassani F, Moayyeri A, Naghavi M, Soltani A, Larijani B, Shalmani HT (2006) Incidence and characteristics of falls leading to hip fracture in Iranian population. Bone 39:408–413 5. Owings TM, Pavol MJ, Foley KT, Grabiner PC, Grabiner MD (1999) Exercise: is it a solution to falls by older adults? J Appl Biomech 15(1):56–63 6. Tinetti ME (2003) Clinical practice: preventing falls in elderly persons. N Engl J Med 348(1):42–49 7. Hausdorff JM, Rios DA, Edelberg HK (2001) Gait variability and fall risk in community-living older adults: a 1-year prospective study. Arch Phys Med Rehabil 82(8):1050–1056 8. Inouye SK, Brown CJ, Tinetti ME (2009) Medicare nonpayment, hospital falls, and unintended consequences. N Engl J Med 360(23):2390–2393 9. Alexander BH, Rivara FP, Wolf ME (1992) The cost and frequency of hospitalization for fall-related injuries in older adults. Am J Public Health 82:1020–1023 10. Tinnetti ME, Speechley M (1989) Prevention of falls among the elderly. N Engl J Med 320(16):1055–1059 11. Gill TM, Williams CS, deLeon CFM, Tinetti ME (1997) The role of change in physical performance in determining risk for dependence in activities of daily living among nondisabled community-living elderly persons. J Clin Epidemiol 50(7):765–772 12. Gill TM, Williams CS, Tinetti ME (2000) Environmental hazards and the risk of nonsyncopal falls in the homes of communityliving older persons. Med Care 38(12):1174–1183 13. Lin CH, Liao KC, Pu SJ, Chen YC, Liu MS (2011) Associated factors for falls among the community-dwelling older people assessed by annual geriatric health examinations. PLoS One 6(4):e18976 14. Bekibele CO, Gureje O (2010) Fall incidence in a population of elderly persons in Nigeria. Gerontology 56:278–283 15. Chu LW, Chi I, Chiu AY (2005) Incidence and predictors of falls in the Chinese elderly. Ann Acad Med Singap 34:60–72 16. Panel on prevention of falls in older persons, American Geriatrics Society and British Geriatrics Society (2011) Summary of the updated American Geriatrics Society/British Geriatrics Society clinical practice guideline for prevention of falls in older persons. J Am Geriatr Soc 59:148–157 17. Ensrud KE, Ewing SK, Cawthon PM, Fink HA, Taylor BC, Cauley JA, Dam TT, Marshall LM, Orwoll ES, Cummings SR (2009) A comparison of frailty indexes for the prediction of falls, disability, fractures, and mortality in older men. J Am Geriatr Soc 57:492–498 18. Cesari M, Landi F, Torre S, Onder G, Lattanzio F, Bernabei R (2002) Prevalence and risk factors for falls in an older community-dwelling population. J Gerontol A Biol Sci Med Sci 57:M722–M726 19. Russell MA, Hill KD, Blackberry I, Day LL, Dharmage SC (2006) Falls risk and functional decline in older fallers discharged directly from emergency departments. J Gerontol A Biol Sci Med Sci 61:1090–1095 20. Bloch F, Thibaud M, Dugue B, Breque C, Rigaud AS, Kemoun G (2011) Psychotropic drugs and falls in the elderly people: updated literature review and meta-analysis. J Aging Health 23:329–346 21. Kwan MM, Close JC, Wong AK, Lord SR (2011) Falls incidence, risk factors, and consequences in Chinese older people: a systematic review. J Am Geriatr Soc 59:536–543 22. Gardner MM, Robertson MC, Campbell AJ (2000) Exercise in preventing falls and fall related injuries in older people: a review of randomized controlled trials. Br J Sports Med 34:7–17

23. Rizzo JA, Baker DI, McAvay G, Tinetti ME (1996) The costeffectiveness of a multifactorial targeted prevention program for falls among community elderly persons. Med Care 34:954–969 24. American Geriatrics Society, British Geriatrics Society, and American Academy of Orthopedic Surgeons panel on falls prevention (2001) Guideline for the prevention of falls in older persons. J Am Geriatr Soc 49:664–772 25. Yeung TS, Wessel J, Stratford PW, MacDermid JC (2008) The timed up and go test for use on an inpatient orthopaedic rehabilitation ward. J Orthop Sports Phys Ther 38:410–417 26. Melzer D, McWilliams B, Brayne C, Johnson T, Bond J (1999) Profile of disability in elderly people: estimates from a longitudinal population study. BMJ 318(7191):1108–1111 27. Takahashi T, Ishida K, Yamamoto H, Takata J, Nishinaga M, Doi Y, Yamamoto H (2006) Modification of the functional reach test: analysis of lateral and anterior functional reach in communitydwelling older people. Arch Gerontol Geriatr 42:167–173 28. Vellas BJ, Wayne SJ, Romero L, Baumgartner RN, Rubenstein LZ, Garry PJ (1997) One-leg balance is an important predictor of injurious falls in older persons. J Am Geriatr Soc 45(6):735–738 29. Duncan PW, Weiner DK, Chandler J, Studenski SA (1990) Functional reach: a new clinical measure of balance. J Gerontol A Biol Sci Med Sci 45(6):192–197 30. Isles RC, Choy NL, Steer M, Nitz JC (2004) Normal values of balance tests in women aged 20–80. J Am Geriatr Soc 52:1367–1372 31. Podsiadlo D, Richardson S (1991) The timed ‘Up & Go’: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 39:142–148 32. Shumway-Cook A, Brauer S, Woollacott M (2000) Predicting the probability for falls in community-dwelling older adults using the timed-up and go test. Phys Ther 80:896–903 33. Vereeck L, Truijen S, Wuyts F, Van de Heyning P (2007) The Dizziness Handicap Inventory and its relationship with functional balance performance. Otol Neurotol 28:87–93 34. Kantardzic M (2003) Data mining: concepts, models, methods, and algorithms. Wiley, New York 35. Masud T, Morris RO (2001) Epidemiology of falls. Age Aging 30:3–7 36. Lehtola S, Koistinen P, Luukinen H (2006) Falls and injurious falls late in home-dwelling life. Arch Gerontol Geriatr 42:217–224 37. Masui T, Hasegawa Y, Matsuyama Y, Sakano S, Kawasaki M, Suzuki S (2005) Gender differences in platform measures of balance in rural community-dwelling elders. Arch Gerontol Geriatr 41:201–209 38. Zhang Y, Chen W (2008) Research overview and progress of the elderly falls. Chin J Gerontol 9:929–931 39. Brauer S, Burns Y, Galley PA (2000) A prospective study of laboratory and clinical measures of postural stability to predict community-dwelling fallers. J Gerontol A Biol Sci Med Sci 55A:M469–M476 40. Steffen TM, Hacker TA, Mollinger L (2002) Age- and genderrelated test performance in community dwelling elderly people: six-minute walk test, berg balance scale, timed up & go test, and gait speeds. Phys Ther 82:128–137 41. Costarella M, Monteleone L, Steindler R, Zuccaro S (2010) Decline of physical and cognitive conditions in the elderly measured through the functional reach test and the mini-mental state examination. Arch Gerontol Geriatr 50:332–337 42. Woollacoott M, Shumway-Cook A (2002) Attention and the control of posture and gait: a review of an emerging area of research. Gait Posture 16:1–14 43. Lin S, Liao C (2011) Age-related changes in the performance of forward reach. Gait Posture 33:18–22 44. Aslan UB, Cavlak U, Yagci N, Akdag B (2008) Balance performance, aging and falling: a comparative study based on a Turkish sample. Arch Gerontol Geriatr 46(3):283–292

123

Aging Clin Exp Res 45. Bongue B, Dupre´ C, Beauchet O, Rossat A, Fantino B, Colvez A (2011) A screening tool with five risk factors was developed for fall-risk prediction in community-dwelling elderly. J Clin Epidemiol 64(10):1152–1160 46. McMichael KA, Vander Bilt J, Lavery L, Rodriguez E, Ganguli M (2008) Simple balance and mobility tests can assess falls risk when cognition is impaired. Geriatr Nurs 29(5):311–323 47. Agrawal Y, Carey JP, Hoffman HJ, Sklare DA, Schubert MC (2011) The modified Romberg Balance Test: normative data in US adults. Otol Neurotol 32(8):1309–1311 48. Jonsson E, Henriksson M, Hirschfeld H (2003) Does the functional reach test reflect stability limits in elderly people? J Rehabil Med 35:26–30 49. Thomas JI, Lane JV (2005) A pilot study to explore the predictive validity of 4 measures of falls risk in frail elderly patients. Arch Phys Med Rehabil 86(8):1636–1640 50. Medley A, Thompson M (2014) Contribution of age and balance confidence to functional mobility test performance: Diagnostic

123

51.

52.

53.

54.

accuracy of L test and normal-paced Timed Up and Go. J Geriatr Phys Ther Boye´ ND, Mattace-Raso FU, Van Lieshout EM, Hartholt KA, Van Beeck EF, Van der Cammen TJ (2014) Physical performance and quality of life in single and recurrent fallers: data from the Improving Medication Prescribing to Reduce Risk of Falls study. Geriatr Gerontol Int Barry E, Galvin R, Keogh C, Horgan F, Fahey T (2014) Is the timed Up and Go test a useful predictor of risk of falls in community dwelling older adults: a systematic review and metaanalysis. BMC Geriatr 14:14 Cummings SR, Nevitt MC, Kidd S (1988) Forgetting falls. The limited accuracy of recall of falls in the elderly. J Am Geriatr Soc 36:613–616 Tinetti ME, Williams CS (1998) The effect of falls and fall injuries on functioning in community-dwelling older persons. J Gerontol A Biol Sci Med Sci 53(2):112–119

Balance performance in older adults and its relationship with falling.

A normal consequence of aging is a general deterioration in a number of musculoskeletal and sensory systems that affect postural control and balance. ...
290KB Sizes 2 Downloads 7 Views