Table of contents

Editorial
Meet the team
Dr Alan Walker
Original Contribution/Clinical Investigation
The discriminative value of two functional assessment scales compared for predicting falls in institutionalized elderly people in Iran
The relationship between depression and educational levels, the retirement years and chronic diseases of elderly men in area 6 of Tehran's municipality 2005
Evaluating Infectious Etiologies of Hospitalization in Elderly Population of Iran
Models and Systems of Elderly Care
The Economic Impact of Treating Geriatric Hip Fracture - A Study at Rustaq Referral Hospital , South Batinah Region, Oman
Clinical Research and Methods
Acceptable Satisfaction after Carpal Tunnel Decompression in Elderly Patients

 

The discriminative value of two functional assessment scales compared for predicting falls in institutionalized elderly
people in Iran

Authors
Ahmad Ali Akbari-Kamrani, MD.
Farhad Azadi, MsC.
Mahyar Salavati, PhD.
Behruz Kazemi, MD.
Behnam Shakiba, Medical student

Correspondence:
Ahmad Ali Akbari-Kamrani, MD.
Iranian Research Centre on Aging, University of Social Welfare and Rehabilitation, Tehran Iran.
Email: m_akbari@razi.tums.ac.ir
Telephone: +98 21 8877 7387
Fax: +98 21 2242 3244

Farhad Azadi, MsC.
Physiotherapist. Iranian Research Centre on Aging, University of Social Welfare and Rehabilitation, Tehran Iran.
Email: fa_azadi@yahoo.com
Telephone: +98 21 2242 3244
Fax: +98 21 2242 3244

Mahyar Salavati, PhD.
Iranian Research Centre on Aging, University of Social Welfare and Rehabilitation, Tehran Iran.
Email: salavati@uswr.ac.ir
Telephone: +98 21 2242 3244
Fax: +98-21 2242 3244

Behruz Kazemi, MD.
Department of internal medicine, Shiraz University of Medical Sciences. Shiraz, Iran.
Email: pmr@sums.ac.ir
Telephone: +98 21 2242 3244
Fax: +98 21 2242 3244

Behnam Shakiba, Medical student.
 Students' Scientific Research centre, Tehran University of Medical Sciences, Tehran, Iran.
Email: b_shakiba@razi.tums.ac.ir
Telephone: +98 21 8474 1125
Fax: +98 21 8474 1125
Address University of Social Welfare and Rehabilitation, Daneshjou Boulevard, Evin Ave, Tehran, Iran: P.O. Box: 198571383


Abstract:

Background: The purpose of the present study is to compare the discriminative values of two clinical assessments methods (Berg Balance Test and Tinetti Balance Scale) to identify fallers among institutionalized elderly people with/without history of falling in an elderly population in Iran.

Methods: Fifty-four subjects, 17 fallers and 37 non-fallers were evaluated throughout this study. The frequency of falls within 6 months was prospectively studied. At the end of 6 months, those having had two or more falls were enrolled in the faller group and those having no falls were enrolled in the non-faller group and then the Berg balance test and Tinetti Scale were administered for all participants in both groups.

Results: Independent t-test indicated a significant difference between the two groups of fallers and non-fallers in the mean scores on Berg Balance Scale (P=0.0001) and Tinetti Scale (P=0.0001). The results indicated much more discriminative value for Berg balance test scores than Tinetti.

Conclusions: This study demonstrates that the Berg Balance Scale assessment method is of more discriminative validity in differentiating those elderly people with history of fall from those without such history. Results from the following study would seem rather valuable as an assessment tool for health care professionals in the identification and monitoring of potential fallers within the elderly community.

Keywords: elderly, fall, Berg balance test, Tinetti scale.

Introduction

Decreased mortality and increased life expectancy has led to enlargement of the population of the elderly. At the beginning of 2000, the population of people older than 65 comprised one-eighth of the world's total population, i.e. 750 million people [1]. In light of this increase, health care professionals must address several problems occurring within this particular population. One of the most important and common problems is related to falls [2].

In 1998 the National Institute of Health reported, among Americans age 65 and older, fall-related injuries are the leading cause of death due to unintentional injuries [3]. The studies demonstrate that 25-47% of the elderly in the community have one or more falls per year and this even reaches 50% among those institutionalized. [1, 4]

Approximately in 5-20% of older persons who fall, the injuries suffered caused serious problems or death, and the psychological effects can lead to impaired mobility, loss of function, and an overall decrease in a person's quality of life [5]. The population lifetime cost of injuries associated with falls averages $12.6 billion, and the average hospital charge for fall related injuries in an older adult is $11,800 [6].

Fall prevention has, therefore been recognized as a priority area for research and intervention [7]. Many studies have reported several factors that have been found to be rather successful in the prediction of falls [8]. The most important of these factors include aging, chronic illness, sedentary lifestyle [9], orthopedic impairments, cardiac disorders [10], visual impairment [11], muscle weakness [12] and impaired balance [10], Balance has been shown to be an important predictor of falls within the elderly population [13]. Balance is required for maintaining a static posture, stabilizing dynamic movements, performing daily activities, and moving around in the community [14]. Throughout the years, several instruments have been developed as a means of quantitatively measuring balance in the elderly population [8]. Several performance balance measures, such as the timed up and go (TUG) [15], one-leg stand (OLS) [16], Activities-specific Balance Confidence (ABC) scale [17], Tinetti balance (TB) [18], and Berg balance test (BBT) [19], are available for evaluating community-dwelling older people but a standardized and valid screening instrument to identify people at risk of falling is still unavailable [20]. The Berg Balance test was designed to be an easy to administer, simple, safe and reasonably brief measurement of balance for elderly people [21]. The Tinetti balance scale (TB) is one part of the performance-oriented assessment of mobility problems [14] and this test is a simple clinical balance scale which measures characteristics associated with falls [20].

Some investigators have suggested that the BBT appears to be the best single predictor of fall status [22] and BBT have the best potential for detecting balance impairment [23]. Another study suggested that BBT is more appropriate for older people [20]. These suggestions were not justified by sufficient empirical evidence.

The purpose of the present study is to compare the discriminative values of these two clinical assessments methods (Berg Balance Test and Tinetti Balance Scale) to identify fallers among institutionalized elderly people with/without history of falling in an elderly population in Iran.


Methods

Study Subjects

Seventy elderly people institutionalized in the "Institute for the Elderly and Handicapped, Kahrizak, Tehran", independent in their daily activities and able to walk at least 10 meters with/without assistive devices volunteered to participate in this study after giving informed consent. Table 1 illustrates the descriptive statistics with regards to our study population. The participants were followed for 6 months, from September 2004 to March 2004, and the frequency of falls per month was registered by physiotherapists through direct monthly visits to the participants and the care personnel.

At the end of 6 months, eight of the participants having fallen once within the period were excluded in order to increase the precision of the study (n1=8), four participants died (n2=4), two were hospitalized in the special care unit for injuries due to falls and diseases (n3=2), one had become wheelchair-bound (n4=1), and one was unable to walk (n5=1).
The subject pool was subsequently divided into two groups one of which consisted of 17 subjects and was categorized as fallers while the other group consisted of 37 non-fallers. Inclusion criteria for the faller's category consisted of two or more falls within the study period. A fall was defined as an "event which results in a person coming to rest inadvertently on the ground or other lower level and other than as a consequence of a violent blow, loss of consciousness or sudden onset of paralysis" [24].

Procedure and Measurements

At the end of 6-month follow-up, demographic characteristics and medical history of the participants including age, gender, weight, height, number of medicines per day and number of the diseases, were collected by interview and referring to their medical files.
Then, each participant was assessed by an investigator (blinded to the allocation of participants) using the two assessment methods according to the instructions.

In Berg Balance Scale, 14 activity items, including standing up from seated position, standing without support, sitting without support, sitting down, transfer, standing without support with the eyes closed, standing without support with the legs fixed together, bending forward with the arms stretched out, lifting an object off the ground, turning to the left and the right and looking back, turning 360 degrees, touching the stool with the legs for several times, standing without support with the feet along each other, and standing on a single leg, were measured according to the instructions.

Each activity item was scored as 0-4, where score of 0 meant inability to perform the item, and score of 4 meant complete ability to perform the item. The total score of this method was 56.

In Tinetti Scale method, the following activity items were measured according to the instructions:

  • On balance: balance at sitting, standing up and effort to stand up, balance at immediate standing (the first 5 seconds), balance at standing and pushing, standing with the eyes closed, turning 360 degrees, and sitting down.
  • On gait: starting to walk, length, height, symmetry and succession of the steps, route, oscillation of the trunk, and width of gait.

Each activity item was scored 0-1 or 0-2, where a score of 0 meant inability to perform the item and a score of 1 or 2 meant complete ability to perform it. The score was 16 for the balance items and 12 for the gait items, and the total score was 28.

Statistical Analysis

Data analysis of the experimental data was made possible by means of the Statistical Package of Social Science (SPSS Inc., Chicago, IL) for Windows version 11.5. Student T test was used to test the statistical significance of mean differences between fallers and non-fallers with regards to weight, height, age, number of medicines per day, number of the diseases, total score of the Tinetti balance scale as well as the mean score of the total score of the Berg balance scale. Chi square test was used to find out if there's any difference in sex and use of assistive device between the two groups.

A backward Wald logistic regression was used in order to determine the best predictive model for falling in the elderly population. For this particular model, fall history was coded as 0 for non-fallers, 1 for fallers and was used as the dependent variable. A P-value of <0.05 was considered statistically significant.

Then, a cut-off score, the sensitivity and specificity of the best predictor for predicting fall during the year following its administration were calculated. A receiver operating characteristics (ROC) curve was constructed with these results. Finally, to make the best clinical decision, validity indices of the best predictor method, including positive and negative likelihood ratio were determined.

Results

Sixteen participants were lost at the end of follow up. The discriminative value of the two assessment methods (Tinetti Scale and Berg Balance Scale) was studied in 54 elderly divided to fallers and non-fallers.

There were no significant differences in sex, weight, height, number of the medicines and number of the diseases between fallers and non-fallers [Table1]. Table1 shows that the fallers had significantly lower Tinetti Scale and Berg Balance scores, higher age and more use of assistive devise compared with non-fallers.

Table 1: Association of demographic characteristics and medical history with fall classification

Risk factor Non fallers
(n=37)		 Fallers
(n=17)		 P value
Age factor
X(mean)
SD
73.32
7.72
80.53
8.11		 0.003
Gender (%)
Female
Male
35
65
52.94
7.1		 0.404
Weight (Kg)
X(mean)
SD
58.43
11.48
56.29
11.21		 0.525
Height (Cm)
X(mean)
SD
156.51
8.40
154.06
10.50		 0.261
Assistive device (%) 0 53 0.0001
No. of medications per day X(mean)
SD
3.79
2.17
4.06
3.05		 0.722
No. of diseases
X(mean)
SD
1.27
0.8
1.50
1.21		 0.437
Berg Balance Scale
X(mean)
SD
51.46
4.25
39.29
5.07		 0.0001
Tinetti Scale
X(mean)
SD
25.62
2.96
18.29
3.64		 0.0001

A backward Wald logistic regression was performed in order to produce a model with the underlying purpose of predicting those individuals at risk of falling. In the logistic regression model, the predictors entered into the analysis were age, the total Berg balance score and the total Tinetti score. The logistic regression analysis showed that only the total Berg balance score contributed significantly to the prediction of falls (P<0.05). Table 2 displays the regression coefficients (B), the standard errors, as well as the intercept (constant), for the significant predictor variables.

Table 2: Forward logistic regression of fall predictor variables on fall status in a geriatric population

Predictor variables B S.E P. value
Age 1.15 0.95 0.14
Berg Balance Scale 0.48 0.36 0.04
Tinetti Scale 1.44 0.34 0.28

The result from the Berg Balance scale was further assessed in the hope of producing significant cut-off scores that will successfully classify those at risk of falling.
With regards to the Berg balance scale, a cut-off of 46 and above was determined. With this value, sensitivity, those with a history of fall (sensitivity) and specificity values, those who do not present a history of falling, are 94 and 86.5%, respectively.

The validity indices including the sensitivity, specificity, predictive value, and likelihood ratio of Berg Balance Scale are displayed in Table 3 for each cut-off point.

Table 3: Validity indices of Berg Balance Scale according to different cut-off points

Cut off point 43 44 46 48
Sensitivity (%) 76.5
[76.30-76.70]		 88.2
[88.05-88.35]		 94.1
[93.99-94.12]		 94.1
[93.99-94.12]
Specificity (%) 89.2
[89.10-89.30]		 89.2
[89.10-89.30]		 86.5
[86.39-86.61]		 83.8
[83.68-83.92]
Likelihood ratio for a positive test 7.08 8.16 6.97 5.81
Likelihood ratio for a negative test 0.26 0.13 0.07 0.07

 

Discussion

There is no consistent evidence of effective interventions to prevent falls among hospital inpatients [25]. The purpose of the present study was to evaluate the several risk factors linked to the falling and to develop a model that can be used by the vast majority of health care professionals, with the hope of quantifying fall risk among elderly people and preventing this terrible occurrence in the aged.

This study also indicated that the frequency of use of assistive devices and mean age are more in the elderly individuals with history of falling than those without such history. Therefore, the risk of falling for an individual can be estimated according to use or nonuse of assistive devices and his or her age.

This study demonstrates that the fallers had significantly lower Tinetti Scale and Berg Balance scores than non-fallers but Berg Balance Scale assessment method is of more discriminative validity in differentiating those elderly people with history of fall from those without such history. This finding contradicts the findings of Riddle et al [21] and Thorbahn et al [10], and favors the findings of Shumway-Cook et al [22]. Among the causes of difference between the findings of this study and of other studies is the difference in properties of the subjects and their life environments.

Through further investigation of the Berg balance scale, we proceeded to the determination of a cut-off score that would be statistically successful in predicting elderly fallers from non-fallers. Previous studies have determined an inconsistent cut off point. Lajoie et al found that a cut-off score of 46 was statistically effective in predicting falls in the elderly community [8] and a cut-off of 49 was determined with a population of 44 community-dwelling older adults by Shumway-Cook et al [22]. In the present study, we found that a cut-off score of 46 was statistically effective in predicting falls in the elderly population and the sensitivity and specificity of classifying fallers and non-fallers with this cut-off of score is 94 and 86.5%, respectively.

Since therapists should make their clinical decisions according to the results of diagnostic tests and not those tests based on gold standard measures, some researchers believe that positive and negative predictive values are more beneficial than sensitivity and specificity.
Unfortunately, predictive values do not estimate the risk of falling according to the patient's signs and symptoms, and are affected by prevalence. If the prevalence of falls in the study is different from the whole society to which the individual belongs, the predictive values calculated in the study do not make an accurate estimation of the risk of falls for the individual.

Two other validity indices which should be used in clinical decision-making are positive and negative likelihood ratios. The results of likelihood ratios indicate how much the test results change the risk of falling estimated before performance of pretest probability test.
Since likelihood ratios can also be used in tests with continued formats for measuring the distances, Riddle and Stratford [21] believe that they are more beneficial than sensitivity, specificity and predictive values, which are confined to tests with double formats.

According to the results of the study, the risk of falling in an elderly individual with a Berg balance score of 44 and less is 8.16 times (positive likelihood ratio) greater than the probability of not falling, and in an elderly individual with a score more than 44, the negative likelihood ratio is 0.13, i.e. the risk of falling in a patient with a negative Berg Balanced Scale score (equal to or more than 44) is 0.13 times greater than the probability of not falling; in other words, positive likelihood ratio of 8.16 in a patient with a score equal to or less than 44 increases the pretest probability, and negative likelihood ratio of 0.13 in a patient with a score more than 44 decreases the pretest probability. Since determining the prevalence does not change the likelihood ratios, they can be generalized to other patients too.

This study was performed in a small number of elderly populations. The application of this study to individuals living in another setting would be speculative. However, further prospective studies are warranted to confirm these results with a large sample size of the elderly representing various lifestyles.

References
1. Newton RA. Balance abilities in individuals with moderate and severe traumatic brain injury. Brain Inj 1995, 9(5):445-51.
2. Baker SP, O'Neill B, Karpf RS. The Injury Fact Book 1984. Lexington Books, MA, USA.
3. Osteo.org [database online], National Institutes of Health Osteoporosis and Related Bone Diseases - National Resource Center, Washington, DC, 1997.
4. Gu MJ, Schultz AB, Shepard NT, Alexander NB. Postural control in young and elderly adults when stance is perturbed: dynamics. J Biomech 1996, 29(3):319-29.
5. Zwick D, Rochelle A, Choksi A, Domowicz J. Evaluation and treatment of balance in the elderly: A review of the efficacy of the Berg Balance Test and Tai Chi Quan. NeuroRehabilitation 2000,15(1):49-56.
6. Greubel DL, Stokesberry C, Jelley MJ. Preventing costly falls in long-term care.Nurse Pract 2002, 27(3):83-5.
7. Myers H. Hospital fall risk assessment tools: a critique of the literature. Int J Nurs Pract 2003, 9(4):223-35.
8. Lajoie Y, Gallagher SP. Predicting falls within the elderly community: comparison of postural sway, reaction time, the Berg balance scale and the Activities-specific Balance Confidence (ABC) scale for comparing fallers and non-fallers. Arch Gerontol Geriatr 2004, 38(1):11-26.
9. Hauer K, Rost B, Rutschle K, Opitz H, Specht N, Bartsch P, Oster P, Schlierf G. Exercise training for rehabilitation and secondary prevention of falls in geriatric patients with a history of injurious falls. J Am Geriatr Soc 200, 49(1):10-20.
10. Bogle Thorbahn LD, Newton RA. Use of the Berg Balance Test to predict falls in elderly persons. Phys Ther 1996, 76(6):576-83.
11. Okuzumi H, Tanaka A, Haishi K, Meguro KI, Yamazaki H, Nakamura T. Age-related changes in postural control and locomotion. Percept Mot Skills 1995, 81: 991-4.
12. Shumway-Cook A, Gruber W, Baldwin M, Liao S. The effect of multidimensional exercises on balance, mobility, and fall risk in community-dwelling older adults. Phys Ther 1997, 77(1):46-57.
13. Berg KO, Maki BE, Williams JI, Holliday PJ, Wood-Dauphinee SL. Clinical and laboratory measures of postural balance in an elderly population. Arch Phys Med Rehabil 1992, 73(11):1073-80.
14. Lin MR, Hwang HF, Hu MH, Wu HD, Wang YW, Huang FC. Psychometric comparisons of the timed up and go, one-leg stand, functional reach, and Tinetti balance measures in community-dwelling older people. J Am Geriatr Soc 2004, 52(8):1343-8.
15. 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):142-8.
16. Vellas BJ, Wayne SJ, Romero L, Baumgartner RN, Rubenstein LZ, Garry PJ. One-leg balance is an important predictor of injurious falls in older persons. J Am Geriatr Soc 1997, 45(6):735-8.
17. Powell LE, Myers AM. The Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci 1995, 50A(1):M28-34.
18. Tinetti ME. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986,34(2):119-26.
19. Berg KO, Wood-Dauphinee SL, Williams JI, Maki B. Measuring balance in the elderly: validation of an instrument. Can J Public Health 1992, 83 Suppl 2:S7-11.
20. Raiche M, Hebert R, Prince F, Corriveau H. Screening older adults at risk of falling with the Tinetti balance scale. Lancet 2000, 356(9234):1001-2.
21. Riddle DL, Stratford PW. Interpreting validity indexes for diagnostic tests: an illustration using the Berg balance test. Phys Ther 1999, 79(10):939-48.
22. Shumway-Cook A, Baldwin M, Polissar NL, Gruber W. Predicting the probability for falls in community-dwelling older adults. Phys Ther 1997, 77(8):812-9.
23. Harada N, Chiu V, Damron-Rodriguez J, Fowler E, Siu A, Reuben DB. Screening for balance and mobility impairment in elderly individuals living in residential care facilities. Phys Ther 1995, 75(6):462-9.
24. Kellogg International Work Group on the Prevention of Falls by the Elderly. The prevention of falls in later life. Dan Med Bull 1987, 34 (Suppl 4): 1-23.
25. Oliver D, Daly F, Martin FC, McMurdo ME. Risk factors and risk assessment tools for falls in hospital in-patients: a systematic review. Age Ageing 2004, 33(2):122-30.