Screening for balance and mobility impairment in elderly individuals living in residential care facilities.

[Harada N, Chiu V, Damron-Rodriquez J, et al. Screening for balance and mobility impairment in elderly individuals living in residential care facilities. Phys Ther. 1995;75:462-469]

Key Words: Balance, Disability, Elderly, Gait, Screening.

The number of persons over the age of 65 years has increased since the turn of the century, with the most dramatic increase occurring in the number of persons 85 years of age and over. As the number of elderly persons has grown, there has been a corresponding rise in the number of older persons with disability. Based on data from the 1987 National Medical Expenditure Survey, an estimated 9.5 million noninstitutionalized individuals experience difficulty in the performance of basic life activities such as walking, self-care, and home management activities.[1] Out of this total of 9.5 million people, approximately 5.6 million individuals (59%) are over the age of 65 years.(1) The likelihood of having difficulty in carrying out basic life activities increases as an individual ages. In the 65- to 74-year-old age group, 1 in 9 individuals has difficulty performing basic activities.[1] This ratio rises to 1 in 4 individuals in the 75- to 84-year-old age group and to 3 in 5 individuals aged 85 years of age and over.[1] As the number of individuals with disability rises, there will be a subsequent rise in the demand for rehabilitation services to assist these individuals in maintaining the highest functional level possible.

Despite this increase in demand for services, many elderly individuals may not receive needed care because of inconsistent referral to physical therapists by primary care physicians,[2] the inability of elderly individuals to gain access to physical therapy on a routine basis, and the shortage of physical therapists to meet the growing demand for services by community dwelling elders.[3] To alleviate this growing problem, screening methods can be used in the community to identify elderly individuals who have gross limitations in mobility and who may be in need of referral to a physical therapist for more detailed evaluation and possible intervention.

In epidemiology, screening methods are often used to identify a group of individuals with a higher probability of having disease than the general population.[4] General characteristics of a screening test include cost, convenience, reliability, and safety. The most useful characteristics of a screening test, however, are its sensitivity and specificity. Sensitivity is defined as the chance that a test will be positive when applied to someone known to have the disease or disability under consideration. Specificity is defined as the chance that the test will be negative when applied to someone known to be disease- or disability-free. Higher sensitivity and specificity indicate a better screening test. These measures can therefore be used to determine how well a test performs in screening a group of individuals for a certain disability.[4]

Clinically based methods that have been developed to measure physical function may act as screening tests to identify older individuals with limitations in mobility who may benefit from physical therapy. Clinical assessment methods may be particularly suitable for screening because they can detect specific impairments, have established reliability and validity, and can be administered by medical or nonmedical personnel who are trained in their application.[5] if a clinical assessment method is demonstrated to have high sensitivity and specificity in detecting mobility impairment in the elderly, it could be administered by other medical or trained nonmedical personnel to determine whether more detailed evaluation by a physical therapist is warranted.[6]

Several clinical assessment methods have been developed to assess mobility and balance function in the elderly. These methods measure different domains of function such as physical performance on specific tasks,[7-12] gait mechanics,[13,14] or the patient's subjective perceptions of his or her ability to balance.[15,16] The validity of these methods has been tested by (1) determining the correlation between a patient's performance and biomechanical measures (such as measures of sway as determined by force plates)[17,18]; (2) determing the assessment method's ability to predict an event, such as whether a patient will fall (predictive validity)[13]; and (3) determing the correlation of these intruments with other established measures of balance or mobility (concurrent validity).[9] Furthermore, many of these instruments have demonstrated test-retest and interrater reliabilities on groups of elderly persons or patients with specific medical conditions.

The aim of our study was to test the ability of four clinical assessment methods to act as screening tests for detecting elderly individuals with balance and mobility impairments who should be referred for a detailed physical therapy evaluation and possible intervention. These four assessment methods were selected to correspond to the three domains of mobility: Two of the selected instruments measured functional balance, one instrument measured gait speed, and one instrument measured subjective fear of falling. The usefulness of each of these assessment methods as a screening test for referral to a physical therapist for detailed evaluation and possible intervention was determined by calculating sensitivity and specificity using a physical therapist's brief evaluation of each individual as the standard.

Method

Subjects

A convenience sample of elderly subjects was obtained from two licensed residential care facilities located in the Los Angeles, Calif, area. Prior to initiation of the study, the facility administrator contacted conservators and informed family members of the study. in addition, the primary care physician of each potential subject was contacted to determine whether there were any medical problems that would preclude participation in the study.

Out of a total of 109 residents in two residential care facilities, 53 (40%) consented to participate in the study. Subject characteristics are presented in Table 1. The mean age of the subjects was 83.3 years. The majority (87%) of the subjects were female, with an average length of stay in the facility of 2.3 years. Half of the sample required the use of assistive devices for ambulation. There were an average of 2.2 diagnoses per patient, with the top five diagnostic categories being cardiovascular, neurologic, psychiatric, musculoskeletal, and endocrine. Specific diagnoses included hypertension, dementia, depression, stroke, arthritis, and chronic obstructive pulmonary disease. The mean score on the Folstein Mini Mental State Examination was 21.2, indicating mild cognitive impairment.

Table 1. Sample Characteristics(a)
(N= 53)
Variable
Age (y)
 [bar X]                     83.3
SD                            7.7
Range                          62-96
Gender (%)
 Male                         13  (7)
 Female                       87 (46)
Length of stay (y)
 [bar X                       2.3
 SD                           2.2
 Range                         0-9
Diagnosis (%)
 Cardiovascular               36 (41)
 Neurologic                   25 (28)
 Psychiatric                  12 (14)
 Musculoskeletal              12 (13)
 Endocrine                      8 (9)
 Respiratory                    4 (5)
 Digestive                      2 (2)
 Ophthalmologic                  1 (1)
Folstein Mini Mental State
 Examination score
 [bar X]                      21.2
 SD                            5.8
 Range                           5-29
Grip strength (kg)
 Right
 [bar X]                      14.3
 SD                            5.6
 Left
 [bar X]                      13.1
 SD                            5.5
ADL[b] (%)
 Medication                    70 (37)
 Bathing                       21 (11)
 Dressing                       4  (2)
 Feeding                        2  (1)
 Toileting                      0  (0)
Walking aids (%)
 No aid                        51 (27)
 Canes                         26 (14)
 Walkers                       23 (12)
(a) Number of subjects shown in parentheses.
(b) ADL=activities of daily living; values represent
percentages and numbers of subjects
needing assistance.

Clinical Measures

The clinical measures that were tested for their feasibility as screening tests included the following: (1) Berg Balance Scale,[7,18-20] (2) balance subscale of the Tinetti Performance-Oriented Mobility Assessment (POMA),[8,21-23] (3) gait speed,[14]and (4) Tinetti Fall Efficacy Scale.[15,16,24] The characteristics of each of these measures are presented in Table 2.

[TABULAR DATA 2 OMITTED]

The Berg Balance Scale, which measures "functional balance," has three dimensions: maintenance of a position, postural adjustment to voluntary movements, and reaction to external disturbances.[7,18-20] Subject performance on each of 14 activities is measured on a five-point ordinal scale ranging from 0 to 4 (0=unable to perform, 4=independent) so that the aggregate score runges from 0 to 56. Correlations between the Berg Balance Scale and other measures of balance have been determined to be moderate to high. The correlations between the Berg Balance Scale and laboratory tests of postural sway, Tinetti balance subscale, Barthel mobility subscale, and timed "up and go" tests are -.55, .91, .67, and -.76, respectively.[18] Berg et al[20] have found high interrater and intrarater reliabilities (interrater and intrarater reliability intraclass correlation coefficients=.98 and .99, respectively) and high internal consistency (Cronbach's alpha=.96). The average time to administer the scale in these studies was 10 to 15 minutes.

Tinetti's POMA balance subscale measures an individual's position changes and ability to balance while performing certain activities, and is usually used in conjunction with a gait subscale to derive an aggregate score of gait and balance.[8,21] The total score on the POMA balance subscale can range from 0 to 16, with a higher score indicating better balance. Tinetti has reported both interrater and test-retest reliability of .95 for the aggregate score on the gait and balance subscales.[24] The POMA gait and balance subscales have been shown to be highly predictive of falls and fall-related injuries in community-dwelling elderly individuals and residents of intermediate care facilities.[22,23,25] In addition, the POMA gait and balance subscales have been shown to be predictive of nursing home placement and mortality.[26] The mean time to administer the gait and balance subscales is 15 minutes.

Gait speed was measured by an insole footswitch system.(*) This system measures gait characteristics such as speed, cadence, stride length, swing and stance times, single-limb support, and double-limb support. As the subject walks a specified distance, footswitches record foot-floor contact, and these gait characteristics are timed and automatically calculated.[14] Speed was the only gait characteristic to be considered as a screening tool because decreased gait speed has been associated with falling in elderly individuals.[27]

Tinetti's Fall Efficacy Scale measures the degree of confidence an individual has in performing certain activities. [17,18,24] Tinetti suggests that self-efficacy may be an appropriate model to investigate an individual's fear of falling. The total score on the Fall Efficacy Scale can range from 10 to 40, with lower scores indicating greater confidence in walking. Test-retest reliability of this scale in a sample of community-dwelling elderly individuals was found to be .71 (Pearson's correlation).[16,28] The validity of the instrument is suggested by the finding that total scores increase progressively as subjects report an increase in fear of falling.[28]

Data Collection

A "health fair" day was scheduled at each facility to collect baseline data. The "health fair" consisted of three stations at which residents were interviewed and assessed on performance-based measures of gait and balance. Interested residents could attend the health fair at any time throughout the day. As each resident arrived, the study was described, and, if informed consent was obtained, the resident was entered into the study.

All subjects were initially tested for cognitive impairment using the Folstein Mini Mental State Examination.[29] Those individuals who achieved a score of 20 or higher were further interviewed about their walking abilities and administered the Fall Efficacy Scale to assess their confidence in walking. A cutoff score of 20 was selected to allow individuals who were moderately cognitively impaired to be evaluated by the physical therapist.

Following the interview, each subject visited two measurement stations to be evaluated on gait and balance/functional activities. The order of testing was variable. Three physical therapists assessed balance and function by administering the Berg Balance Scale and the Tinetti balance subscale and by evaluating the subject's ability to transfer and walk on level surfaces and ramps. These therapists received prior training to standardize their administration of these tests. Gait speed was assessed by another physical therapist and a research assistant.

Footswitches were inserted into the subject's shoes, and a recorder was strapped onto the subject's waist. Gait characteristics were recorded as the subject walked a distance of 6.1 m (20 ft). The average time to prepare a subject for testing and to test the subject on all activities was 45 minutes.

Two weeks following the initial data collection, another physical therapist, who was blinded to the results of the clinical measures, visited each facility to briefly assess each subject's functional level and the need for further evaluation and possible intervention. The physical therapist's assessment took between 5 to 10 minutes per subject and included an interview followed by observation of the subject's ability to transfer and walk on level surfaces, ramps, stairs, and outdoors. The assessment was tailored to the functional capacity of the subject. For example, subjects who had difficulty walking a short distance indoors were not assessed walking outdoors. Subjects who exhibited difficulty while performing any of these transfer and ambulation activities were judged to be appropriate for physical therapy intervention.

To assess the degree of agreement among physical therapists' judgments of appropriateness for treatment, three of the subjects were videotaped while performing the transfer and ambulation activities. Ten physical therapists, including the physical therapist who performed the assessment on all patients, subsequently rated each subject's appropriateness for physical therapy based on their observation of the videotaped performances. The level of clinical experience of these physical therapists ranged from 0.5 to 47 years, with a mean of 13 years. The physical therapists were asked to elaborate on their evaluation criteria by listing each functional component included in their assessment of subject performance.

Data Analysis

Distributions, frequencies, and measures of central tendency were examined for each clinical measure using the SAS statistical software system.([dagger]) Next, the sensitivity and specificity levels of each clinical measure of balance and mobility were calculated using established methods.[30,31] In general, the number of subjects scoring above and below a specified score (cutoff value) on each measure were counted and categorized according to whether they were true-positive, true-negative, false-positive, or false-negative using the physical therapist's brief assessment as the standard. The selection of cutoff values by which to calculate sensitivity and specificity is arbitrary; therefore, optimal cutoff values were determined by plotting receiver operating characteristic (ROC) curves for each assessment method to determine the point that provided the best tradeoff between sensitivity and specificity.[32] The ROC curve plots sensitivity on the y-axis and specificity (1-specificity) on the x-axis for a range of score values. The point that provides the best tradeoff between sensitivity and specificity is determined by viewing the slope of the ROC curve.[30] In this analysis, the optimal point occurred where the slope of the curve was the closest to 1. Finally, the sensitivity and specificity levels achieved by using two clinical assessment measures, gait speed and the Berg Balance Scale, were determined in a parallel testing situation in which a subject was labeled positive if diagnosed as positive on at least one test.[33] These two measures were selected because they demonstrated the best sensitivity and specificity levels when determined individually.

To assess the degree of physical therapist agreement of appropriateness for treatment based on observation of the videotaped patient performances, the kappa statistic was calculated using STATA statistical software.[double dagger] Kappa is a chance-corrected measure of agreement that can yield values ranging from -- 1 to 1, depending on the strength of agreement.[34] The frequency of items assessed by the physical therapists during their assessment was also examined.

Results

The acceptability of each clinical measure, as determined by the subject's ability to complete the test, was the highest for the evaluation by the physical therapist, followed by both tests of balance, gait speed, and fear of falling. All 53 subjects completed the evaluation by the physical therapist. Forty-three subjects (81%) completed both balance tests, 44 (83%) completed the gait speed test, and 28 (53%) completed the Fall Efficacy Scale. For both tests of balance, the most common reason for noncompletion was subject refusal. The majority of the subjects who failed to complete the gait speed tests did so secondary to decreased cognition. For the Fall Efficacy Scale, almost one half of the subjects could not complete the measure because of cognitive impairment or communication difficulties.

Table 3 provides performance results for each clinical test. Frequency distributions on both the Berg Balance Scale and the Tinetti POMA balance subscale tended to be skewed to the right. In this study, the Berg Balance Scale required 15 minutes to administer, whereas the POMA balance subscale averaged 10 minutes.

[TABULAR DATA 3 OMITTED]

The gait speed measurements were normally distributed, with a mean of 33.6 m/min. The slowest speed was recorded for a subject who had experienced a severe stroke several years earlier. This subject was walking with an ankle-foot orthosis and a quad cane, and had severe left-side weakness.

In the 28 subjects who could be evaluated on the Fall Efficacy Scale, the mean score was 18.3. Out of 53 subjects, 28 (53%) were judged by the physical therapist to be appropriate for inclusion in a physical therapy mobility training program based on the brief assessment described previously.

The ROC curves for the Berg Balance Scale and gait speed are plotted in Figures 1 and 2. The optimal cutoff points, based on ROC curves for each clinical assessment instrument and measure, are presented in Table 4.

[TABULAR DATA 4 OMITTED]

At a cutoff score of 48, the Berg Balance Scale demonstrated an equal specificity level of 78% and a better sensitivity level, 84% versus 68%, than the POMA balance subscale at a cutoff score of 14. The measure of gait speed demonstrated good sensitivity and specificity levels (80% and 89%, respectively) at a cutoff score of 34. The sensitivity and specificity levels of the Fall Efficacy Scale, using a cutoff score of 16 and administered to cognitively intact individuals, were 59% and 82%, respectively. Using a combination of the Berg Balance Scale and gait speed yielded a sensitivity of 91% and a specificity of 70% when a subject was positive on at least one test.

Analysis of the physical therapists' ratings of the videotaped patient performances yielded moderate agreement beyond chance (K=.47, P<.0001). The items assessed by the physical therapists to arrive at their determination of appropriateness for treatment are listed in Table 5. The most frequently reported items assessed on videotape included difficulty with transfers, difficulty with ability to balance, and difficulty with ambulation on level surfaces and stairs.

Table 5. Items Included in a Brief
Physical Therapist Assessment
                                        No. of
                                        Physical
                                        Therapists
                                        Assessing
Item Assessed                           Item(a)
Difficulty with transfers
Difficulty with ambulation on level     9
surfaces
Difficulty with ambulation on stairs    9
Difficulty with balance                 9
Difficulty with ambulation on ramps     8
Assistive device fit/use                7
Posture                                 6
Strength (functional)                   6
Gait deviations                         6
Range of motion                         3
Cognitive impairment(b)                 2
(a) Ten physical therapists participated in the
videotaped patient ratings.
(b) Cognitive impairment could not be directly
assessed on videotape although two physical
therapists reported that they would assess this
in person.

Discussion

The results of this study show that clinical assessment instruments that detect balance and mobility impairments are useful for screening elderly individuals who may be in need of a detailed physical therapy evaluation and possibly intervention. These screening methods demonstrated good sensitivity and specificity for reproducing a physical therapist's judgment. The two clinical measures that showed the strongest potential for use as screening methods were the Berg Balance Scale and gait speed. Our results showed that the Berg Balance Scale was more sensitive than the Tinetti POMA balance subscale and had comparable specificity. The Berg Balance Scale, however, takes longer to administer than the POMA balance subscale (15 minutes versus 10 minutes). The strength of the Berg Balance Scale lies in its detailed grading scale, which appears to be better at detecting balance impairment than the POMA balance subscale. Topper et al[35] also describe this limitation of the POMA in identifying individuals who are at risk for falling.

In developing the screening methods, we included individual measures of balance, gait, and subjective perceptions of fear of falling. The combination of two clinical measures, balance (as measured by the Berg Balance Scale) and gait speed, yielded the highest sensitivity level of 91%, suggesting that a combination of clinical tests most accurately reflects the physical therapist's judgment and thus might be the best for developing screening methods. In a screening situation involving the Berg Balance Scale and gait speed, a subject would be administered the second test only if the first test did not indicate the need for further evaluation.

In this study, physical therapists were used to conduct screening tests to maximize efforts of ensuring that these tests were performed consistently. The use of non-physical therapists to perform these tests could decrease the likelihood that the tests were performed consistently because physical therapists are specially trained to assess function. In a screening situation, these screening tests could be administered by health care personnel who are in constant contact with elderly people, such as primary care physicians during a routine office visit or facility personnel. These individuals, however, would require in-depth training to consistently administer the screening tests. The next step is to have physical therapists, or other individuals who are knowledgeable about these tests, educate other health care personnel on how to conduct the tests for elderly individuals. if physical therapists can teach lower-level personnel to conduct these tests, the screening procedure would be less costly.

This study showed that tests of balance and gait speed had high sensitivity and specificity in screening the older community-dwelling population for balance and mobility impairments. These specific tests were selected because they have established reliability validity in the geriatric literature. Other comparable, yet simpler, tests could be used as screening methods for the detection of balance and mobility deficits. For example, gait speed could be measured by using a stopwatch rather than the footswitch system, and balance could be measured using the forward-reach technique rather than the Berg Balance Scale. The use of simpler tests would facilitate their administration. Simpler methods, however, would have to show acceptable sensitivity and specificity to be able to identify appropriate individuals.

There are several limitations of this study. First, the sample was relatively small and drawn from the residential care facility population. This study is therefore not generalizable to other community-dwelling elderly individuals, such as those living in their own homes. The residential care facility population, however, represents a sizable community-dwelling population with a high prevalence of balance and mobility deficits who often go undetected in their need for physical therapy services. Second, there were only three physical therapists to assess balance using the Berg and Tinetti scales. These therapists, however, received prior training to standardize the administration of these tests. Third, the standard for identifying subjects with balance and mobility deficits was limited to the assessment of one licensed physical therapist, which was brief and may not have been fully diagnostic. Subsequent ratings of videotaped patient performances by 10 physical therapists, including the initial rater, however, yielded moderate agreement beyond chance, indicating that physical therapists tend to evaluate patients similarly. Finally, there was a 2-week lag time between testing and the physical therapists' evaluations. There could have been some clinical changes between the ratings, although such changes should be minimal in a stable population.

Further studies should be performed that incorporate larger sample sizes and different sites in developing screening methods to identify older persons with balance and mobility deficits who may be in need of more detailed physical therapy evaluation by a skilled therapist. To facilitate the administration of screening tests by clinicians or facility personnel, further studies should be performed to identify simpler and easier-to-administer methods with comparable sensitivity and specificity in the community-based setting.

Conclusion

As the US population over the age of 65 years continues to grow, there will be a corresponding rise in the level of functional disability. Physical therapists can play an important role in delaying the onset of disability and prolonging health into older ages. It is therefore imperative that appropriate screening methods are developed to identify community-dwelling elderly individuals with functional impairment who should be referred for a detailed physical therapy evaluation.

Acknowledgments

We thank Susie Mais, PT, and the Physical Therapy Department at the Veterans Administration Medical Center-West Los Angeles for their expertise in the evaluation and treatment of elderly individuals. Maria Duarte provided ongoing support in the administration of this project. We also thank Theodore Hahn, MD, and Larry Rubenstein, MD, for their thoughtful comments in the preparation of this manuscript. (*) B&L Engineering, 12309 E Florence Ave, Santa Fe Springs, CA 90670 ([dagger]) SAS Institute, SAS Campus Dr, Cary, NC 27513. ([double dagger]) Computing Resource Center, 1640 Fifth St, Santa Monica, CA 90401.

[Figure 1 to 3 ILLUSTRATION OMITTED]

References

[1] Disability Statistics Program, University of California, San Francisco. Disability Statistics Abstract, Number 3. Washington, DC: US Department of Education, National Institute on Disability and Rehabilitation Research; April 1992. [2] Hoenig H, Mayer-Oaks SA, Siebens H, et al. Geriatric rehabilitation: What do physicians know about it and how should they use it? J Am Geriatr Soc. 1994;42:341-347. [3] Koska MT. Rehabilitation growth fuels PT shortages. Hospitals. 1989;63:32. [4] Rogan WJ, Gladen B. Estimating prevalence from the results of a screening test. Am J Epidemiol. 1978;107:71-76. [5] Applegate WB, Blass JP, Williams TF. Instruments for the functional assessment of older patients. N Engl J Med. 1990;322:1207-1214. [6] Neufeld Bloom S. The frail and institutionalized elderly. In: Guccione A, ed. Geriatric Physical Therapy. St Louis, Mo: Mosby-Year Book; 1993:377-390. [7] Berg K, Wood-Dauphinee S, Willialm JI, Maki B. Measuring balance in the elderly: validation of an instrument. Can J Public Health. 1992;83:S7-S11. [8] Tinetti ME. Performance-oriented assessment of mobility problems in the elderly. Am Geriatr Soc. 1986;34:119-126, [9] Weiner DK, Duncan PW, Chandler J, Studenski SA. Functional reach: a marker of physical frailty. JAM Geriatr Soc. 1992;40:203-207. [10] Ring C, Nayak USL, Isaacs B. Balance function in elderly people who have and who have not fallen. Arch Phys Med Rehabil. 1988; 69:261-264. [11] Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the "Get-up and Go" test. Arch Phys Med Rehabil. 1986;67:387-389. [12] Studenski S, Duncan PW, Chandler J. Postural responses and effector factors in persons with unexplained falls: results and methodologic issues. J Am Geriatr Soc. 1991,39:229-234. [13] Feltner ME, MacRae PG, McNitt-Gray JL. Quantitative gait assessment as a predictor of prospective and retrospective falls in community-dwelling older women. Arch Phys Med Rehabil. 1994,75:447-453. [14] Perry J. Gait Analysis: Normal and Pathoclogical Function. Thorofare, NJ: SLACK Inc; 1992:431. [15] Tinetti ME, Richman D, Powell L. Falls efficacy as a measure of fear of falling. J Gerontol. 1990;45:P239-P243. [16] Tinetti ME, Powell L. Fear of falling and low self-efficacy: a cause of dependence in elderly persons. J Gerontol. 1993;48:35-38. [17] Lichtenstein MJ, Burger MC, Shields SL, Shiavi RG. Comparison of biomechanics platform measures of balance and videotaped measures of gait with a clinical mobility scale in elderly women. J Gerontol. 1990;45:M49-M54. [18] Berg K, Maki B, Willialm JI, et al. Clinical and laboratory measures of postural balance in an elderly population. Arch Phys Med Rehabil. 1992;73:1073-1080. [19] Berg K. Balance and its measure in the elderly; a review. Physiotherapy Canada. 1989;41:240-246. [20] Berg K, Wood-Dauphinee S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiotherapy Canada. 1989;41:304-311. [21] Tinetti ME, Ginter SF. Identifying mobility dysfunction in the elderly. JAMA. 1988;259: 1190-1193. [22] Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med. 1988;319: 1701-1707. [23] Tinetti ME, Williams TF, Mayewski R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med. 1986;80: 429-434. [24] Tinetti ME, Baker DI, Garrett PA, et al. Yale FICSIT: risk factor abatement strategy for fall prevention. J Am Geriatr Soc. 1993;41: 315-320. [25] Robbins AS, Rubenstein LZ, Josephson KR, et al. Predictors of falls among elderly people: results of two population-based studies. Arch Int Med. 1989;149:1628-1633. [26] Reuben DB, Siu A, Kimpau S. The predictive validity of self-report and performance-based measures of function and health. J Gerontol. 1992;47:M106-M110. [27] Guimaraes RM, Issacs B. Characteristics of the gait in old people who fall. Int Rehab Med. 1980;2:177-180. [28] Buchner DM, Hombrook MC, Kutner NG, et al. Development of the common data base for the FICSIT trials. J Am Geriatr Soc. 1993; 41:297-308. [29] Folstein MF, Folstein SE, McHugh PR. "Mini-mental state": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 1 2:189-198. [30] Fletcher RH, Fletcher SW, Wagner EH. Clinical Epidemiology: The Essentials. Baltimore, Md: Williams and Wilkins; 1982:41-58. [31] Feinstein AR. On the sensitivity, specificity, and discrimination of diagnostic tests. Clin Pharmacol Ther. 1975;17:104-116. [32] Swets JA, Pickett RM. Evaluation of Diagnostic Systems: Methods From Signal Detection Theory. New York, NY: Academic Press Inc; 1982. [33] Mausner JS, Bahn AK. Epdemiology. Philadelphia, Pa: WB Saunders Co; 1974:247-248. [34] Woolson RF. Statistical Methods for the Analysis of Biomedical Data. New York, NY: John Wiley & Sons Inc; 1987:255-260. [35] Topper AK, Maki BE, Holliday PJ. Are activity-based assessments of balance and gait in the elderly predictive of risk of falling and/or type of fall? J Am Geriatr Soc. 1993;41: 479-487.

N Harada, Phd, PT, is Health Services Research Associate, Geriatric Research, Education, and Clinical Center (11G), Veterans Administration Medical Center-West Los Angeles, and Assistant Professor, School of Medicine, University of California, Los Angeles, Los Angeles, CA 90073 (USA). She is also Associate Director, UCLA/MEDTEP Outcomes Research Center for Asians and Pacific Islanders. Address all correspondence to Dr Harada.

V Chiu, is Research Associate, UCLA/MEDTEP Outcomes Research Center for Asians and Pacific Islanders.

J Damron-Rodriguez, PhD, is Associate Chief of Evaluation and Education, Geriatric Research, Education, and Clinical Center, Veterans Administration Medical Center-West Los Angeles, and Assistant Professor, School of Public Policy and Social Research, University of California, Los Angeles.

E Fowler, PhD, PT, is Adjunct Associate Professor, School of Medicine, and Director, Functional Assessment Laboratory, Department of Rehabilitation Services, University of California, Los Angeles.

A Siu, MD, MSPH, is Deputy Commissioner for Clinical Affairs, New York State Department of Health, Empire State Plaza, Coming Tower, Albany, NY 12220.

DB Reuben, is Chief, Division of Geriatrics; Director, Multicampus Program in Geriatric Medicine and Gerontology; and Associate Professor, School of Medicine, University of California, Los Angeles.

This study was approved by the Human Subjects Protection Committee of the School of Medicine, University of California, Los Angeles, and the Veterans Administration Medical Center-West Los Angeles.

This study was supported by the UCLA Older Americans Independence Center, Grant 5 P60 AG10415-02, and the UCLA/MEDTEP Outcomes Research Center for Asians and Pacific Islanders, Grant HS07370.

This article was submitted July 6, 1994, and was accepted January 4, 1995.