- Email: [email protected]
The Relationship Between Falls History and Computerized Dynamic Posturography in Persons With Balance and Vestibular Disorders
402
ORIGINAL ARTICLE
The Relationship Between Falls History and Computerized Dynamic Posturography in Persons With Balance and Vestibular Disorders Susan L. Whitney, PhD, PT, NCS, ATC, Gregory F. Marchetti, PhD, PT, Annika I. Schade, MD ABSTRACT. Whitney SL, Marchetti GF, Schade AI. The relationship between falls history and computerized dynamic posturography in persons with balance and vestibular disorders. Arch Phys Med Rehabil 2006;87:402-7.
© 2006 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
Objective: To describe the relationship between Sensory Organization Test (SOT) scores and reported falls in persons with vestibular and balance disorders. Design: Descriptive cross-sectional. Setting: Outpatient tertiary balance and falls clinic. Participants: One hundred physical therapy (PT) charts of people referred to a balance and falls clinic were reviewed. Criteria for inclusion were that the patients had completed the SOT of computerized dynamic posturography (CDP), had a vestibular diagnosis, and had the numbers of falls recorded from patient report within the last 6 months at the initial examination. Rotational chair, caloric testing, oculomotor test, and Activities-Specific Balance Confidence (ABC) results were recorded, if available. Interventions: Not applicable. Main Outcome Measures: The SOT composite and ABC scores as they related to reports of falls in the last 6 months. Results: The overall proportion of persons with vestibular disorders with positive 6-month history for 1 or more falls was 30% and for recurrent falls (ⱖ2) was 17%. Vestibular laboratory findings, age, sex, and vestibular diagnosis were not related to reported falls status. Patients who reported multiple falls prior to the PT examination had lower SOT composite scores than patients who reported 1 or no falls in the previous 6 months. The receiver operating characteristic curve identified a composite SOT score of less than 38 as demonstrating the highest likelihood ratio for differentiating between those people who reported no falls in the past 6 months and those who reported 2 or more falls. Conclusions: Persons who are recurrent fallers perform worse on SOT than either nonfallers or 1-time fallers. CDP performance can help guide the clinician in the development of a safe exercise program. Key Words: Accidental falls; Balance; Electronystagmography; Posture; Rehabilitation; Vestibular diseases.
HE RELATIONSHIP BETWEEN FALLS and vestibular/ T balance dysfunction is not well understood, yet recent reports suggest that falling in persons with vestibular disorders
From the Departments of Physical Therapy (Whitney) and Otolaryngology (Whitney), University of Pittsburgh, Pittsburgh, PA; Centers for Rehab Services, University of Pittsburgh Medical Center, Pittsburgh, PA (Whitney); Department of Physical Therapy, Duquesne University, Pittsburgh, PA (Marchetti); and ORL Klinik, Universitätsspital Zürich, Zürich, Switzerland (Schade). Presented in part to the Barany Society, September 29, 2002, Seattle, WA. Supported in part by the National Institutes of Health (grant no. DC 05384). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Correspondence to Susan L. Whitney, PhD, PT, NCS, ATC, University of Pittsburgh, 6035 Forbes Tower, Pittsburgh, PA 15260, e-mail: [email protected]. 0003-9993/06/8703-10348$32.00/0 doi:10.1016/j.apmr.2005.11.002
Arch Phys Med Rehabil Vol 87, March 2006
is a concern. Whitney et al1 reported that 35% of patients with various vestibular disorders under 65 years of age and 39% of those over the age of 65 reported falls during their initial examination. Fifty-one percent of persons with bilateral vestibular disorders between the ages of 65 and 74 reported falls.2 Recently, Pothula et al3 suggested that vestibular dysfunction may be responsible for unexplained falls in persons reporting to an emergency department. Better identification of persons at risk for falling may help the clinician make more targeted intervention decisions. The relationship between vestibular laboratory tests and computerized dynamic posturography (CDP) has been reported by various authors.4,5 Keim4 retrospectively compared the results of CDP and electronystagmography in persons who had peripheral, central, or both peripheral and central vestibular disorders. He reported that 42% had abnormal vestibular findings and 83% had an abnormal CDP score. Keim4 reported a higher prevalence of abnormal CDP results in persons with central vestibular disorders, suggesting that CDP is more sensitive at detecting deficits with persons with brain dysfunction.6 In a meta-analysis performed on CDP by Di Fabio,6 the sensitivity and specificity of CDP was approximately 50% when compared with other tests of vestibular function, yet Di Fabio also reports that CDP can detect abnormalities in some patients who have normal vestibulo-ocular functioning. Others4,7-9 have all suggested that CDP might be used to gain insight into the patient’s balance dysfunction that may be caused by biomechanic or neurologic disorders. Jacobson et al10 have suggested that only testing the vestibulo-ocular reflex may not provide a comprehensive evaluation of the patient’s functional deficits. Girardi et al5 reported on the relationship between selfreported falls in persons presenting to a balance and falls clinic over the age of 65 who had fallen at least 1 time in the past year (mean fall rate, 3.5 times) and scores from the Balance Trak 500, which included 4 of the 6 tests of CDP. CDP was more sensitive at identifying people who had fallen compared with ocular motor, rotational chair, positional, or caloric findings.5 Others have tested older adults using platform posturography devices that include 4 of the 6 conditions of CDP.11,12 Rose and Clark11 reported significant Sensory Organization Test (SOT) equilibrium score changes after intervention in older adults with a history of falls. The SOT mean composite score of 4 of the original 6 conditions of CDP has been shown to differentiate between fallers and nonfallers in older persons. In contrast, Wolfson et al13 demonstrated that there was no relationship between performance on the SOT and reported falls within
403
POSTUROGRAPHY AND FALLS, Whitney Table 1: Patient Demographics and Characteristics by Patient Reported Fall Status Patients Classified by Reported Fall Status Variable
All Patients
Nonfaller
1-Time Faller
Recurrent Faller
n (%) Mean age ⫾ SD (range) (n⫽100) Vestibular diagnostic category, n (%) Peripheral Central Combined (peripheral and central) No. aged ⱖ65y (%) (n⫽100) No. of women (%) (n⫽62) Mean months of symptom duration ⫾ SD (range) (n⫽99) Mean ABC score ⫾ SD (range) (n⫽91) Mean composite SOT score ⫾ SD (range) (n⫽100)
100 (100) 59⫾17 (14–90)
70 (70) 58⫾15 (14–86)
13 (13) 63⫾21 (27–90)
17 (17) 60⫾19 (22–89)
52 (52) 42 (42) 6 (6) 43 (43) 62 (62) 36⫾60 (1–360) 60⫾24 (0–100) 52⫾18 (6–85)
36 (52) 31 (44) 3 (4) 28 (40) 44 (63) 37⫾67 (1–360) 64⫾23 (0–100) 55⫾17 (6–85)
9 (69) 3 (23) 1 (8) 7 (53) 8 (62) 30⫾38 (1–96) 44⫾24 (12–96) 52⫾18 (16–77)
7 (41) 8 (47) 2 (12) 8 (47) 10 (59) 36⫾39 (1–144) 51⫾19 (20–86) 39⫾18 (7–82)
Abbreviations: ABC, Activities-Specific Balance Confidence scale; SD, standard deviation.
a 12-month period in nursing home residents, although the SOT total number of losses of balance on conditions 1 through 6 were related to ankle strength and age. Black et al14 have reported improvements in SOT scores and fewer falls during CDP testing over the course of rehabilitation in persons with peripheral vestibular disorders. Others have reported changes in SOT scores over the course of vestibular rehabilitation as an outcome measure, which have correlated with changes in impairment and various functional performance measures.15-17 O’Neill et al18 have also reported that SOT scores improved over the course of rehabilitation, yet there was not a relationship between improvements demonstrated on SOT and changes in functional performance in persons with peripheral vestibular disorders. No studies to date suggest a relationship between CDP that includes all 6 tests of the SOT and reported falls in people with vestibular disorders. In the cohort of Girardi et al,5 all subjects had fallen at least 1 time in the last year and Girardi only included 4 of the 6 SOT conditions. Therefore, the purpose of this project was to determine if there is a relationship between CDP and reported falls in persons with balance and vestibular disorders. METHODS We reviewed 100 charts of patients with vestibular disorders who had been referred to physical therapy (PT). All patients who had undergone CDP during their first PT session and who had information about reported falls in the previous 6 months were included in the study. The University of Pittsburgh Institutional Review Board approved this study as exempt. Seventy percent of the patients during the data collection period met the above 3 criteria for inclusion. Patients were not tested on CDP for the following reasons: they were unable to stand for 10 to 15 minutes unsupported, it was not possible to explain the test
due to mental status in order to obtain legitimate consent, they arrived late for the PT evaluation, they refused, or it was felt that CDP would make the patient ill because of severe motion intolerance based on the physical therapist’s judgment. The first author administered all of the CDP testing as part of the standard PT assessment and had over 15 years experience with the CDP equipment. The first author, who was also the treating physical therapist, asked all of the patients if they had fallen in the last 6 months in a systematic manner that was recorded within the patient record. None of the falls reported suggested that the patient was experiencing a Tumarkin fall from an otolith crisis.19,20 There were 70 patients who did not report any falls, 13 reported 1 fall, and 17 persons who reported more than 1 fall. Patient demographics by fall status are detailed in table 1. Between 74% and 85% of the patients underwent portions of electronystagmography testing that included rotational chair testing, positional testing, and caloric testing (table 2). The physician prescribing the laboratory tests specifically ordered the number of tests needed to make a diagnosis based on the patients presenting symptoms. The protocol of the laboratory testing has previously been reported with the criteria for what was considered a normal laboratory test.21 Based on the published criteria, an otolaryngologist determined if the individual vestibular laboratory test was normal or abnormal. A trained physical therapist performed the SOT of CDP on the EquiTest equipment. The physical therapists followed the standard EquiTest protocol for testing (version 4.0). Test-retest reliability (intraclass correlation coefficient) of the SOT has been reported to vary between .15 and .93.22,23 Percentage agreement for losses of balance during testing ranged from 77% to 100%.22 An in-depth description of the SOT protocol has been provided by Ford-Smith et al.22 The platform consists of 2 forceplates that move together and rotate when necessary,
Table 2: The Electronystagmographic Results Categorized by Reported Falls Prior to the Start of PT for Persons Who Had Completed the Full or Partial Electronystagmography Battery Fall Status
Nonfaller
1-Time Faller
Recurrent Faller
Diagnostic Test
Total
Abnormal, n (%)
Total
Abnormal, n (%)
Total
Abnormal, n (%)
Rotational chair test (n⫽74) Oculomotor test (n⫽82) Caloric test (n⫽75) Positional test (n⫽79)
50 55 52 54
24 (48) 6 (11) 23 (44) 13 (24)
9 12 10 12
3 (33) 1 (8) 5 (50) 2 (17)
15 15 13 13
8 (53) 2 (13) 4 (13) 2 (15)
NOTE. The total subject number reveals how many patients completed each test.
Arch Phys Med Rehabil Vol 87, March 2006
POSTUROGRAPHY AND FALLS, Whitney
while the patient is secured in an overhead harness to prevent a fall while viewing a visual surround. Conditions 1 and 2 include standing on a solid support surface with the eyes open and eyes closed, respectively. Condition 3 consists of having the subject stand on a solid support surface with the visual surround sway referenced (the walls move in the same direction as the subject’s sway). During conditions 4 and 5, the floor is sway referenced (the floor moves in the same direction as the subject’s sway) with eyes open and closed. Condition 6 includes both a sway referenced floor and visual surround. If a person fell on 1 of the 3 trials, it was recorded as a fall for that condition. A fall was defined as a patient taking a step, grabbing the wall or side of the platform, opening their eyes during an eyes closed condition, or when the patient had exceeded their limits of stability.22 If the operator stopped the trial, the trial was scored a 0. In addition, the composite equilibrium score was recorded for analysis that was automatically computed by the standard algorithm in the computer software. The equilibrium scores can range from 0 to 100 with 100 indicating no movement on the posture platform. Higher SOT scores indicate greater stability. Patients also completed the Activities-Specific Balance Confidence (ABC) scale, a questionnaire that is used to assess balance confidence.24 Scores of 67 and lower have recently been related to falling in older adults in the community.25 Fear of falling has been related to future fall events.26 ABC scores were recorded to determine if there was any association between balance confidence and performance on the SOT of CDP. Data Analysis We analyzed the characteristics of the study sample using descriptive statistics for continuous variables (age, symptom duration, ABC score) and frequency analysis for categoric variables (diagnostic category, vestibular test results [rotational, oculomotor, caloric, and positional tests]). We classified the subjects based on their number of self-reported falls in the last 6 months as a nonfaller, a 1-time faller, or a recurrent faller. The differences between fall groups were analyzed using 1-way analysis of variance F test with Bonferroni post hoc multiple comparisons for continuous variables and contingency table analysis with chi-square test for categoric variables. Statistical significance was determined at P less than .05 for all analyses. We analyzed the association between CDP and self-reported fall status groups using a generalized linear model with Bonferroni multiple group comparison to test the effect of fall group status on SOT score plus the effect of age as a covariate. The CDP score that best discriminated self-reported fall status group was determined using a receiver operating characteristic (ROC) curve analysis. In addition, likelihood ratios were calculated based on the optimal sensitivity and specificity of the CDP score for the self-reported fall groups. The ROC and likelihood ratio analyses were based on the self-reported fall group and SOT differences identified in the generalized linear model. CDP performance was dichotomously measured based on fall behavior on each SOT condition. Any loss of balance during each of the 6 SOT conditions (3 trials per condition) was recorded categorically as a fall for that SOT condition. The significance of the association between self-reported fall status groups (nonfaller, 1-time faller, recurrent faller) and falling on 1 or more SOT conditions as defined above was determined using a contingency table with chi-square analysis. Arch Phys Med Rehabil Vol 87, March 2006
RESULTS Table 1 contains demographic data for the total sample plus for all patients as classified by their reports of falling in the 6 months prior to examination. There was no significant difference among patient self-reported fall status groups on the variables of age or symptom duration (P⬎.72). In addition, there was no difference in the proportion of patients’ age 65 and older by self-reported falls status groups. There was no relationship between sex or diagnostic category (peripheral, central, and both peripheral and central) and the nonfaller, 1-time faller, or recurrent faller groups. There was a significant difference in mean ABC score among the 3 self-reported falls status groups (F2⫽5.17, P⬍.01). Multiple group comparisons (Bonferroni) indicated that patients who reported no falls in the past 6 months demonstrated greater balance confidence, as measured by the ABC, than patients reporting 1 fall in the past 6 months (P⬍.02). There was no significant difference in balance confidence (mean ABC score) between the patient group who reported no past falls and the recurrent fall group. There was no significant difference in mean ABC score between the patient group reporting a single fall and the group reporting recurrent falls in the 6 months before examination. Figure 1 illustrates the group means and 95% confidence intervals (CIs) for the ABC score for patients stratified by self-reported fall history in the previous 6 months. Mean SOT scores and self-reported falls history within the past 6 months were significantly related (F3⫽5.81, P⬍.01). Figure 2 illustrates the means and 95% CIs for the SOT score by reported fall history in the past 6 months. Multiple group comparisons (Bonferroni) indicated that the patient group who reported no falls in the previous 6 months demonstrated higher mean SOT scores than the recurrent self-reported falls group (P⬍.01). There was no difference in mean SOT score between the patient group who reported no falls and the group reporting 1 fall in the 6 months prior to examination. The patient group
80.00
95% CI of Mean ABC Score
404
*
70.00
60.00
50.00
40.00
30.00
20.00 Recurrent Faller
One-Time Faller
Nonfaller
Fall History Past 6 Months
Fig 1. Mean ABC scores with 95% CIs for patients stratified by self-reported falls history for 6 months prior to the start of PT (recurrent faller, 1-time faller, nonfaller). *Significant difference (P<.01).
405
POSTUROGRAPHY AND FALLS, Whitney
0.9 60.00
50.00
0.8 Sensitivity (tru e p o sitives)
95% CI of Composite SOT Score
1
*
40.00
0.7 0.6
*
0.5 0.4 0.3 0.2 0.1
30.00
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 - Specificity (false positives) Recurrent Faller
One-Time Faller
Nonfaller
Fall History Past 6 Months Fig 2. The mean composite SOT scores with the 95% CIs for patient stratified by self-reported fall status (recurrent, 1-time, nonfaller). *Significant difference (P<.01).
who reported falling 1 time in the 6 months prior to PT examination demonstrated a higher mean SOT score than the group who reported 2 or more falls. This difference, however, was not statistically significant (P⫽.08). This failure of this difference to reach statistical significance may represent a type II error based on the small sample size in the 1-time and recurrent fall groups. Age was a significant covariate in the relationship between the SOT and the self-reported falls groups (P⫽.02). Patient age was found to account for 5% of the total SOT score variance. Due to this significant covariance, the self-reported falls group differences in mean SOT were determined based on adjustment for age. The ROC curve was constructed to determine if the composite SOT score could be used to discriminate between the nonfalling and recurrent falling self-reported falls groups. The 1-time self-reported fall group was not considered in the ROC curve analysis because of group similarities to both the nonfalling group (similar age-adjusted mean SOT scores) and the recurrent falling group (similar mean ABC scores). Analysis of the ROC curve (fig 3) revealed that optimal identification of people who report recurrent falls was associated with a composite SOT score of less than 38 (sensitivity, 53%; specificity, 87%). The likelihood ratio at the SOT value of less than 38 was 4.12, indicating an over 4-fold increase in the pre- to posttest odds of detecting a person who reported recurrent falls in the 6 months prior to examination (24% pretest odds to 99.8% posttest odds). Among the people who did not report a fall and those people who reported multiple falls, the probability that any individual patient reported multiple falls was 20% (17/87). Applying the likelihood ratio of 4.12 increases that probability to 50% (150% increase over baseline). After examining the relationship between SOT conditions in which 1 or more falls occurred over the 3 trials per condition by nonfaller, 1-time faller, and recurrent faller, the strongest association was seen with falls on SOT conditions 5 and 6
Fig 3. ROC curve (with point of highest likelihood ratio) for the discrimination of patients with self-reported recurrent fall history from patients with no history of self-reported falls using the SOT score. Area under curve equals .75 (P<.01). Diagonal line indicates the area of no discrimination. *Cutoff below 38 points as the location of the maximum likelihood ratio (4.12).
(table 3). There were also significant relationships between falls on SOT conditions 6, and falls on SOT conditions 4, 5, and 6. In interpreting these associations, it is imperative to note that the categoric SOT falling conditions (eg, falling on conditions 5 and 6, condition 6, or conditions 4, 5, and 6) are not mutually exclusive for each patient. For example, a patient identified as having fallen on conditions on 5 and 6 may or may not have also fallen on conditions 2, 3, and 4. Table 2 provides the vestibular test results categorized by patient status (nonfaller, 1-time faller, recurrent faller) based on the number of self-reported falls in the 6 months prior to the PT examination. There was no relationship between diagnostic test results (normal versus abnormal findings of each of the vestibular tests) and self-reported fall group status. DISCUSSION There is a relationship between the composite score of the SOT and reported falls in persons with vestibular disorders. The use of the SOT composite score can help confirm the clinician’s other examination findings about the patient’s pos-
Table 3: Proportion of Patients in Each Self-Reported Fall History Group Who Fell on SOT Conditions 5 and 6, Condition 6, and Conditions 4, 5, and 6* Self-Reported Fall History
Fall on SOT Condition
Recurrent
1-Time
No Falls
P†
5 and 6 (%) 6 (%) 4, 5, and 6 (%)
88 88 47
54 69 31
46 53 17
⬍.01 ⬍.03 ⬍.03
*Persons who fell on conditions 5 and 6, condition 6, and conditions 4, 5, and 6 might have also fallen on other conditions. These 2 categories were the only significant associations with self-reported fall status. † Chi-square test with 2 degrees of freedom.
Arch Phys Med Rehabil Vol 87, March 2006
406
POSTUROGRAPHY AND FALLS, Whitney
tural stability. Although the sensitivity of the SOT cut score at less than 38 was only 53% when comparing nonfallers versus recurrent fallers, the specificity of the cut score was 87%. To date, there has not been a published cut score on the SOT that has been related to past reports of falls. Obviously, prospective work would be ideal to identify future falls. The question of how useful the information gained from the SOT in the prediction of future falls can certainly not be answered from a retrospective study. There is also the potential source of error with recall bias in using patient self-report to determine fall history.27 A prospective study in this setting could not be conducted in the absence of PT intervention to reduce fall risk for these patients. Additionally, even prospective studies of falls rely on patient self report to ascertain the occurrence of a fall event and are not free of recall bias. While the retrospective design limits the predictive value of the results, we feel that descriptive information gained from an analysis of the clinical and fall history associations with CDP does provide a link between balance impairments and disability in patients with vestibular dysfunction. From these data, no link to future fall risk can be made. There was a relationship between observed falling on conditions 2 through 6 on the SOT and self-reported falls, which has previously not been reported in persons with vestibular disorders. Age, sex, and length of symptoms did not affect the patient’s reported fall status. There was no significant association between patient self-reported fall status and age (⬍65 or 65 and older), unlike the work of Cohen28 and Shimada29 and colleagues, which indicated that SOT number of falls were related to age. Cohen28 and Shimada29 included people without balance dysfunction, unlike the present sample, which consisted of people with balance disorders, possibly suggesting that the age effect on SOT scores is no longer a factor because all patients had balance disorders. Typically as one increases in age, there is an increased risk of falling.30 Wolfson et al13 reported an increase in the number of SOT falls with increasing age, although their subjects were nursing home residents with a mean age of 77 or older. The subjects in the present study were younger, had balance and vestibular disorders, and were ambulatory outpatients. In the study by Girardi et al,5 all subjects had to have fallen at least once in the last year and had to be over the age of 65 in order to meet the criteria for admission to their study. Our data suggest that, regardless of age, people with vestibular disorders report falling. Cohen28 noted that no normative subjects fell on conditions 1, 2, or 3 and only 1 person fell on condition 4, indicating that conditions 1 through 4 were fairly easy for Cohen’s sample of healthy community-living older adults. Similar to Cohen,28 none of our subjects fell on condition 1. Four people fell on at least 1 trial on condition 2, 13 people on condition 3, and 33 (33%) persons fell on at least 1 trial of condition 4 in the present study. It was interesting that a similar number of persons with vestibular disorders fell on condition 4 (30% vs 37%), whether or not they were 65 years of age or older. Wolfson13 suggested that older community-dwelling women may be more likely to fall on CDP during sway referenced conditions with visual compromise than men. Sixty-three percent of the nonfallers, 62% of the 1-time fallers, and 59% of the 2 or more fallers were reported by women. Overall, “dizziness” is reported more in women than in men, yet sex did not affect their fall category (non, 1-time, or recurrent faller).31,32 Girardi’s subjects5 had a mean age of 78 and had experienced at least 1 “significant” fall within the last year. The average fall rate of their sample was 3.5 falls in the previous year versus 30% of our subjects who fell in the last 6 months, Arch Phys Med Rehabil Vol 87, March 2006
indicating that Girardi’s sample was older and possibly more frail than those patients that are typically referred to our PT clinic.5 Patients at our clinic are outpatients who must be willing to travel into a major city to a tertiary care center. Although Girardi5 did not include all 6 conditions of the SOT, the study reported that 79% of their subjects had abnormal CDP scores, which is similar to 77% in this study (defined as falling at least once on an SOT condition). Keim reported that 83% of people with various vestibular diagnoses had abnormalities on the SOT,4 with persons with central vestibular dysfunction demonstrating the highest prevalence at 64%. Girardi5 demonstrated that, in the older subjects, 91% had at least 1 abnormal vestibular laboratory finding. Keim4 reported abnormal electronystagmographic findings in 42% of his patients with various vestibular disorders, which is similar to the 50% to 55% abnormal vestibular results reported in this study. Girardi5 tested older persons who had reported at least 1 significant fall during the initial examination at the falls and balance clinic, but did not report the diagnoses of their subjects, making comparison more difficult. Keim’s study4 and the present study consisted of persons with various vestibular diagnoses. The ABC findings were not expected. The people with vestibular disorders who fell 2 or more times in the past 6 months had higher (better) ABC scores by 7 percentage points than those who reported 1 fall in the previous 12 months (44% vs 51%). The 1-time fall group had the smallest n value (13) of the 3 fall-group categories, suggesting that subject number might have affected the mean ABC score. The nonfaller, 1-time, and recurrent fall groups’ mean composite SOT scores decreased as the number of patient reported falls increased. It was expected that composite SOT scores would be related to patient reported falls. No studies have previously reported that there is a relationship between the 6-item SOT composite score and patient reported falls. Girardi5 recorded the 4 conditions of the SOT as normal or abnormal rather than as a limit of stability composite score, making comparison impossible. Baloh et al33 suggested that sway velocity of CDP was no different in older adults who had fallen versus those who had not fallen, but did not report SOT composite scores for comparison. None of the above studies compared the SOT composite score to reported falls, which is the easiest measure available to clinicians as it is provided as part of the typical CDP clinical protocol. There was no relationship noted between reported falls in persons with vestibular dysfunction and vestibular laboratory test results (calorics, oculomotor, rotational chair, and positional testing). Girardi5 reported a relationship between abnormal oculomotor findings (42% of their patients [n⫽33]) and at least 1 significant fall in the previous year. Only 16% of the 55 people tested in the present study had abnormal oculomotor findings, regardless of their fall category. Girardi5 included 33 subjects in the study who were chosen because they had fallen in the last year, yet in the present study a history of falling was not an admission criteria. The data suggest that persons who are recurrent fallers are identified with SOT, similar to Girardi’s findings.5 Those people who fell 2 or more times in the last year also fell more frequently during the following conditions: 4, 5, and 6; 5 and 6; and 6. There are several limitations in this study. Retrospective work can help the reader to understand the relationship between self-reported falls and SOT scores, but cannot help the clinician to predict future fall events. A prospective study that standardizes PT intervention pre- versus posttherapy could address whether patients fall less frequently as a result of exercise. Also the same physical therapist obtained the fall
POSTUROGRAPHY AND FALLS, Whitney
history and performed the SOT. This knowledge could have introduced bias into this retrospective report. The SOT may be especially important when examining and interviewing persons who are afraid to disclose that they fall frequently. Tinetti and Williams34 have suggested that falling in community-living older persons is related to long-term care admissions, and patients may be afraid that if they report their falling honestly, they may not be able to continue to live in their desired residence. The use of the SOT may also be important as there is the potential that patients will be unable to recall a fall,27 thereby underreporting recent falls during history taking. CONCLUSIONS There was an association between those who report recurrent falls and increasing number of falls on conditions 2 through 6 of the SOT of CDP. These data suggest that people who had SOT composite scores of less than 38 on the SOT have reported an increased number of falls. The score of less than 38 on the composite SOT has the highest sensitivity and specificity for identifying persons who have reported previous falls. References 1. Whitney SL, Hudak MT, Marchetti GF. The dynamic gait index relates to self-reported fall history in individuals with vestibular dysfunction. J Vestib Res 2000;10:99-105. 2. Herdman SJ, Blatt P, Schubert MC, Tusa RJ. Falls in patients with vestibular deficits. Am J Otol 2000;21:847-51. 3. Pothula VB, Chew F, Lesser TH, Sharma AK. Falls and vestibular impairment. Clin Otolaryngol 2004;29:179-82. 4. Keim RJ. Clinical comparisons of posturography and electronystagmography. Laryngoscope 1993;103:713-6. 5. Girardi M, Konrad HR, Amin M, Hughes LF. Predicting fall risks in an elderly population: computer dynamic posturography versus electronystagmography test results. Laryngoscope 2001;111: 1528-32. 6. Di Fabio RP. Meta-analysis of the sensitivity and specificity of platform posturography. Arch Otolaryngol Head Neck Surg 1996; 122:150-6. 7. Goebel JA, Paige GD. Dynamic posturography and caloric test results in patients with and without vertigo. Otolaryngol Head Neck Surg 1989;100:553-8. 8. Voorhees RL. The role of dynamic posturography in neurotologic diagnosis. Laryngoscope 1989;99:995-1001. 9. Monsell EM, Furman JM, Herdman SJ, Konrad HR, Shepard NT. Computerized dynamic platform posturography. Otolaryngol Head Neck Surg 1997;117:394-8. 10. Jacobson GP, Newman CW, Hunter L, Blazer GK. Balance function test correlates of the dizziness handicap inventory. J Am Acad Audiol 1991;2:253-60. 11. Rose DJ, Clark S. Can the control of bodily orientation be significantly improved in a group of older adults with a history of falling? J Am Geriatr Soc 2000;48:275-82. 12. Wallmann HW. Comparison of elderly nonfallers and fallers on performance measures of functional reach, sensory organization, and limits of stability. J Gerontol A Biol Sci Med Sci 2001;56: M580-3. 13. Wolfson L, Judge J, Whipple R, King M. Strength is a major factor in balance, gait, and the occurrence of falls. J Gerontol A Biol Sci Med Sci 1995;50 Spec No:64-7. 14. Black FO, Angel C, Pesznecker SC, Gianna C. Outcome analysis of individualized vestibular rehabilitation protocols. Am J Otolaryngol 2000;21:543-51.
407
15. Telian SA, Shepard NT, Smith-Wheelock M, Hoberg M. Bilateral vestibular paresis: diagnosis and treatment. Otolaryngol Head Neck Surg 1991;104:67-71. 16. Horak FB, Jones-Rycewicz C, Black FO, Shumway-Cook A. Effects of vestibular rehabilitation on dizziness and imbalance. Otolaryngol Head Neck Surg 1992;106:175-80. 17. Brown KE, Whitney SL, Wrisley DM, Furman JM. Physical therapy outcomes for persons with bilateral vestibular loss. Laryngoscope 2001;111:1812-7. 18. O’Neill DE, Gill-Body KM, Krebs DE. Posturography changes do not predict functional performance changes. Am J Otol 1998;19: 797-803. 19. Black FO, Effron MZ, Burns DS. Diagnosis and management of drop attacks of vestibular origin: Tumarkin’s otolithic crisis. Otolaryngol Head Neck Surg 1982;90:256-62. 20. Tumarkin A. The otolithic catastrophe: a new syndrome. Br Med J 1936;1:175-7. 21. Jacob R, Furman J, Balaban C. Psychiatric aspects of vestibular disorders: disorders of the vestibular system. New York: Oxford Univ Pr; 1996. 22. Ford-Smith CD, Wyman JF, Elswick RK Jr, Fernandez T, Newton RA. Test-retest reliability of the sensory organization test in noninstitutionalized older adults. Arch Phys Med Rehabil 1995; 76:77-81. 23. Tsang WW, Hui-Chan CW. Effect of 4- and 8-week intensive tai chi training on balance control in the elderly. Med Sci Sports Exerc 2004;36:648-58. 24. Powell LE, Myers AM. The Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci 1995;50:M28-34. 25. 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;38:11-26. 26. Cumming RG, Salkeld G, Thomas M, Szonyi G. Prospective study of the impact of fear of falling on activities of daily living, SF-36 scores, and nursing home admission. J Gerontol A Biol Sci Med Sci 2000;55:M299-305. 27. Lachenbruch PA, Reinsch S, MacRae PG, Tobis JS. Adjusting for recall bias with the proportional hazard model. Methods Inf Med 1991;30:108-10. 28. Cohen H, Heaton LG, Congdon SL, Jenkins HA. Changes in sensory organization test scores with age. Age Ageing 1996;25: 39-44. 29. Shimada S, Obuchi S, Shiba Y, Okamoto M, Kakurai S. Relationship with dynamic balance function during standing and walking. Am J Phys Med Rehabil 2003;82:511-6. 30. Lord SR. Falls in the elderly: admissions, bed use, outcome and projections. Med J Aust 1990;153:117-8. 31. Colledge NR, Wilson JA, Macintyre CC, Maclennan WJ. The prevalence and characteristics of dizziness in an elderly community. Age Ageing 1994;23:117-20. 32. Sloane PD, Linzer M, Pontinen M, Divine GW. Clinical significance of a dizziness history in medical patients with syncope. Arch Intern Med 1991;151:1625-8. 33. Baloh RW, Fife TD, Zwerling L, et al. Comparison of static and dynamic posturography in young and older normal people. J Am Geriatr Soc 1994;42:405-12. 34. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med 1997;337:1279-84.
Arch Phys Med Rehabil Vol 87, March 2006
ORIGINAL ARTICLE
The Relationship Between Falls History and Computerized Dynamic Posturography in Persons With Balance and Vestibular Disorders Susan L. Whitney, PhD, PT, NCS, ATC, Gregory F. Marchetti, PhD, PT, Annika I. Schade, MD ABSTRACT. Whitney SL, Marchetti GF, Schade AI. The relationship between falls history and computerized dynamic posturography in persons with balance and vestibular disorders. Arch Phys Med Rehabil 2006;87:402-7.
© 2006 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
Objective: To describe the relationship between Sensory Organization Test (SOT) scores and reported falls in persons with vestibular and balance disorders. Design: Descriptive cross-sectional. Setting: Outpatient tertiary balance and falls clinic. Participants: One hundred physical therapy (PT) charts of people referred to a balance and falls clinic were reviewed. Criteria for inclusion were that the patients had completed the SOT of computerized dynamic posturography (CDP), had a vestibular diagnosis, and had the numbers of falls recorded from patient report within the last 6 months at the initial examination. Rotational chair, caloric testing, oculomotor test, and Activities-Specific Balance Confidence (ABC) results were recorded, if available. Interventions: Not applicable. Main Outcome Measures: The SOT composite and ABC scores as they related to reports of falls in the last 6 months. Results: The overall proportion of persons with vestibular disorders with positive 6-month history for 1 or more falls was 30% and for recurrent falls (ⱖ2) was 17%. Vestibular laboratory findings, age, sex, and vestibular diagnosis were not related to reported falls status. Patients who reported multiple falls prior to the PT examination had lower SOT composite scores than patients who reported 1 or no falls in the previous 6 months. The receiver operating characteristic curve identified a composite SOT score of less than 38 as demonstrating the highest likelihood ratio for differentiating between those people who reported no falls in the past 6 months and those who reported 2 or more falls. Conclusions: Persons who are recurrent fallers perform worse on SOT than either nonfallers or 1-time fallers. CDP performance can help guide the clinician in the development of a safe exercise program. Key Words: Accidental falls; Balance; Electronystagmography; Posture; Rehabilitation; Vestibular diseases.
HE RELATIONSHIP BETWEEN FALLS and vestibular/ T balance dysfunction is not well understood, yet recent reports suggest that falling in persons with vestibular disorders
From the Departments of Physical Therapy (Whitney) and Otolaryngology (Whitney), University of Pittsburgh, Pittsburgh, PA; Centers for Rehab Services, University of Pittsburgh Medical Center, Pittsburgh, PA (Whitney); Department of Physical Therapy, Duquesne University, Pittsburgh, PA (Marchetti); and ORL Klinik, Universitätsspital Zürich, Zürich, Switzerland (Schade). Presented in part to the Barany Society, September 29, 2002, Seattle, WA. Supported in part by the National Institutes of Health (grant no. DC 05384). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Correspondence to Susan L. Whitney, PhD, PT, NCS, ATC, University of Pittsburgh, 6035 Forbes Tower, Pittsburgh, PA 15260, e-mail: [email protected]. 0003-9993/06/8703-10348$32.00/0 doi:10.1016/j.apmr.2005.11.002
Arch Phys Med Rehabil Vol 87, March 2006
is a concern. Whitney et al1 reported that 35% of patients with various vestibular disorders under 65 years of age and 39% of those over the age of 65 reported falls during their initial examination. Fifty-one percent of persons with bilateral vestibular disorders between the ages of 65 and 74 reported falls.2 Recently, Pothula et al3 suggested that vestibular dysfunction may be responsible for unexplained falls in persons reporting to an emergency department. Better identification of persons at risk for falling may help the clinician make more targeted intervention decisions. The relationship between vestibular laboratory tests and computerized dynamic posturography (CDP) has been reported by various authors.4,5 Keim4 retrospectively compared the results of CDP and electronystagmography in persons who had peripheral, central, or both peripheral and central vestibular disorders. He reported that 42% had abnormal vestibular findings and 83% had an abnormal CDP score. Keim4 reported a higher prevalence of abnormal CDP results in persons with central vestibular disorders, suggesting that CDP is more sensitive at detecting deficits with persons with brain dysfunction.6 In a meta-analysis performed on CDP by Di Fabio,6 the sensitivity and specificity of CDP was approximately 50% when compared with other tests of vestibular function, yet Di Fabio also reports that CDP can detect abnormalities in some patients who have normal vestibulo-ocular functioning. Others4,7-9 have all suggested that CDP might be used to gain insight into the patient’s balance dysfunction that may be caused by biomechanic or neurologic disorders. Jacobson et al10 have suggested that only testing the vestibulo-ocular reflex may not provide a comprehensive evaluation of the patient’s functional deficits. Girardi et al5 reported on the relationship between selfreported falls in persons presenting to a balance and falls clinic over the age of 65 who had fallen at least 1 time in the past year (mean fall rate, 3.5 times) and scores from the Balance Trak 500, which included 4 of the 6 tests of CDP. CDP was more sensitive at identifying people who had fallen compared with ocular motor, rotational chair, positional, or caloric findings.5 Others have tested older adults using platform posturography devices that include 4 of the 6 conditions of CDP.11,12 Rose and Clark11 reported significant Sensory Organization Test (SOT) equilibrium score changes after intervention in older adults with a history of falls. The SOT mean composite score of 4 of the original 6 conditions of CDP has been shown to differentiate between fallers and nonfallers in older persons. In contrast, Wolfson et al13 demonstrated that there was no relationship between performance on the SOT and reported falls within
403
POSTUROGRAPHY AND FALLS, Whitney Table 1: Patient Demographics and Characteristics by Patient Reported Fall Status Patients Classified by Reported Fall Status Variable
All Patients
Nonfaller
1-Time Faller
Recurrent Faller
n (%) Mean age ⫾ SD (range) (n⫽100) Vestibular diagnostic category, n (%) Peripheral Central Combined (peripheral and central) No. aged ⱖ65y (%) (n⫽100) No. of women (%) (n⫽62) Mean months of symptom duration ⫾ SD (range) (n⫽99) Mean ABC score ⫾ SD (range) (n⫽91) Mean composite SOT score ⫾ SD (range) (n⫽100)
100 (100) 59⫾17 (14–90)
70 (70) 58⫾15 (14–86)
13 (13) 63⫾21 (27–90)
17 (17) 60⫾19 (22–89)
52 (52) 42 (42) 6 (6) 43 (43) 62 (62) 36⫾60 (1–360) 60⫾24 (0–100) 52⫾18 (6–85)
36 (52) 31 (44) 3 (4) 28 (40) 44 (63) 37⫾67 (1–360) 64⫾23 (0–100) 55⫾17 (6–85)
9 (69) 3 (23) 1 (8) 7 (53) 8 (62) 30⫾38 (1–96) 44⫾24 (12–96) 52⫾18 (16–77)
7 (41) 8 (47) 2 (12) 8 (47) 10 (59) 36⫾39 (1–144) 51⫾19 (20–86) 39⫾18 (7–82)
Abbreviations: ABC, Activities-Specific Balance Confidence scale; SD, standard deviation.
a 12-month period in nursing home residents, although the SOT total number of losses of balance on conditions 1 through 6 were related to ankle strength and age. Black et al14 have reported improvements in SOT scores and fewer falls during CDP testing over the course of rehabilitation in persons with peripheral vestibular disorders. Others have reported changes in SOT scores over the course of vestibular rehabilitation as an outcome measure, which have correlated with changes in impairment and various functional performance measures.15-17 O’Neill et al18 have also reported that SOT scores improved over the course of rehabilitation, yet there was not a relationship between improvements demonstrated on SOT and changes in functional performance in persons with peripheral vestibular disorders. No studies to date suggest a relationship between CDP that includes all 6 tests of the SOT and reported falls in people with vestibular disorders. In the cohort of Girardi et al,5 all subjects had fallen at least 1 time in the last year and Girardi only included 4 of the 6 SOT conditions. Therefore, the purpose of this project was to determine if there is a relationship between CDP and reported falls in persons with balance and vestibular disorders. METHODS We reviewed 100 charts of patients with vestibular disorders who had been referred to physical therapy (PT). All patients who had undergone CDP during their first PT session and who had information about reported falls in the previous 6 months were included in the study. The University of Pittsburgh Institutional Review Board approved this study as exempt. Seventy percent of the patients during the data collection period met the above 3 criteria for inclusion. Patients were not tested on CDP for the following reasons: they were unable to stand for 10 to 15 minutes unsupported, it was not possible to explain the test
due to mental status in order to obtain legitimate consent, they arrived late for the PT evaluation, they refused, or it was felt that CDP would make the patient ill because of severe motion intolerance based on the physical therapist’s judgment. The first author administered all of the CDP testing as part of the standard PT assessment and had over 15 years experience with the CDP equipment. The first author, who was also the treating physical therapist, asked all of the patients if they had fallen in the last 6 months in a systematic manner that was recorded within the patient record. None of the falls reported suggested that the patient was experiencing a Tumarkin fall from an otolith crisis.19,20 There were 70 patients who did not report any falls, 13 reported 1 fall, and 17 persons who reported more than 1 fall. Patient demographics by fall status are detailed in table 1. Between 74% and 85% of the patients underwent portions of electronystagmography testing that included rotational chair testing, positional testing, and caloric testing (table 2). The physician prescribing the laboratory tests specifically ordered the number of tests needed to make a diagnosis based on the patients presenting symptoms. The protocol of the laboratory testing has previously been reported with the criteria for what was considered a normal laboratory test.21 Based on the published criteria, an otolaryngologist determined if the individual vestibular laboratory test was normal or abnormal. A trained physical therapist performed the SOT of CDP on the EquiTest equipment. The physical therapists followed the standard EquiTest protocol for testing (version 4.0). Test-retest reliability (intraclass correlation coefficient) of the SOT has been reported to vary between .15 and .93.22,23 Percentage agreement for losses of balance during testing ranged from 77% to 100%.22 An in-depth description of the SOT protocol has been provided by Ford-Smith et al.22 The platform consists of 2 forceplates that move together and rotate when necessary,
Table 2: The Electronystagmographic Results Categorized by Reported Falls Prior to the Start of PT for Persons Who Had Completed the Full or Partial Electronystagmography Battery Fall Status
Nonfaller
1-Time Faller
Recurrent Faller
Diagnostic Test
Total
Abnormal, n (%)
Total
Abnormal, n (%)
Total
Abnormal, n (%)
Rotational chair test (n⫽74) Oculomotor test (n⫽82) Caloric test (n⫽75) Positional test (n⫽79)
50 55 52 54
24 (48) 6 (11) 23 (44) 13 (24)
9 12 10 12
3 (33) 1 (8) 5 (50) 2 (17)
15 15 13 13
8 (53) 2 (13) 4 (13) 2 (15)
NOTE. The total subject number reveals how many patients completed each test.
Arch Phys Med Rehabil Vol 87, March 2006
POSTUROGRAPHY AND FALLS, Whitney
while the patient is secured in an overhead harness to prevent a fall while viewing a visual surround. Conditions 1 and 2 include standing on a solid support surface with the eyes open and eyes closed, respectively. Condition 3 consists of having the subject stand on a solid support surface with the visual surround sway referenced (the walls move in the same direction as the subject’s sway). During conditions 4 and 5, the floor is sway referenced (the floor moves in the same direction as the subject’s sway) with eyes open and closed. Condition 6 includes both a sway referenced floor and visual surround. If a person fell on 1 of the 3 trials, it was recorded as a fall for that condition. A fall was defined as a patient taking a step, grabbing the wall or side of the platform, opening their eyes during an eyes closed condition, or when the patient had exceeded their limits of stability.22 If the operator stopped the trial, the trial was scored a 0. In addition, the composite equilibrium score was recorded for analysis that was automatically computed by the standard algorithm in the computer software. The equilibrium scores can range from 0 to 100 with 100 indicating no movement on the posture platform. Higher SOT scores indicate greater stability. Patients also completed the Activities-Specific Balance Confidence (ABC) scale, a questionnaire that is used to assess balance confidence.24 Scores of 67 and lower have recently been related to falling in older adults in the community.25 Fear of falling has been related to future fall events.26 ABC scores were recorded to determine if there was any association between balance confidence and performance on the SOT of CDP. Data Analysis We analyzed the characteristics of the study sample using descriptive statistics for continuous variables (age, symptom duration, ABC score) and frequency analysis for categoric variables (diagnostic category, vestibular test results [rotational, oculomotor, caloric, and positional tests]). We classified the subjects based on their number of self-reported falls in the last 6 months as a nonfaller, a 1-time faller, or a recurrent faller. The differences between fall groups were analyzed using 1-way analysis of variance F test with Bonferroni post hoc multiple comparisons for continuous variables and contingency table analysis with chi-square test for categoric variables. Statistical significance was determined at P less than .05 for all analyses. We analyzed the association between CDP and self-reported fall status groups using a generalized linear model with Bonferroni multiple group comparison to test the effect of fall group status on SOT score plus the effect of age as a covariate. The CDP score that best discriminated self-reported fall status group was determined using a receiver operating characteristic (ROC) curve analysis. In addition, likelihood ratios were calculated based on the optimal sensitivity and specificity of the CDP score for the self-reported fall groups. The ROC and likelihood ratio analyses were based on the self-reported fall group and SOT differences identified in the generalized linear model. CDP performance was dichotomously measured based on fall behavior on each SOT condition. Any loss of balance during each of the 6 SOT conditions (3 trials per condition) was recorded categorically as a fall for that SOT condition. The significance of the association between self-reported fall status groups (nonfaller, 1-time faller, recurrent faller) and falling on 1 or more SOT conditions as defined above was determined using a contingency table with chi-square analysis. Arch Phys Med Rehabil Vol 87, March 2006
RESULTS Table 1 contains demographic data for the total sample plus for all patients as classified by their reports of falling in the 6 months prior to examination. There was no significant difference among patient self-reported fall status groups on the variables of age or symptom duration (P⬎.72). In addition, there was no difference in the proportion of patients’ age 65 and older by self-reported falls status groups. There was no relationship between sex or diagnostic category (peripheral, central, and both peripheral and central) and the nonfaller, 1-time faller, or recurrent faller groups. There was a significant difference in mean ABC score among the 3 self-reported falls status groups (F2⫽5.17, P⬍.01). Multiple group comparisons (Bonferroni) indicated that patients who reported no falls in the past 6 months demonstrated greater balance confidence, as measured by the ABC, than patients reporting 1 fall in the past 6 months (P⬍.02). There was no significant difference in balance confidence (mean ABC score) between the patient group who reported no past falls and the recurrent fall group. There was no significant difference in mean ABC score between the patient group reporting a single fall and the group reporting recurrent falls in the 6 months before examination. Figure 1 illustrates the group means and 95% confidence intervals (CIs) for the ABC score for patients stratified by self-reported fall history in the previous 6 months. Mean SOT scores and self-reported falls history within the past 6 months were significantly related (F3⫽5.81, P⬍.01). Figure 2 illustrates the means and 95% CIs for the SOT score by reported fall history in the past 6 months. Multiple group comparisons (Bonferroni) indicated that the patient group who reported no falls in the previous 6 months demonstrated higher mean SOT scores than the recurrent self-reported falls group (P⬍.01). There was no difference in mean SOT score between the patient group who reported no falls and the group reporting 1 fall in the 6 months prior to examination. The patient group
80.00
95% CI of Mean ABC Score
404
*
70.00
60.00
50.00
40.00
30.00
20.00 Recurrent Faller
One-Time Faller
Nonfaller
Fall History Past 6 Months
Fig 1. Mean ABC scores with 95% CIs for patients stratified by self-reported falls history for 6 months prior to the start of PT (recurrent faller, 1-time faller, nonfaller). *Significant difference (P<.01).
405
POSTUROGRAPHY AND FALLS, Whitney
0.9 60.00
50.00
0.8 Sensitivity (tru e p o sitives)
95% CI of Composite SOT Score
1
*
40.00
0.7 0.6
*
0.5 0.4 0.3 0.2 0.1
30.00
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 - Specificity (false positives) Recurrent Faller
One-Time Faller
Nonfaller
Fall History Past 6 Months Fig 2. The mean composite SOT scores with the 95% CIs for patient stratified by self-reported fall status (recurrent, 1-time, nonfaller). *Significant difference (P<.01).
who reported falling 1 time in the 6 months prior to PT examination demonstrated a higher mean SOT score than the group who reported 2 or more falls. This difference, however, was not statistically significant (P⫽.08). This failure of this difference to reach statistical significance may represent a type II error based on the small sample size in the 1-time and recurrent fall groups. Age was a significant covariate in the relationship between the SOT and the self-reported falls groups (P⫽.02). Patient age was found to account for 5% of the total SOT score variance. Due to this significant covariance, the self-reported falls group differences in mean SOT were determined based on adjustment for age. The ROC curve was constructed to determine if the composite SOT score could be used to discriminate between the nonfalling and recurrent falling self-reported falls groups. The 1-time self-reported fall group was not considered in the ROC curve analysis because of group similarities to both the nonfalling group (similar age-adjusted mean SOT scores) and the recurrent falling group (similar mean ABC scores). Analysis of the ROC curve (fig 3) revealed that optimal identification of people who report recurrent falls was associated with a composite SOT score of less than 38 (sensitivity, 53%; specificity, 87%). The likelihood ratio at the SOT value of less than 38 was 4.12, indicating an over 4-fold increase in the pre- to posttest odds of detecting a person who reported recurrent falls in the 6 months prior to examination (24% pretest odds to 99.8% posttest odds). Among the people who did not report a fall and those people who reported multiple falls, the probability that any individual patient reported multiple falls was 20% (17/87). Applying the likelihood ratio of 4.12 increases that probability to 50% (150% increase over baseline). After examining the relationship between SOT conditions in which 1 or more falls occurred over the 3 trials per condition by nonfaller, 1-time faller, and recurrent faller, the strongest association was seen with falls on SOT conditions 5 and 6
Fig 3. ROC curve (with point of highest likelihood ratio) for the discrimination of patients with self-reported recurrent fall history from patients with no history of self-reported falls using the SOT score. Area under curve equals .75 (P<.01). Diagonal line indicates the area of no discrimination. *Cutoff below 38 points as the location of the maximum likelihood ratio (4.12).
(table 3). There were also significant relationships between falls on SOT conditions 6, and falls on SOT conditions 4, 5, and 6. In interpreting these associations, it is imperative to note that the categoric SOT falling conditions (eg, falling on conditions 5 and 6, condition 6, or conditions 4, 5, and 6) are not mutually exclusive for each patient. For example, a patient identified as having fallen on conditions on 5 and 6 may or may not have also fallen on conditions 2, 3, and 4. Table 2 provides the vestibular test results categorized by patient status (nonfaller, 1-time faller, recurrent faller) based on the number of self-reported falls in the 6 months prior to the PT examination. There was no relationship between diagnostic test results (normal versus abnormal findings of each of the vestibular tests) and self-reported fall group status. DISCUSSION There is a relationship between the composite score of the SOT and reported falls in persons with vestibular disorders. The use of the SOT composite score can help confirm the clinician’s other examination findings about the patient’s pos-
Table 3: Proportion of Patients in Each Self-Reported Fall History Group Who Fell on SOT Conditions 5 and 6, Condition 6, and Conditions 4, 5, and 6* Self-Reported Fall History
Fall on SOT Condition
Recurrent
1-Time
No Falls
P†
5 and 6 (%) 6 (%) 4, 5, and 6 (%)
88 88 47
54 69 31
46 53 17
⬍.01 ⬍.03 ⬍.03
*Persons who fell on conditions 5 and 6, condition 6, and conditions 4, 5, and 6 might have also fallen on other conditions. These 2 categories were the only significant associations with self-reported fall status. † Chi-square test with 2 degrees of freedom.
Arch Phys Med Rehabil Vol 87, March 2006
406
POSTUROGRAPHY AND FALLS, Whitney
tural stability. Although the sensitivity of the SOT cut score at less than 38 was only 53% when comparing nonfallers versus recurrent fallers, the specificity of the cut score was 87%. To date, there has not been a published cut score on the SOT that has been related to past reports of falls. Obviously, prospective work would be ideal to identify future falls. The question of how useful the information gained from the SOT in the prediction of future falls can certainly not be answered from a retrospective study. There is also the potential source of error with recall bias in using patient self-report to determine fall history.27 A prospective study in this setting could not be conducted in the absence of PT intervention to reduce fall risk for these patients. Additionally, even prospective studies of falls rely on patient self report to ascertain the occurrence of a fall event and are not free of recall bias. While the retrospective design limits the predictive value of the results, we feel that descriptive information gained from an analysis of the clinical and fall history associations with CDP does provide a link between balance impairments and disability in patients with vestibular dysfunction. From these data, no link to future fall risk can be made. There was a relationship between observed falling on conditions 2 through 6 on the SOT and self-reported falls, which has previously not been reported in persons with vestibular disorders. Age, sex, and length of symptoms did not affect the patient’s reported fall status. There was no significant association between patient self-reported fall status and age (⬍65 or 65 and older), unlike the work of Cohen28 and Shimada29 and colleagues, which indicated that SOT number of falls were related to age. Cohen28 and Shimada29 included people without balance dysfunction, unlike the present sample, which consisted of people with balance disorders, possibly suggesting that the age effect on SOT scores is no longer a factor because all patients had balance disorders. Typically as one increases in age, there is an increased risk of falling.30 Wolfson et al13 reported an increase in the number of SOT falls with increasing age, although their subjects were nursing home residents with a mean age of 77 or older. The subjects in the present study were younger, had balance and vestibular disorders, and were ambulatory outpatients. In the study by Girardi et al,5 all subjects had to have fallen at least once in the last year and had to be over the age of 65 in order to meet the criteria for admission to their study. Our data suggest that, regardless of age, people with vestibular disorders report falling. Cohen28 noted that no normative subjects fell on conditions 1, 2, or 3 and only 1 person fell on condition 4, indicating that conditions 1 through 4 were fairly easy for Cohen’s sample of healthy community-living older adults. Similar to Cohen,28 none of our subjects fell on condition 1. Four people fell on at least 1 trial on condition 2, 13 people on condition 3, and 33 (33%) persons fell on at least 1 trial of condition 4 in the present study. It was interesting that a similar number of persons with vestibular disorders fell on condition 4 (30% vs 37%), whether or not they were 65 years of age or older. Wolfson13 suggested that older community-dwelling women may be more likely to fall on CDP during sway referenced conditions with visual compromise than men. Sixty-three percent of the nonfallers, 62% of the 1-time fallers, and 59% of the 2 or more fallers were reported by women. Overall, “dizziness” is reported more in women than in men, yet sex did not affect their fall category (non, 1-time, or recurrent faller).31,32 Girardi’s subjects5 had a mean age of 78 and had experienced at least 1 “significant” fall within the last year. The average fall rate of their sample was 3.5 falls in the previous year versus 30% of our subjects who fell in the last 6 months, Arch Phys Med Rehabil Vol 87, March 2006
indicating that Girardi’s sample was older and possibly more frail than those patients that are typically referred to our PT clinic.5 Patients at our clinic are outpatients who must be willing to travel into a major city to a tertiary care center. Although Girardi5 did not include all 6 conditions of the SOT, the study reported that 79% of their subjects had abnormal CDP scores, which is similar to 77% in this study (defined as falling at least once on an SOT condition). Keim reported that 83% of people with various vestibular diagnoses had abnormalities on the SOT,4 with persons with central vestibular dysfunction demonstrating the highest prevalence at 64%. Girardi5 demonstrated that, in the older subjects, 91% had at least 1 abnormal vestibular laboratory finding. Keim4 reported abnormal electronystagmographic findings in 42% of his patients with various vestibular disorders, which is similar to the 50% to 55% abnormal vestibular results reported in this study. Girardi5 tested older persons who had reported at least 1 significant fall during the initial examination at the falls and balance clinic, but did not report the diagnoses of their subjects, making comparison more difficult. Keim’s study4 and the present study consisted of persons with various vestibular diagnoses. The ABC findings were not expected. The people with vestibular disorders who fell 2 or more times in the past 6 months had higher (better) ABC scores by 7 percentage points than those who reported 1 fall in the previous 12 months (44% vs 51%). The 1-time fall group had the smallest n value (13) of the 3 fall-group categories, suggesting that subject number might have affected the mean ABC score. The nonfaller, 1-time, and recurrent fall groups’ mean composite SOT scores decreased as the number of patient reported falls increased. It was expected that composite SOT scores would be related to patient reported falls. No studies have previously reported that there is a relationship between the 6-item SOT composite score and patient reported falls. Girardi5 recorded the 4 conditions of the SOT as normal or abnormal rather than as a limit of stability composite score, making comparison impossible. Baloh et al33 suggested that sway velocity of CDP was no different in older adults who had fallen versus those who had not fallen, but did not report SOT composite scores for comparison. None of the above studies compared the SOT composite score to reported falls, which is the easiest measure available to clinicians as it is provided as part of the typical CDP clinical protocol. There was no relationship noted between reported falls in persons with vestibular dysfunction and vestibular laboratory test results (calorics, oculomotor, rotational chair, and positional testing). Girardi5 reported a relationship between abnormal oculomotor findings (42% of their patients [n⫽33]) and at least 1 significant fall in the previous year. Only 16% of the 55 people tested in the present study had abnormal oculomotor findings, regardless of their fall category. Girardi5 included 33 subjects in the study who were chosen because they had fallen in the last year, yet in the present study a history of falling was not an admission criteria. The data suggest that persons who are recurrent fallers are identified with SOT, similar to Girardi’s findings.5 Those people who fell 2 or more times in the last year also fell more frequently during the following conditions: 4, 5, and 6; 5 and 6; and 6. There are several limitations in this study. Retrospective work can help the reader to understand the relationship between self-reported falls and SOT scores, but cannot help the clinician to predict future fall events. A prospective study that standardizes PT intervention pre- versus posttherapy could address whether patients fall less frequently as a result of exercise. Also the same physical therapist obtained the fall
POSTUROGRAPHY AND FALLS, Whitney
history and performed the SOT. This knowledge could have introduced bias into this retrospective report. The SOT may be especially important when examining and interviewing persons who are afraid to disclose that they fall frequently. Tinetti and Williams34 have suggested that falling in community-living older persons is related to long-term care admissions, and patients may be afraid that if they report their falling honestly, they may not be able to continue to live in their desired residence. The use of the SOT may also be important as there is the potential that patients will be unable to recall a fall,27 thereby underreporting recent falls during history taking. CONCLUSIONS There was an association between those who report recurrent falls and increasing number of falls on conditions 2 through 6 of the SOT of CDP. These data suggest that people who had SOT composite scores of less than 38 on the SOT have reported an increased number of falls. The score of less than 38 on the composite SOT has the highest sensitivity and specificity for identifying persons who have reported previous falls. References 1. Whitney SL, Hudak MT, Marchetti GF. The dynamic gait index relates to self-reported fall history in individuals with vestibular dysfunction. J Vestib Res 2000;10:99-105. 2. Herdman SJ, Blatt P, Schubert MC, Tusa RJ. Falls in patients with vestibular deficits. Am J Otol 2000;21:847-51. 3. Pothula VB, Chew F, Lesser TH, Sharma AK. Falls and vestibular impairment. Clin Otolaryngol 2004;29:179-82. 4. Keim RJ. Clinical comparisons of posturography and electronystagmography. Laryngoscope 1993;103:713-6. 5. Girardi M, Konrad HR, Amin M, Hughes LF. Predicting fall risks in an elderly population: computer dynamic posturography versus electronystagmography test results. Laryngoscope 2001;111: 1528-32. 6. Di Fabio RP. Meta-analysis of the sensitivity and specificity of platform posturography. Arch Otolaryngol Head Neck Surg 1996; 122:150-6. 7. Goebel JA, Paige GD. Dynamic posturography and caloric test results in patients with and without vertigo. Otolaryngol Head Neck Surg 1989;100:553-8. 8. Voorhees RL. The role of dynamic posturography in neurotologic diagnosis. Laryngoscope 1989;99:995-1001. 9. Monsell EM, Furman JM, Herdman SJ, Konrad HR, Shepard NT. Computerized dynamic platform posturography. Otolaryngol Head Neck Surg 1997;117:394-8. 10. Jacobson GP, Newman CW, Hunter L, Blazer GK. Balance function test correlates of the dizziness handicap inventory. J Am Acad Audiol 1991;2:253-60. 11. Rose DJ, Clark S. Can the control of bodily orientation be significantly improved in a group of older adults with a history of falling? J Am Geriatr Soc 2000;48:275-82. 12. Wallmann HW. Comparison of elderly nonfallers and fallers on performance measures of functional reach, sensory organization, and limits of stability. J Gerontol A Biol Sci Med Sci 2001;56: M580-3. 13. Wolfson L, Judge J, Whipple R, King M. Strength is a major factor in balance, gait, and the occurrence of falls. J Gerontol A Biol Sci Med Sci 1995;50 Spec No:64-7. 14. Black FO, Angel C, Pesznecker SC, Gianna C. Outcome analysis of individualized vestibular rehabilitation protocols. Am J Otolaryngol 2000;21:543-51.
407
15. Telian SA, Shepard NT, Smith-Wheelock M, Hoberg M. Bilateral vestibular paresis: diagnosis and treatment. Otolaryngol Head Neck Surg 1991;104:67-71. 16. Horak FB, Jones-Rycewicz C, Black FO, Shumway-Cook A. Effects of vestibular rehabilitation on dizziness and imbalance. Otolaryngol Head Neck Surg 1992;106:175-80. 17. Brown KE, Whitney SL, Wrisley DM, Furman JM. Physical therapy outcomes for persons with bilateral vestibular loss. Laryngoscope 2001;111:1812-7. 18. O’Neill DE, Gill-Body KM, Krebs DE. Posturography changes do not predict functional performance changes. Am J Otol 1998;19: 797-803. 19. Black FO, Effron MZ, Burns DS. Diagnosis and management of drop attacks of vestibular origin: Tumarkin’s otolithic crisis. Otolaryngol Head Neck Surg 1982;90:256-62. 20. Tumarkin A. The otolithic catastrophe: a new syndrome. Br Med J 1936;1:175-7. 21. Jacob R, Furman J, Balaban C. Psychiatric aspects of vestibular disorders: disorders of the vestibular system. New York: Oxford Univ Pr; 1996. 22. Ford-Smith CD, Wyman JF, Elswick RK Jr, Fernandez T, Newton RA. Test-retest reliability of the sensory organization test in noninstitutionalized older adults. Arch Phys Med Rehabil 1995; 76:77-81. 23. Tsang WW, Hui-Chan CW. Effect of 4- and 8-week intensive tai chi training on balance control in the elderly. Med Sci Sports Exerc 2004;36:648-58. 24. Powell LE, Myers AM. The Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci 1995;50:M28-34. 25. 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;38:11-26. 26. Cumming RG, Salkeld G, Thomas M, Szonyi G. Prospective study of the impact of fear of falling on activities of daily living, SF-36 scores, and nursing home admission. J Gerontol A Biol Sci Med Sci 2000;55:M299-305. 27. Lachenbruch PA, Reinsch S, MacRae PG, Tobis JS. Adjusting for recall bias with the proportional hazard model. Methods Inf Med 1991;30:108-10. 28. Cohen H, Heaton LG, Congdon SL, Jenkins HA. Changes in sensory organization test scores with age. Age Ageing 1996;25: 39-44. 29. Shimada S, Obuchi S, Shiba Y, Okamoto M, Kakurai S. Relationship with dynamic balance function during standing and walking. Am J Phys Med Rehabil 2003;82:511-6. 30. Lord SR. Falls in the elderly: admissions, bed use, outcome and projections. Med J Aust 1990;153:117-8. 31. Colledge NR, Wilson JA, Macintyre CC, Maclennan WJ. The prevalence and characteristics of dizziness in an elderly community. Age Ageing 1994;23:117-20. 32. Sloane PD, Linzer M, Pontinen M, Divine GW. Clinical significance of a dizziness history in medical patients with syncope. Arch Intern Med 1991;151:1625-8. 33. Baloh RW, Fife TD, Zwerling L, et al. Comparison of static and dynamic posturography in young and older normal people. J Am Geriatr Soc 1994;42:405-12. 34. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med 1997;337:1279-84.
Arch Phys Med Rehabil Vol 87, March 2006