Prediction of Functional Limitations in Balance after Tests of Tandem Walking and Standing Balance in Seniors

Abstract

Objective:

Seniors who may not complain of balance problems may be developing subtle balance problems which may affect future function. This study sought to determine if subtle problems would be predicted by standard balance testing.

Setting:

The Geriatric Medicine clinic at an academic tertiary care center.

Subjects:

Geriatricians referred 27 people who were ambulatory without gait aids and had no apparent neurologic or musculoskeletal impairments affecting their balance or gait performance.

Design:

Initially subjects were screened with some questions about vertigo, tests of standing balance on foam with eyes closed (Romberg tests) with head still, and head nodding (pitch) and shaking (yaw) at 0.3 Hz, and tandem walking with eyes closed (TW). Approximately two years later 22 subjects were available to be re-screened.

Results:

At the second test date subjects’ scores declined slightly but significantly on all 3 variations of the Romberg tests: with head still, head pitch, and head yaw. p≥ 0.02, but TW did not change significantly. TW was the best predictor of balance performance; the lowest-scoring subjects at the second test used a cane, had had hip surgery, or were scheduled for hip surgery: sensitivity = 1.0, specificity = .84. The second-best predictor was Romberg on foam with eyes closed and head still.

Conclusions:

These data suggest that some seniors may have subtle musculoskeletal impairments that might be indicated by simple balance tests. A finding on these tests might indicate the need for further work-up for balance-related problems. These quick, inexpensive tests can be administered by ancillary staff and may add important information to the initial visit and annual follow-up visits.

Brief description:

Seniors who could walk unassisted were tested on heel-toe walking and standing on soft foam with head still or moving slowly, all with eyes closed. When they were retested 1.5 years later people with the lowest scores on heel-toe walking and standing on foam with head still were mostly likely to be walking with a cane, or to have had or need to have hip surgery. Poor scores on these simple balance tests may indicate subclinical problems.

Introduction

Many studies describe measures to predict catastrophic balance events (i.e., falls and hip fractures) in elderly patients. Less catastrophic changes that affect balance may precede falls and may impair functional performance and quality of life. These changes, too, are worth predicting, such as performance one year after surgery for hip fracture¹. The literature has guidelines on fall prevention and what clinicians should do after a patient falls².

Screening for all of the problems that might cause falls or other balance problems within the limited time available to physicians in their clinic visits is challenging. Specific types of balance can be readily screened. The literature has many studies of balance tests done in the laboratory or in the clinic for screening; some of them have been shown to be useful for predicting vestibular disorders. For example, new norms have been published recently for two common, useful screening tests of balance that may be useful in predicting vestibular disorders^3–5. The value of these quick screening tests for predicting future functional performance or balance skill is not clear. The goal of this study was to determine if poor performance on standing and walking balance tests predicted future balance problems or related changes in health status. We used a well-known and normed modification of the Romberg test, the Clinical Test of Sensory Interaction and Balance (CTSIB), and Tandem Walking (TW) with eyes closed ^3,4. We hypothesized that CTSIB and TW scores would decline significantly over time.

Methods

Subjects and consent

Subjects were 27 adults, including 6 males and 21 females. At the initial test (Test 1) mean ages were 75.3 yrs., SD, 7.2, range 63 – 85. At the second test (Test 2), mean ages were 77.9, SD, 7.2, range 65.6 to 87.6. All subjects were referred to the study by one of three board-certified geriatricians when they visited the geriatric medicine clinic for routine care. Criteria were: no known neurologic or otologic problems other than presbycusis, no significant orthopedic problems, e.g. joint replacement or arthritis needing medication, ability to stand and walk unassisted and without a cane or rollator for at least 3 minutes, no history of falls, and competence to understand the instructions and to give informed consent. These geriatricians know their patients well, so they were able to refer patients who they considered to be essentially healthy for their ages. At Test 1 all subjects met the referral criteria.

All subjects gave written informed consent prior to participation in Test 1. All subjects were re-consented prior to Test 2. This study was approved by the Institutional Review Board for Human Subjects Research at our institution.

Methodology

All subjects were given Test 1 during a 6-week period from September 2015 to November 2015 when they visited their geriatricians for routine care. They were given Test 2 at the same facility during a 12-week period from September 2017 to January 2018 when they returned to visit their geriatricians for routine care.

At Test 1 subjects were screened for balance impairments by testing on two measures, CTSIB with three conditions and TW. For both tests subjects removed their shoes but wore their socks and during the tests they crossed their arms. TW subjects took 10 steps in tandem (heel-to-toe) fashion with their eyes open for practice, to learn the task. Then they took 5 steps in tandem fashion with eyes closed, with close guarding from a staff member, to learn the task. Finally, they took 10 tandem steps with eyes closed. The dependent measure was the total number of correct tandem steps out of 10. Errors were opening their eyes, taking more than one step prior to the tandem step, moving out of tandem step positions or moving their arms ³.

For CTSIB, subjects stood on 10 cm-thick, medium density Sunmate foam (Dynamic Systems, Leicester, NC: sunmatecushions.com), which is continuously compliant. They stood with their feet together, i.e. side-by-side. The foam was placed near a wall for safety and the staff member provided close guarding. For the first 5 seconds subjects stood with their eyes open to get used to the foam. Then they were asked to close their eyes and perform the test conditions. In between test conditions they opened their eyes and stepped off the foam to rest for 5 seconds or more, as needed. CTSIB was performed under three conditions: 1) standing still without moving the head (head still); 2) shaking the head left-right (i.e. “No”) in yaw rotations (head yaw); 3) nodding the head up-down (i.e. “Yes”) in pitch (head pitch). For Conditions 2 and 3, (the yaw and pitch head movement trials) subjects heard an oscillating sound at approximately 0.3 Hz. While standing on the floor with eyes open they practiced moving their heads in time to the sound, left-right (yaw) and then up-down (pitch). Head movements were made through a moderate range of motion, comfortable for the subject. The dependent measure for all trials was the amount of time the subject was able to perform the test, up to 30 seconds per trial ⁴.

Subjects were asked questions from the United States National Health Interview Survey (NHIS) at both test dates. Using the NHIS wording, which included excluding times when drinking alcohol, at Test 1 subjects were asked questions about several symptoms, using the same initial phrase each time. Those questions were: “During the past 12 months, have you had a problem with dizziness or balance”, “dizziness, lightheadedness, feeling as if you are going to pass out or faint, or with unsteadiness or imbalance”, “a spinning or vertigo sensation, a floaty or disconnected sensation, feeling lightheaded without a sense of motion, feeling as if you are going to pass out or faint, blurring of your vision when you move your head, feeling off balance or unsteady, other dizziness or balance problem” ⁶.

At Test 2, subjects were re-tested on TW and CTSIB, with the same practice trials as at Test 1. They also answered questions about history of falls since Test 1, current use of a cane or rollator, orthopedic surgery, or vestibular, neurologic or musculoskeletal problems they had developed since Test 1. They were asked the NHIS questions about vertigo and balance again, At Test 2 the initial phrase was modified by asking them to respond how they felt since the date of Test 1, rather than “during the past 12 months”.

The research assistant who performed most of the tests was trained by the senior investigator and she practiced until their inter-rater reliability was at least 0.90. She had an undergraduate degree but had no special expertise in balance or geriatric medicine. The senior investigator also tested some subjects. CTSIB was timed with a hand-held stopwatch. Balance tests were observed but not video-recorded. Scores and responses were recorded on paper forms and later transferred the data to the database.

Statistical Methods

Pearson correlations were used to examine the association between Tests 1 and 2. Paired t-tests were also used to determine significant changes from Test 1 to Test 2 for various parameters. The ability of CTSIB and TW scores at Test 1 to predict various outcomes at Test 2 (e.g., falls/use of cane, lightheadedness and vertigo problems) was measured by use of area under the curve via logistic regression analysis. P<0.05 was considered significant. All analysis were performed using SAS statistical software (version 9.4, Cary, NC)

Results

Although 27 subjects were tested on balance at Test 1, 22 subjects were tested on balance at Test 2. The other 5 subjects were no longer able to stand and walk independently. Two subjects had recently had hip surgery, one subject had recently had knee surgery, one subject had to use a cane and was scheduled to have hip surgery, and one other subject had to use a cane to stand or walk due to arthritis. Those subjects were able to participate by answering the questions, however.

We used norms from published tables. For TW the cutpoints were: age 60 to 69, 4 steps; ages 70 to 79, 3 steps, age ≥ 80 1 step. For CTSIB head still, the cut-points were: age 60 to 60, 17 seconds; age 70 to 79, 16 seconds, age ≥ 80, 6 seconds. For CTSIB head yaw cut-points were: age 60 to 60, 12 seconds; age 70 to 79, 12 seconds, age ≥ 80, 6 seconds. For CTSIB, head pitch, cut-points were: age 60 to 60, 14 seconds; age 70 to 79, 12 seconds, age ≥ 80, 5 seconds.

Males and females did not differ significantly on Test 1 and 2 scores on TW or CTSIB scores. With the limited number of males that finding should be consider with caution. The entire cohort was considered as a single group for the rest of the analyses. Age was not significantly related to Tandem Walking or any of the CTSIB trials.

When Tests 1 and 2 were compared, TW scores did not change significantly over tests, p=0.38. CTSIB time scores, however, decreased significantly for head still (p=0.0007), head yaw (p=0.0004), and for head pitch (p=0.02). (See Table 1.) Test 1 scores for CTSIB were only weakly to moderately related to Test 2 scores for TW or CTSIB trials. Test 2 CTSIB scores were not related or were only weakly related to Test 2 TW scores. Test 2 CTSIB scores were moderately to strongly related to other Test 2 CTSIB scores. (See Table 2.)

Table 1.

Test 1 and Test 2 scores on TW (median number of correct tandem steps, range) and CTSIB (mean trial duration (sec), SD, range).

	Test 1 (N=27)	Test 2 (N=22)
TW	4 to 8	5 to 7
CTSIB head still, time	11.6 ± 10.2, 1.6 to 30	6.5 ± 8.0, 0 to 30
CTSIB head yaw, time	10.5 ± 9.9, 0 to 30	5.8 ± 7.7, 0 to 30
CTSIB head pitch, time	10.2 ± 9.9, 0 to 30	5.4 ± 7.6, 0.9 to 30

Table 2.

Correlations between age, TW and CTSIB. Pearson r correlations are shown.

	Test 2 TW	Test 2 CTSIB head still	Test 2 CTSIB head yaw	Test 2 CTSIB head pitch
Test 1 age	−0.43 (p=0.04)	−0.4 (p=0.06)	−0.29 (p=.1)	−0.37 (p=0.08)
Test 1 TW	0.23 (p=0.29)	0.23 (p=0.29)	0.43 (p=0.04)	0.27 (p=0.24)
Test 1 CTSIB head still	0.23 (p=.29)	0.40 (p=0.04)	0.13 (p=0.50)	0.29 (p=0.17)
Test 1 CTSIB head yaw	0.43 (p=0.04)	0.41 (p=0.06)	0.47 (p=0.03)	0.60 (p=0.003)
Test 1 CTSIB head pitch	0.27 (p=0.24)	0.09 (p=0.70)	0.25 (p=0.29)	0.33 (p=0.16)
Test 1 Tandem Walking	1.00	0.23 (p=0.29)	0.43 (p=0.04)	0.27 (p=0.24)
Test 2 CTSIB head still	0.36 (p=0.09)	1.00	0.79 (p<0.0001)	0.67 (p=0.0006)
Test 2 CTSIB head yaw	0.39 (p=0.08)		1.00	0.91 p<0.0001
Test 2 CTSIB head pitch	0.44 (p=0.04)			1.00

The best predictor for using a cane and/ or having had hip surgery was the Test 1 TW score: ROC = 0.97; CI .91, 1.0; sensitivity = 1.0, specificity = .84. The second-best predictor for cane use and/ or having had hip surgery was the Test 1 CTSIB head still score: ROC = 0.71; CI .25, 1.0; sensitivity, 0.67, specificity, 0.92. Test 1 CTSIB head pitch was a moderate predictor of lightheadedness: ROC=0.75; CI=.56. .95; sensitivity, 0.83, specificity, 0.64. Test 1 CTSIB head pitch was also a moderate predictor of feeling off balance: ROC=0.72; CI .52, .92; sensitivity, 0.70, specificity, 0.73. The norms for these tests were based on published data ^3,4.

Fifteen people reported having had one or more falls. No tests at Test 1 predicted future falls or dizziness/ vertigo reported at Test 2.

Discussion

These tests are useful for quick balance screening in the clinic by ancillary health staff. The inability of seniors to perform these tests indicates the occurrence of some balance problems. When a patient presents with low scores, lower than appropriate for that person’s age group, the primary care physician should consider exploring these issues further. For example, the physician might want to determine if the patient has an underlying musculoskeletal problems by referring for imaging studies of the hip or knee or a consult with orthopedics or physical medicine. If the problem does not require surgery, the physician might consider referral for occupational therapy and physical therapy, to improve strength, motor control, and independence in activities of daily living. If improvement in strength and motor control are not possible therapy may still be useful for developing adaptive strategies to improve safety, independence and participation in daily life activities.

Lower extremity disability such as decreased strength, sensation or balance, perhaps necessitating use of a cane or walker, is associated with a high risk for falls ⁷. Falls in seniors are associated with high morbidity and mortality, including hip fractures⁸. The rate of death from falls has been increasing⁹. The literature on hip fractures and pre-fracture screening has conflicting findings. One study showed that TW is not a good predictor of future hip fractures¹⁰ but another study showed that TW is useful in combination with a mathematical model¹¹. Ability to perform single leg stance predicted hip fractures in one study¹², but not in another one¹¹. Many patients with vestibular disorders cannot perform single leg stance. The evidence about the optimal test methodologies for predicting hip fracture are unclear. Therefore, rather than try to predict hip fractures, specifically, perhaps predicting future significant balance problems would be more efficacious. Upon finding balance problems the physician could explore the underlying problems in more detail.

Screening takes very little time, even for the busy physician. Ancillary clinic staff can easily ask questions during routine intake at the start of the visit. With minimal extra time, i.e. 2 to 3 minutes, after 5 to 10 minutes of training clinic staff such as medical assistants would be able to perform TW and CTSIB after they take vital signs. The investment in equipment is also minimal because a slab of foam and a stopwatch are inexpensive. Therefore, we recommend incorporating these simple, effective tests into the routine at the initial visit with seniors and periodically after that, such as during the annual wellness visit.

The geriatrician’s goal for screening is to identify functional limitations that may ultimately affect the patient’s independence and quality of life. Such limitations may contribute to the need for a higher level of care, possibly including hospitalization or institutionalization. Therefore, although we recommend follow-up at the annual visit, the specific decision about follow-up is part of the physician’s treatment plan. The physician may want to recommend follow-up sooner if some developing condition is suspected or when the patient has been completed therapy.

These test results may be useful as a starting point for helping the physician to determine the treatment plan for patients with multifactorial problems but these tests do not provide data about the underlying problems. Knowing if the balance problem is caused by a vestibular impairment, presbystasis (disequilibrium of aging), or arthritic changes to the hip, for example, is essential in selecting the optimal treatment. A diagnosis of vestibular impairment or presbystasis may lead to referrals for vestibular and/ or balance rehabilitation by occupational and physical therapists, without medication, device use, or further medical intervention. A diagnosis of hip arthritis, however, may lead to a different treatment plan.

The study has some limitations. We would have liked to have followed all of our subjects who were tested initially, but not all of them were available later. Therefore, the sample size is relatively small, although larger than a pilot study. Also, in the limited context of this study we were unable to follow-up subjects through referrals to other services, such as physical medicine or orthopedics. The study had more female than male subjects due to the use of a convenience sample over a limited period of time. Thus generalization for males should be made with caution.

Because subjects were referred to the study by geriatricians who screened their patients for ability to participate, we did not use all consecutive patients in the geriatric medicine clinic over a certain period of time. The system could have included unknown referral biases by the geriatricians.

Conclusion

These simple balance tests can be given quickly by ancillary staff, without expensive technology. These tests are known to be useful for clinical assessment. We have shown here for the first time that they may also give the physician insight into developing subclinical problems, too. This information may be the first step in early identification and treatment of problems that may lead to falls and disablement.

Key points.

• Community-dwelling, independently ambulatory seniors can be tested on balance quickly and easily as part of standard annual exam.

• Poor balance scores may indicate the presence of subclinical musculoskeletal problems.

• Poor-scoring patients might benefit from further work-up of these problems.