One-third of people older than 65 and one-half of people older than 80 fall at least once per year.[1-4] Research indicates that the decrease in functional fitness caused by reduced sensation and vestibular dysfunction[5] is a risk factor for falls and fall-related injuries,[6] as is possible age-related decline in the quality and speed of integration of the balance system and unconscious processes.[7] The increase in body sway related to poor tactile sensitivity and poor joint sense on a firm surface, along with reduced reflex speeds, poor coordination, and loss of flexibility and strength, all lead people to rely more on upper leg and trunk strength to balance.[8] Reduced power capacity, which is another result of reduced speed of movement, leads to a slower gait that creates excessive muscle fatigue.[9] A decrease in functional fitness also occurs with sarcopenia (age-related muscle mass loss), which results from deconditioning, physical inactivity, or chronic disease.[5]

Increased fall rate is associated with mobility impairment, dizziness,[10] and complex interactions of intrinsic and extrinsic risk factors.[7,11] Intrinsic risk factors may lead to insufficient muscle strength and flexibility to lift one’s body mass from a chair or toilet seat, dress, or climb into a bath.[12] Hip and knee extensor strength is needed to stand from a sitting position, and hip and knee flexors must retain flexibility in order to maintain posture and not limit walking ability.[5]

Exercise assists in the maintenance of personal independence and may enhance it in older adults by increasing muscular strength and endurance.[13,14] Older adults are capable of significantly improving lower extremity strength and muscle power with resistance training.[15] Muscle power is important in preventing a fall: one must be able to get the stabilizing leg out fast enough to prevent or reduce the severity of the fall.[7]

We know that poor balance has been identified as a risk factor for falling,[4,16] and that balance relies on three main inputs: visual, vestibular, and somatosensory, all of which require certain baseline strength to be used to their potential. We also know that the “hip strategy” or fall-prevention strategy most widely used by the elderly requires coordinating and activating the muscles about the ankle, knee, and hip joints to resist destabilization as the abdominals are activated to flex the hip and move the centre of gravity backward.[17]

Objective

The main goal of this research was to determine if strength training can improve one’s hip strategy for regaining balance. Strength training in general has been shown to induce modest improvements in strength and profound effects in functional independence.[12,14,15] Improvements in self-confidence and coordination, overall health, quality of life, and general well-being have all been demonstrated.[18] This study sought to establish whether increasing the strength of muscles surrounding the hip girdle could contribute to improvements in static balance and well-being.

Method

Forty-four women aged 70 to 86 participated in the study. All subjects included in the study were considered untrained (defined as participating in exercise less than 2 hours per week) and did not meet any of the exclusion criteria. Each subject was informed of all procedures and signed a consent form approved by the human ethics committee of the University of British Columbia.

Subjects were placed into the control group or the training group, depending on their preference. The ability of some of the subjects to attend the training sessions at the university training facility was limited by a lengthy public transportation strike, which negated randomization. It should be noted that although subjects self-selected their group, each subject volunteered for the study initially with the intent and motivation to exercise in response to recruitment posters expressing the need for volunteers to participate in a training study. Each group was characteristically homogeneous in their motivation and desire to exercise. Subjects in the training group (n = 22) were assigned to 1-hour workouts consisting of strengthening and stretching exercises three times per week. Participants in the control group (n = 22) were asked to maintain their current level of activity.

Testing protocol

Testing of both training and control groups was carried out at 0 weeks and at 12 weeks. The training group was also tested at 6 weeks. Tests used included the physical activity scale for the elderly (PASE) questionnaire, a questionnaire about well-being, and three performance tests: timed two-legged squat (TLS), timed chair stand (TCS), and KAT balance system. Anonymity was maintained by assigning each subject a number for use on the questionnaires and assessment sheets.

The PASE questionnaire was devised to measure activity levels of adults aged 65 and older.[19] It was used in this study as a marker to explain any differences between the training and control groups and to discern any test score discrepancies between pretesting and posttesting in the control group.

The timed two-legged squat required the subject to perform a squat as low as possible without deviating from a plumb line while maintaining biomechanically efficient alignment of knees, nose, and toes. The subject had to maintain a starting position facing the wall with toes 10 cm from the wall, hands on shoulders, and arms abducted 90 degrees. The subject could not touch the wall with any part of the body and the head had to remain parallel to the floor. The test was considered complete when the subject deviated from the squat position.

The timed chair stand measured lower extremity function as total time in seconds to rise five times from an armless chair positioned 45 cm from the ground. Subjects could not use their arms for support in achieving a standing position. This protocol was repeated three times. Research has shown the chair stand to be a reliable physical performance measure.[20]

The KAT balance measure protocol[21] was used to test stability on a platform with adjustable settings measured in pounds per square inch (increased psi increases platform stability). Prior to the first test condition, stability was set to 6.0 psi to replicate standing on a hard surface. Subjects removed their hands from the handrail upon command and stood as quietly as possible and tried not to grasp the handrail. The measure of balance was recorded as the length of time the subject maintained platform stability within ± 0.2 for a maximum of 30 seconds. Each trial was terminated if the subject grasped the railing, could not maintain stability within ± 0.2 psi, completed the 30-second trial, or opened her eyes during the eyes-closed condition.

Each subject was given six trials of platform stability:

A Eyes open, 1.0 psi 
B Eyes open, 2.5 psi 
C Eyes open, 4.0 psi 
D Eyes closed, 1.0 psi 
E Eyes closed, 2.5 psi 
F Eyes closed, 4.0 psi

As a safety precaution, the eyes-open conditions were given before the eyes-closed conditions. A random numbers table was used to select platform stability settings for the eyes-open condition. The same order was used for the eyes-closed condition.

The questionnaire administered at the end of the program was designed to measure any change in well-being noted by the participants. Subjects were asked: “Since starting the program, has your sense of well-being changed? And if so, how?”

Training protocol

Training consisted of exercises associated with the hip girdle and lower body, started 1 week after preliminary testing and continued for 12 weeks. Training was three times per week for 1 hour, including warm-up (10 minutes), strengthening (35 minutes), and stretching (15 minutes). Muscles targeted were quadriceps, hamstrings, gluteus medius, tensor fasciae latae, gastrocnemius, rectus abdominus, and transversus abdominus. The exercises used were timed two-legged squats, leg press, leg curls, standing hip flexion, standing hip abduction, standing calf raises, controlled knee raises, and tubing abdominal curls. Four women from the training group dropped out of the study because of illness (n = 1), and soreness (n = 3). Two women from the control group dropped out because they did not wish to participate further.

Statistical analysis

SPSS 10.0 for Windows was used to analyze the data. All data were considered continuous. A two-way mixed ANOVA design with repeated measures (pretraining to posttraining) and between subjects (training versus control) factors analysis of variance was conducted to test differences between the groups in TLS, TCS, and timed KAT balance measures. As well, a one-way (3x1) repeated measures ANOVA was conducted to test differences at weeks 0, 6, and 12 within the training group for each test variable. A simple main-effects analysis (post hoc for the interaction) was executed to look at group differences at pretraining and posttraining times. Intraclass correlations and Pearson product moment correlations were used to conduct test-retest reliability between trials for the TLS and TCS, and t tests were used to analyze differences between PASE scores.

Results

There was no significant difference between groups for either mean age (control group 75.9 ± 3.7 years; training group 76.1 ± 4.8 years) or PASE scores (control group 127.1 ± 57.5; training group 108.8 ± 51.4). Table 1 shows the average weight lifted by training group subjects after weeks 1, 6, and 12, demonstrating indirectly an improvement in lower body strength.

Timed two-legged squat

Figure 1 shows a significant difference between the training group and the control group after 12 weeks of training (Interaction test, F1,36 = 20.22, P <.001). As a simple main-effects analysis (post hoc for the interaction) a t test was executed to look at group differences at pretraining and posttraining times (each t test at .025 significance level). The pretraining t test revealed no significant differences between mean TLS times for either group (P = .808). The posttraining t test revealed a significant difference in mean TLS times between the control and training group (P = .025). There was a significant difference (P <.001) over 12 weeks of training with the major significant change occurring within the first 6 weeks (t 18 = 2.61, P = .018) as seen in Table 2.

Timed chair stand

Figure 2 shows the differences between groups for average chair stand times at pretraining and posttraining testing. There was a significant difference between groups over the 12-week period (Interaction test, F 1,36 = 28.61, P < .001). A t test was conducted to confirm group differences at baseline and posttraining. The pretraining t test revealed no differences (P = .263). However, posttraining testing revealed significant differences (P = .015). Table 3 shows differences within the training group at 0, 6, and 12 weeks. The plot revealed significance in TCS means after 12 weeks of training, between 0 and 6 weeks (t18 = 5.34, P < .001) and between 6 and 12 weeks (t17 = 4.21, P =.001

Timed KAT balance measure

Table 4 shows the results of a within-group analysis conducted to see the effects of training on training group subjects at 0, 6, and 12 weeks. A separate repeated measures ANOVA was used for each of the six platform stability conditions, A to F. The results revealed no significance on conditions C, D, or E. However, condition B one-way ANOVA with linear trend, F 1,17 = 6.65, P = .019, and condition F, one-way ANOVA with linear trend, F 1,17 = 4.75, P = .044 both demonstrated significance. When KAT balance measure test results for both the training and control groups were analyzed, only condition F (eyes closed, psi = 4.0) demonstrated significance (Interaction test, F 1,36 = 5.422, P =.026), as shown in .

Self-reported well-being

The responses of all 39 women who finished the program and completed the well-being questionnaire are summarized in Table 5. The 18 “no” responses came from subjects in the control group. Two subjects in the control group provided “yes” responses based on their decision to undertake an exercise program. However, when their PASE questionnaires were analyzed, there was no significant difference between pretest and posttest reported activity. All other positive responses were from the training group.

PASE questionnaire

Statistical analysis of the PASE scores revealed no significant differences between the activity levels of the control group at pretest and posttest intervals. The questionnaire was used solely as a measuring tool to distinguish any changes in control group activity over the 12-week study period.

Discussion

Static squat times, lower body function, static balance at a platform stability level of 4.0 psi, and well-being all improved with the progressive resistance-training program used in this study. Most change occurred during the first 6 weeks, probably as a result of neuromuscular coordination and adaptations, which suggests that to induce strength changes from weeks 6 to12 a periodized exercise program of greater intensity (whereby the last two repetitions of each set are consistently difficult to perform) would be needed.

The finding from the TLS test that indicates neuromuscular changes and adaptations only occurred during the first 6 weeks of training is consistent with other research.[22] It is unknown how well the test differentiates between neuromuscular recruitment and strength gains. Further research is warranted to assess any possible correlations. However, the simplicity of the TLS test and the fact that it can be executed with a stopwatch, goniometer, and a wall makes it accessible and easy to administer. As well, the exercise can be used as a safe introduction to a training program and can eventually be augmented with concentric and eccentric squats. The TCS test is also easy to administer and provides a reliable measure of functional lower body strength and power, a finding consistent with other studies.[20]

Of the three performance tests used, the KAT balance measure provided the results that raise the most questions. The only between-group difference noted was for condition F (eyes closed, psi = 4.0). This is important as condition F lacked visual input, and logic suggests that only one of two inputs could have improved. Although no aspect of the resistance training focused solely on vestibular functioning, it is possible that the platform stability trials could have led to improvements in maintaining balance. More plausibly, the progressive resistance-training program may have improved the somatosensory functioning as well as any related processes, thus allowing only those subjects in the training group to exhibit improved time to balance. But why, then, was there a significant increase only between groups for condition F? Perhaps for the eyes-closed conditions, the platform stability for conditions D and E was too difficult to see any significant changes. For the eyes-open conditions, condition C was considered easy by the subjects and aptly demonstrated ceiling effects. Results for condition A showed minor improvements in mean balance times, and thus changes to both groups may be attributed to the visual input or learning effects on the test. Overall limitations to the KAT balance measure include its limited use in research and the need for additional reliability and validity trials to control for extrinsic or intrinsic variables, such as weather and mood, that may affect balance in older adult women.

The results of the questionnaire used to assess well-being are consistent with research reporting improved quality of life after undertaking a training program.[23] When the questionnaire responses were categorized and quantified, the responses indicated improved health, less aching, and improved function. The open-ended nature of the questions allowed subjects to recall any apparent changes in the 12 weeks of training and to express what was most important to them during training. Although recall ability might be considered a limiting factor, no subject appeared to have difficulty with recall. Of concern are possible confounding variables for well-being, such as the use of healthy older adults rather than frail older adults. Outside events may have had an impact on the balance and well-being outcomes, and improvements in mood may have been the result of the increased social interaction with other subjects or the trainer. However, because the responses of the control group were so consistently unlike the responses of the training group, it is feasible to conclude that 12 weeks of training can have a significant impact on well-being.

Conclusion

The timed chair stand is a reliable measure of lower body strength, while the timed two-legged squat is a reliable measure of muscular recruitment. Gains in muscular recruitment and lower body strength and function in 12 weeks may improve balance time in certain conditions and may have a profound positive effect on well-being in women aged 70 to 86.

Future research could overcome the limitations of this study by using a larger sample size and focusing on more functional testing and testing dynamic balance. Follow-up studies might also examine fall incidence (males, the frail elderly), response to exercise in other populations and other methods for improving balance that involve removing visual stimuli or training in a prone rather than vertical position. Future studies should maintain a focus on improving function and well-being in older adults.

Competing interests
None declared.

 Table 1. Average weight (in pounds) lifted by training group subjects at weeks 1, 6, and 12.

Values shown are means (± SD)
* Keiser Equipment Systems (Fresno, CA)
† All Pro Ankle Weights (Carpinteria, CA)

 Table 2. Group means for average timed two-legged squat (TLS).

Values shown are means (± SD)
* Significant difference compared with week 0 (P < .001)

 Table 3. Group means for average timed chair stand (TCS).

Values shown are means (± SD)
* Significant difference compared with week 0 (P <.001)
† Significant difference compared with week 6 (P = .001).

 Table 4. KAT balance measure means for conditions A to F within the training group (n = 18).

Values are means (± SD)
* Significant difference compared with week 0 (P = .023).
† Significant difference compared with week 6 (P = .044).
‡ Significant difference compared with week 0 (P = .026).

 Table 5. Responses to well-being questionnaire, by category.

J. Rezmovitz, MSc, J.E. Taunton, MD, E. Rhodes, PhD, A. Martin, PhD, B. Zumbo, PhD

Mr Rezmovitz is a master’s graduate from the School of Human Kinetics, UBC, and is now a research assistant in medical oncology at Sunnybrook Hospital, Toronto. Dr Taunton is a primary care sports medicine physician and director of the Allan McGavin Sports Medicine Centre, as well as professor in the Department of Family Practice and School of Human Kinetics, UBC. Dr Rhodes is a professor in the Faculty of Education and School of Human Kinetics at UBC. Dr Martin is a retired professor in the Faculty of Education and School of Human Kinetics at UBC. Dr Zumbo is a professor in the department of Measurement Evaluation and Methodolgy Research in the Faculty of Education, UBC.