Introduction

One out of three adults over 65 years of age living in the community experiences at least one fall each year [1]. Falls are the leading cause of fatal and nonfatal injuries and financial burden among Americans aged 65 and older [2]. Hospitalization from fall-related injuries accounts for 77% of all hospitalizations from injuries [3]. In the United States, the direct medical costs due to fall injuries were $31 billion in 2015 [4].

Successful balance control depends on the integration of three systems: the sensory system which collects information about the position of the body in space, the central nervous system that processes the sensory information, and the motor system that sends signals for appropriate movement execution. With aging, accumulated deficits in these systems contribute to falls [5]. Gait and balance impairment [6,7,8] and decreased muscle strength, especially in the lower extremities, increase the risk for falls [9, 10]. Furthermore, decreased movement speed and decreased force control with aging make it difficult to recover balance after tripping [11]. Given the potential negative physical and financial consequences of falls, it is important to provide an intervention to decrease the likelihood of falling before the balance impairments reach a critical level and falling becomes inevitable [12].

Intervention programs to reduce fall risks are divided into either providing single interventions (e.g., exercise) or providing multiple modes of interventions (e.g., exercise and education). The multi-modes interventions are further divided into individually tailored multifactorial interventions and group tailored multicomponent interventions [13]. An umbrella review of meta analyses by Stubbs et al. [14] concluded that single intervention and individually tailored multifactorial intervention programs were effective in reducing both the rate of falls and the number of individuals classified as fallers with varying levels of balance impairment. Because multiple factors contribute to falls, the American Geriatrics Society and British Geriatrics Society recommend either multifactorial or multicomponent interventions [15]. Noticing that the effectiveness of multicomponent interventions has not been extensively reviewed, Goodwin et al. [16] performed a meta analysis of 15 multicomponent intervention studies and concluded that multicomponent interventions are effective in reducing rate of falls and the number of people who fall. Fall prevention efforts, however, often target people who are already at high risk of falling (e.g., nursing home residents, older adults with musculo-skeletal or neurological impairments) [16]. Only 5 of the 15 multicomponent intervention studies reviewed by Goodwin et al. [16] involved older adults who were not considered to be at high risk for falls. Furthermore, the American Geriatrics Society and British Geriatrics Society recommend community-dwelling older people who have a past history of two or more falls, who have experienced a recent acute fall, or who have difficulty with walking and balance should be considered for participation in an intervention to reduce their fall risks [15]. The same guideline suggests that older adults who do not report falling in the past 12 months, need only be periodically reassessed. However, some studies have emphasized the importance of early intervention for maintenance of unimpaired balance [17,18,19]. For example, Muir et al. [20] found that a significant percentage of community-dwelling older adults who did not report recent falls experienced falls during the next 12 months. They concluded that balance impairment (defined as obtaining a Berg Balance Scale score of less than 50 out of 54) was a contributing factor for future falls. They also found that, for relatively high-functioning older adults who did not report an incident of falling, lower extremity weakness and impaired balance each uniquely contributed to an individual’s moving from being classified as a non-faller to being classified as a faller. Thus, it is important to study the effectiveness of a multicomponent intervention program focused on balance and muscle strength for independently functioning community-dwelling older adults.

The current study investigated the effectiveness of an 8-week multicomponent balance training program on reducing fall risks by improving balance and muscle strength in independently functioning, community-dwelling older adults. The Fullerton Advanced Balance (FAB) scores at baseline were used to classify the participants as either with high fall risks or with low fall risks. We hypothesized that the multicomponent balance training program would (a) improve balance and muscle strength and (b) individuals classified as having high fall risks would experience more improvement than individuals classified as having low fall risks, resulting in a significant number of older adults with high fall risks at pre-testing being classified as having low fall risks at post-testing.

Methods

Participants

Fifty-three older adults (9 males and 44 females) participated in the study through the “Balance and Fall Prevention” class offered at a large urban University and a local senior housing community (Fig. 1). Participation in the Program was advertised through the quarterly newsletter of the university-associated Lifelong Learning Institute. Newsletters were distributed to 1600 members and placed in local senior centers and libraries. Participation in the research arm of the class was voluntary. Included participants were 65 years of age or older, able to stand unassisted for 45 min, able to follow verbal commands, independent in feeding, bathing, dressing, toileting, and transfers. Participants were excluded if they reported any medical condition that affected their cognition, balance, or walking. Six out of fifty-three participants were eliminated from the analysis: Five participants missed 3 or more classes (2 due to illness and 3 due to lost interest), and one participant missed post-testing due to traveling (Fig. 1). The age of the 47 participants who completed the intervention ranged from 65 to 91 years of age (M = 80.09, SD = 6.62). The majority of the participants were Caucasian. The Institutional Review Board of the University approved the study, and all participants signed an informed consent prior to joining the study.

Fig. 1
figure 1

Participants’ flow chart

Full size image

Procedures

During pre-testing, participants completed a health, activity, and demographic questionnaire providing information that was used to confirm eligibility of the subjects to participate in the study. In addition, we obtained blood pressure, dichotomous fall history (whether or not the participants fell in the past 12 months), exercise frequency (how often they exercised per week), and fall worry score (using a scale ranging from 1 = not worried to 7 = extremely worried). The following three outcome measures were recorded at  pre- and post-intervention.

Fullerton Advanced Balance (FAB) scale

The FAB scale was used to assess participants’ balance. The score at pre-testing determined each participant’s designation as an older adult with high fall risks or an older adult with low fall risks [21]. FAB is known to be appropriate for assessing balance of independently functioning community-dwelling older adults, and is considered a useful tool for identifying subtle changes in various components affecting balance abilities [22]. The scale consists of ten items assessing both static and dynamic balance. Each item is scored using a 5-point scale (0 = not able to perform/ needs assistance, 4 = able to perform independently) with a maximum score of 40 points. FAB scores are cross-sectionally associated with faller status as determined from a self-reported fall history such that older adults with FAB scores of 25 or lower are more likely to have a history of two or more falls in the past 6 months [23] or in the past 12 months [22]. A cut-off score of 25 was found to provide the sensitivity of 74.6% and specificity of 52.6% in distinguishing people without a history of falls in the past 12 months from people with a history of two or more falls in the past 12 months among independently functioning older adults [22]. Another study found that a cut-off score of 22 produced the sensitivity of 85% and specificity of 65% in predicting faller status [23]. A recent prospective study found that FAB scores have a sensitivity of 68% in predicting whether people with idiopathic Parkinson Disease will have one or more falls over the next 6 months [24]. The FAB scale has good test–retest reliability (r = 0.96), inter-rater reliability (r = 0.94–0.97), and convergent validity established using the Berg Balance Scale (r = 0.75) [21].

Senior fitness measures

Physical strength was assessed using three measures of the Senior Fitness Test [25]: the number of arm curls in 30 s (arm curl), the number of chair stands in 30 s (chair stand), and the number of in-place steps in 2 min (2-min step). The arm curls and chair stands assess general upper and lower extremity strength and the 2-min step assesses endurance. Although our program does not include an endurance intervention, we included the 2-min step test as an outcome measure to gauge the extent of indirect improvement in endurance that might have occurred as a result of improvement in balance and muscle strength. All measures were assessed once at pre- and post-test. Test–retest reliability for SFT items ranged from 0.80 to 0.98; content validity was established through the literature review and expert opinion; criterion validity ranged from 0.73 to 0.81; and construct validity was demonstrated by superior performance of older adults who were regular exercisers compared to those who were non-exercisers [25].

Intervention

The intervention was a 60-min group exercise class, conducted at an on-campus fitness center twice a week for 8 weeks and one educational session on fall prevention presented halfway through the intervention. A certified balance instructor conducted the exercise class. A licensed physical therapist designed the exercises and supervised the class. Graduate physical therapy students ensured that each participant correctly performed the exercises at a challenging level for that participant. Each student supervised two participants.

Weekly progression of the exercise intervention is presented in Appendix. The warm-up for the class included postural awareness, trunk stabilization, and variable walking activities with arm movement. Participants changed walking direction and speed, made abrupt stops, and performed head turns in response to music. Balance and multisensory training, designed to be progressively more difficult over time, included dynamic weight shifts, toe tapping, tandem walking, and walking in different directions. Sensory challenges during all exercises included eyes open and eyes closed, standing and walking on foam and head movement, as well as reading while walking.

Upper and lower extremity strength exercises were performed from both seated and standing positions using elastic resistance as well as standing against the wall for closed chain exercises. Some participants were able to do the seated exercise sitting on Dyna discs or an exercise ball. The strength of elastic resistance was set for each individual to complete ten repetitions. When a participant was able to complete more than ten repetitions, the participant was given the stronger resistance band. At week 5, exercise intensities were increased from 2 to 3 sets for each exercise.

The class also introduced an activities and games section that included crossing an obstacle course, passing objects in different directions, and kicking balls or balloons in circles or in lines. Head and eye movement training at different speeds and directions was done in the seated positions using popsicle sticks as a focal point.

The education, presented on the 4th week of the program, included instructions on home modifications for safety, proper footwear, the importance of addressing vision problems, proper use of assistive devices, management of medications, and practice on how to get up from the floor after a fall. Participants also received individualized consultation on home safety in response to a completed home safety checklist.

Power analysis

We conducted a priori power analysis using G*POWER [26]. The result showed that the desired total sample size is 34 for an ANOVA with two repeated measures (pre and post) and two groups with a medium effect size (i.e., F = 0.25), α = 0.05, power = 0.8, and an average correlation for repeated measure = 0.5.

Statistical analysis

Statistical analyses were conducted using IBM SPSS (Armonk, NY) version 23. Fisher’s exact test was applied to examine participants’ gender distribution, and a MANOVA and follow-up univariate ANOVA analyses were used to examine other demographic and activity characteristics as a function of fall risk status at pre-testing. An exact McNemar’s test was applied to compare the frequencies of people who classified with low fall risk and high fall risks at pre- and post-testing. A mixed ANOVA with time (pre vs. post) × group (low fall risks vs. high fall risks) was conducted to investigate the effects of the intervention on the FAB scores. Time was a within-subject variable and group was a between-subject variable. A significant time effect was followed with a paired t test for each group. A parallel-mixed MANOVA was conducted to investigate the effects of the intervention on three measures of the SFT. A significant result was followed with a univariate ANOVA on each measure of the SFT. The statistical significance was set at P < 0.05 in all analyses.

Results

Balance

Participants were classified based on the FAB scores. People whose FAB score was equal to or less than 25 were classified as having high fall risks and those who scored above 25 were classified as having low fall risks [22]. There were 28 older adults with high fall risks (MFAB = 20.29, SDFAB = 5.09) and 19 older adults with low fall risks (MFAB = 30, SDFAB = 3.18) at pre-testing. Table 1 shows demographic characteristics of participants at pre-testing as a function of fall risk status. The results of a MANOVA at pre-testing revealed that there was no significant difference between people with high fall risks and low fall risks in any demographic and activity variables except for age, where people with high fall risks tended to be older than people with low fall risks, as expected. Although there were more females than males in general, the result of the Fisher’s exact test showed that males and females were similarly distributed between high and low fall risk groups.

Table 1 Demographic and activity characteristics of participants (n = 47) at pre-testing
Full size table

The means and standard deviations of the balance test scores are provided in Table 2. As predicted, the results of a mixed ANOVA showed a significant effect of time, F1,45 = 128.25, ɳ2 = 0.74, P < 0.001, such that there was an improvement in the FAB scores after the intervention for the combined sample. There was also a significant interaction of time and group, F1,45 = 6.15, ɳ2 = 0.12, P = 0.02. A follow-up paired sample t test for each group demonstrated that both groups showed a significant improvement in balance from pre-testing to post-testing, but people who classified as having high fall risks showed more improvement, t27 = 9.39, P < 0.001, than people who classified as having low fall risks, t18 = 8.15, P < 0.001.

Table 2 Means (and standard deviations) and paired t tests of Fullerton Advance Balance scores and Univariate Analyses of Senior Fitness Test scores at pre- and post-testing for independently functioning community-dwelling older adults who were classified as low (n = 19) and high (n = 28) fall risks at pre-testing
Full size table

When the participants were re-classified according to their fall risk status after the intervention, significantly more people were classified as having low fall risks than having high fall risks (Table 3) with 41 people with low fall risks (MFAB = 32.27, SDFAB = 3.44) and only six people with high fall risks (MFAB = 18.67, SDFAB = 4.50). Twenty-two older adults with high fall risks at pre-testing were classified as having low fall risks at post-testing, whereas no one who classified as having low fall risks at pre-testing was classified as having high fall risks at post-testing. An exact McNemar’s test showed that there was a significantly higher proportion of people being classified as having low fall risks after the intervention compared to the baseline proportion, P < 0.001, showing the effectiveness of the intervention in improving balance.

Table 3 Number of older adults with low and high fall risks at pre- and post-testing
Full size table

Senior fitness measures

The means and standard deviations of the senior fitness test scores are also provided in Table 2. As predicted, the results of mixed MANOVA showed a significant main effect of time, Wilks λ = 0.39, F3,43 = 22.77, ɳ2 = 0.61, P < 0.001, such that there was an improvement in the overall senior fitness test score after the intervention for the combined sample. Univariate analyses showed that the improvement occurred on all three measures of arm curl, chair stand, and 2-min step (P < 0.001 for all). The interaction of time and group was not significant (P = 0.43).

Discussion

This study investigated the effects of a multicomponent intervention program conducted two times per week for 8 weeks on reducing fall risks of independently functioning community-dwelling older adults. The intervention included a progressively difficult exercise component as well as an education component provided at the midway of the intervention. Based on the FAB scores at pre-testing, participants were classified as either having low fall risks or having high fall risks. Confirming our hypothesis, the intervention improved balance, and the participants in the high fall risk category improved more than those in the low fall risk category. Our hypothesis on muscle strength was partially confirmed as the intervention improved muscle strength, but both groups improved equally.

This study aimed at evaluating a balance intervention program for relatively healthy older adults using a sensitive measurement tool for assessing balance ability. Although research on the identification of fall risks often focused on older adults with substantial impairment in balance [27], past studies have also demonstrated the prevalence of future falls among high-functioning older adults. For example, Muir et al. [8] found that 43% of the high-functioning community-dwelling older people with no history of recent falls experienced falls within the next 12 months. Another study found that about half of healthy older women with no past history of falls experienced falls within the next 12 months, and about a quarter of all the participants fell more than once [18]. Similar results were obtained in healthy older men as well [20]. These results underscore the importance of providing a fall risk reduction program for older adults who may not be considered at high risk for falls.

Past studies often determined older adults’ fall risk based on their fall history during the past 12 months [15]. Because the causes of falls are multifactorial, the past history of falls alone may not accurately reflect the degree of impairment in balance [12]. Given that different tools for assessing balance impairment lead to different results in defining who would be at risk of falling [8, 22], it is important to use the tools that are designed to capture the balance ability of healthy older adults. The FAB scale used in this study includes test items that are more challenging than other balance scales (e.g., Berg Balance Scale) and is useful in detecting the subtle changes in balance abilities of functionally independent older adults [22]. Our intervention improved older adults’ FAB scores. Although values for minimal detectable change (MDC) or minimal clinically important difference (MCID) for the FAB scores are not available, it was found that every score increase in FAB is associated with an 8% decrease in the probability of falling [22]. Thus, the average gain of 6 points in the FAB scores in this study implies a significant reduction in the likelihood of falls as a result of the intervention.

This study also found that a multicomponent intervention program was effective in improving older adults’ upper and lower body muscle strength and endurance. Studies have shown that a decrease in muscle strength is a risk factor for falls [9, 10], and functional lower extremity weakness and impaired balance independently contribute to moving from non-faller status to faller status [20]. Group exercise programs have been found to be effective in increasing muscle strength [9]. For example, a 12 week regimen of strength training produced a significant improvement in lower limb strength and balance [28]. The results of this study on muscle strength are consistent with those of past studies and emphasize the importance of strength training in decreasing the risk of future falls. While the MDC or MCID values for the SFT performance with healthy older adults are not available, a study found that the MDC values at the 90% confidence intervals in the community-dwelling older adults with cognitive impairment were 2.3 for the arm curl test and 2.0 for the chair stand test [29]. The current study with healthy older adults found improvements of 2.75 for the arm curl test and 2.11 for the chair stand test. Because the variabilities in performance in these two tests were smaller in our data from healthy older adults than the data from the people with cognitive impairment, we can conclude that the changes in this study exceeded the MDC values, demonstrating clinical importance of our intervention program. Even though our intervention program did not emphasize endurance training, the improvement in the 2-min step test is not surprising considering the variety of walking activities, as well as lower extremity resistance training employed in the exercise intervention.

We also hypothesized that compared to people who were classified as having low fall risks, people who were classified as having high fall risks would show greater improvements after the intervention, because people with low fall risks would have less room for improvement. Our results, however, supported this hypothesis only for balance, but not for muscle strength. Both groups equally improved in muscle strength. This might be related to the novelty of strength training in both groups. These results underscore the importance of an intervention targeting independently functioning older adults. The older adults who did not seem to have an immediate danger of falling still had ample room for further improvement in overall balance. Given that fall-related injuries can lead to admission to a long-term care facility that could cause additional financial burden as well as negative psychological consequence such as depression [30, 31], it is important to provide an early intervention that will allow healthy older adults to continue their independent living without experiencing falls.

This study has some limitations. One limitation is that this study did not have a control group that did not receive intervention and where participants were randomly assigned to either a control group or an intervention group. Instead, we classified older adults as either having low fall risks or having high fall risks using the FAB cut-off score that was determined based on older adults’ self-report of past history of falls, not based on fall data obtained through a prospective study [22]. It would be desirable for future studies to employ a randomized controlled trial and also utilize measures for group assignment that have been shown to be predictive of future falling among community-dwelling older adults. In addition, a more comprehensive assessment of fall likelihood could provide an improved classification of balance ability [8]. For example, in studying the effectiveness of a physical therapist-prescribed home exercise intervention on balance, one study applied several balance assessment measures (e.g., the functional reach test, the step test, walking speed, the step quick turn test) to identify older adults with mild balance dysfunction [12]. Another limitation is that this study only measured the clinical outcomes of balance and senior fitness. Other physical measures such as gait characteristics [32], the occurrence of future falls [32], and psychological measures such as fall efficacy [33] would have provided a deeper understanding of the effects of an intervention on reducing fall risks. In addition, the outcome measures were assessed only just before and after the intervention. Follow-up testing would be useful to determine if the participants maintained these improvements. Finally, participants were volunteers from two large senior communities, but the sample size was relatively small (n = 47). Despite the small sample size, after the participants were separated into having high fall risks and having low fall risks, there was enough power to detect a significant interaction effect of time and group on balance. Also, there were more females than males, consistent with literature on fall prevention [1] and with the population of older adults from which the sample was derived (68% female, 32% male). Future studies could employ a similar proportion of males and females to investigate potential sex-related effects in fall prevention. This study investigated the effects of an intervention on independently functioning community-dwelling older adults. Given that independently functioning older adults are not necessarily high-functioning older adults [1], future studies could benefit from investigating the effects of interventions aimed at reducing fall risks specifically for high-functioning older adults. In addition, our multicomponent intervention consisted of static and dynamic balance exercises, strengthening exercises, sensory integration (visual, proprioception, and vestibular), and education. Future studies could incorporate other modes of intervention such as vitamin D supplementation or cataract surgery [1] to investigate the effectiveness of a comprehensive multicomponent intervention on the reduction of fall risks.

In conclusion, this study demonstrated the effectiveness of an intervention targeting independently functioning older adults, a group often not a primary target of fall prevention efforts. The intervention improved balance and muscle strength in all participants. People who were classified as having high fall risks based on the FAB scores at pre-testing showed greater improvement in balance compared to those who were classified as having low fall risks.