Daily Use of Bilateral Custom-Made Ankle-Foot Orthoses for Fall Prevention in Older Adults: A Randomized Controlled Trial

Abstract

OBJECTIVE:

To examine the effects of bilateral custom-made ankle-foot orthoses (AFOs) to prevent falls for older adults with concern about or at risk for falling over 12-month daily use.

DESIGN:

Secondary analysis of a randomized controlled trial.

SETTING:

Community-dwelling older adults.

INTERVENTION:

Half of the participants were randomly allocated to an intervention group (IG) that received fitted walking shoes and bilateral custom-made AFOs, and the other half were randomly allocated to a control group (CG) that only received fitted walking shoes.

MEASUREMENTS:

Self-reported fall history of 12-month duration was investigated at baseline and 12-month follow-up for both groups. Fall incidence rate and proportion of fallers were used as outcome measures to determine effects of 12-month footwear intervention in either group.

PARTICIPANTS:

Adults aged 65 years and older with concern about or at risk for falling (n = 44).

RESULTS:

No significant between-group differences in participant characteristics were observed at the baseline (P = .144−.882). Within the IG, significant reductions were found in the fall incidence rate (P = .039) and the proportion of fallers (P = .036) at the 12-month follow-up compared to the baseline. Within the CG, no significant change was found at the 12-month follow-up compared to the baseline for the fall incidence rate (P = .217) or the proportion of fallers (P = .757). When comparing the IG with the CG, there was no significant difference in the change from the baseline to the 12-month follow-up for the fall incidence rate (P = .572) or the proportion of fallers (P = .080).

CONCLUSION:

This study failed to demonstrate a significant benefit of bilateral custom-made AFOs to reduce falls compared to fitted walking shoes. However, the AFO users had significant reductions in falls compared to the preceding year. A future study with a larger sample size is recommended to confirm these observations.

Age-related changes in foot and ankle characteristics may negatively influence balance, walking ability, and fall risk.¹ Footwear intervention is often recommended as a non-surgical solution to reduce the risk of falling because of lower-extremity problems among older adults.² Ankle-foot orthoses (AFOs) are a footwear intervention to enhance ankle stability by locking the ankle joint at a right angle during walking and other daily activities. AFOs are often clinically prescribed to patient populations with lower-extremity illness for controlling their foot drop and reducing their risk of falling.^3,4 In the hemiparetic stroke patient population, AFOs have been shown to enhance balance,^5–7 enhance walking ability,^6–8 reduce fear of falling,⁹ enhance gait kinematics and kinetics,^10,11 and reduce the energy cost of walking.^12,13 Similarly, positive effects of AFOs on gait biomechanics,^14,15 walking ability,^16,17and energy cost of walking^17,18 were found in hemiplegic cerebral palsy patients. However, few studies reported potential negative effects of AFOs, such as muscle disuse atrophy in hemiparetic stroke patients^19,20 and alteration in dynamic balance in patients with multiple sclerosis.²¹

Given dominant evidence showing positive effects of AFOs in the above patient populations, it is likely that the daily use of AFOs could also prevent falls for community-dwelling older adults, who have concern about or are at risk for falling but are not necessarily diagnosed with any specific condition affecting the lower extremities (eg, stroke, hemiplegia, or foot deformity). Yalla et al²² reported that bilateral custom-made AFOs stabilized the ankle joint and reduced body sway in the mediolateral direction while standing still among older adults. However, the main limitation of the study of Yalla et al²² was its single-arm cross-sectional trial design, which could not reveal the longitudinal effects of daily use of AFOs on the stabilization of the ankle and balance.

To overcome this limitation, we conducted a longitudinal randomized controlled trial (RCT) to investigate the effectiveness of daily use of bilateral custom-made AFOs on fall prevention. Our previous article²³ demonstrated that 6-month daily use of bilateral custom-made AFOs could improve balance in community-dwelling older adults with concern about or at risk for falling, but that article only reported the parameters related to fall risk. It was still unclear whether the daily use of AFOs could directly reduce prospective falls. Therefore, in this article, we did a secondary analysis to examine the effectiveness of AFOs on fall prevention over 12-month daily use in community-dwelling older adults with concern about or at risk for falling. We hypothesized that the 12-month daily use of bilateral custom-made AFOs with fitted walking shoes would reduce falls compared to fitted walking shoes alone and compared to the year before the start of the intervention.

METHODS

Participants

Details of the study design have been described in our previous study.²³ In short, we recruited 44 older adults from outpatient clinics or education centers for seniors/older adults at the Houston metropolitan area, TX. Inclusion criteria included being ambulatory and aged 65 years or older, with self-reported concern about falling or at risk for falling (confirmed by either 13 seconds or more in the Timed Up and Go [TUG] Test²⁴ or a fall in the past 6 months²⁵). Participants were excluded if they: (i) had a wound on either the feet or the ankles; (ii) were taking any medication that might have temporary (less than 1 month) or unstable (fluctuate over time) impact on gait and balance, according to the judgment of the clinical investigator; (iii) had a major foot amputation or any lower-extremity fracture; (iv) had unstable medical conditions affecting gait and balance, according to the judgment of the clinical investigator; (v) were participating in any other interventional study within the past 30 days or planning to participate in any other interventional study during the duration of our study; (vi) were unable to stand without help, were nonambulatory, or were unable to walk a distance of at least 1.8 m (approximately 6 ft) without assistance; (vii) had cognitive impairment (having a score of less than 24 on the Mini-Mental State Examination); or (viii) were unable or unwilling to participate in all procedures and follow-up evaluations (eg, long travel distance, such as living greater than 24 km away from the site of assessment). All participants signed a written consent form approved by the Institutional Review Board of the Baylor College of Medicine (Houston, TX; number H-38050) and were compensated $20 per session for their time in baseline and follow-up sessions. The clinical trial was registered (Clinical Tnal Registration—URL: http://www.clinical-trials.gov. Unique identifier: ).

Study Design

Figure 1 illustrates the consort diagram for the study. An RCT with two arms, a control group (CG) and an intervention group (IG), was conducted to evaluate effects of daily use of bilateral custom-made AFOs on reducing falls at 12 months compared to the preceding year. Eligible participants were allocated into one of these two groups using a random number sequence generated using MATLAB, version 2016a (Mathworks) before the start of participant recruitment.

Figure 1.

Consolidated Standards of Reporting Trials diagram for the study. Abbreviations: AFO, ankle-foot orthosis; CG, control group; IG, intervention group. [Color figure can be viewed at http://wileyonlinelibrary.com]

During the first visit, we assessed the eligibility of each potential participant using the inclusion and exclusion criteria mentioned above. If eligible, a foot measuring device (RALYN Shoe Care) was used to measure shoe size for ordering fitted walking shoes (MW813; New Balance) for every participant. The feet of each participant in the IG were cast using a contoured footboard, provided by the company that made the bilateral custom-made AFO (Moore Balance Brace; Arizona AFO). Participants from both groups reported their demographics (age, sex, height, and weight), provided their health information (medications and use of walking assistive devices), and completed the following questionnaires: Visual Analog Scale of foot pain,²⁶ Fall Efficacy Scale-International (FES-I; where an FES-I score of 23 or greater represents a high concern of falls),²⁷ and Fried Frailty Phenotype.²⁸ To measure the plantar sensation of all participants’ feet, we used a vibration perception threshold test with a biothesiometer (Xilas Medical).²⁹ The Alternate-Step Test³⁰ and TUG Test²⁴ were conducted to assess dynamic stability and mobility performance, respectively. Additionally, we asked all participants how many falls occurred in the 12 months before the first visit. Falls were defined as “an unexpected event in which the participant came to rest on the ground, floor, or lower level.”³¹ All data collected at the first visit were considered as baseline.

The footwear fitting with the prescribed footwear was scheduled at the second visit (1 month after the first visit) for all participants. Participants in the CG received the fitted walking shoes, and participants in the IG received the same brand of walking shoes as the CG and also received bilateral custom-made AFOs. Participants in both groups were encouraged to wear the prescribed footwear (AFOs and fitted walking shoes for the IG and fitted walking shoes only for the CG) at all times, in particular when standing or walking. We conducted follow-up investigations at 3, 6, and 12 months after the second visit (the start of the footwear intervention), at which all participants were asked to self-report the number of falls that occurred since their last visit. The count of falls at each of the follow-up visits was added together at the 12-month follow-up, to get an estimate of 12-month fall history while wearing the prescribed footwear.

The outcome measures used in this article are fall incidence rate (the number of falls per year), proportion of fallers (those who fell on at least one occasion), and proportion of multiple fallers (those who fell on two or more occasions) at the baseline and the 12-month follow-up.

Statistical Analyses

Baseline group differences were compared using one-way analysis of variance for continuous variables that were normally distributed or using the Mann-Whitney U test if they were not normally distributed. Normality was assessed using Shapiro-Wilk tests (P > .05). For categorical variables, we used the χ² test to compare baseline group differences. For between- and within-group comparisons in the outcome measures, we used generalized estimating equations (GEEs). The GEE is a repeated-measurement analysis recommended for nonparametric variables.³² In the context of this article, GEEs were used to analyze effects of Group (two levels: IG and CG), Time (two levels: baseline and 12-month follow-up), and Group × Time interaction on the three outcome measures under the intent-to-treat principle. The Group effect modeled between-group difference (IG compared with CG) for the outcome measures at each time point (baseline or 12-month follow-up). The Time effect modeled within-group change (12-month follow-up compared with baseline) in each group (IG or CG). The Group × Time interaction effect modeled between-group difference (IG compared with CG) for change in the outcome measures from the baseline to the 12-month follow-up. All statistical analyses were performed using IBM SPSS Statistics, version 24 (IBM). For all statistical analyses, significance was accepted at P < .05.

RESULTS

The baseline characteristics of our participants are provided in Table 1. Twenty-two participants were allocated to the IG (age = 73.7 ± 6.3 years; body mass index [BMI] =27.8 ± 4.8 kg/m²; sex = 63.6% female), and 22 were allocated to the CG (age = 75.6 ± 6.5 years; BMI = 29.3 ± 6.4 kg/m²; sex = 77.3% female). No significant difference was observed between the CG and the IG, for any of the characteristics at baseline (P > .050). Twelve participants (approximately 27%) dropped out prior to the 12-month follow-up (CG, n = 1 withdrew due to back pain unrelated to our study, n = 2 loss of contact, n = 1 deceased, n = 1 loss of ambulatory ability unrelated with our study; IG, n = 2 withdrew due to relocation, n = 2 loss of contact, n = 1 loss of ambulatory ability unrelated with our study, n = 2 deceased). None of the participants received any exercise, balance training, or physical therapy that could influence fall-related balance and gait performance during the trial.

Table 1.

Baseline Characteristics of the Participants

Variables	Control Group (n = 22)	Intervention Group (n = 22)	P Value
Demographic characteristics
Age, y	75.6 ± 6.5	73.7 ± 6.3	.328
Female sex	17 (77.3)	14 (63.6)	.322
Body mass index, kg/m2	29.3 ± 6.4	27.8 ± 4.8	.453
Health information
Fall incidence rate during the past 12 mo	1.82 ± 4.19	2.27 ± 4.17	.273
Fear of falling (FES-I score)	33.3 ± 11.5	32.8 ± 10.7	.882
High concern for fall (FES-I score ≥23)	20 (90.9)	17 (77.3)	.216
Frailty phenotypes			.627
Frail	2 (9.5)	3 (14.3)
Prefrail	15 (71.4)	12 (57.1)
Nonfrail	4 (19.0)	6 (28.6)
Depression (CES-D score)	8.0 ± 6.7	8.8 ± 7.2	.655
Depressed (CES-D score ≥16)	3 (13.6)	4 (18.2)	.680
Use of assistive device	9 (40.9)	12 (54.5)	.365
Medications per day
Prescription medications	4.7 ± 4.0	7.5 ± 5.6	.144
Over-the-counter medications	2.9 ± 2.5	5.0 ± 7.2	.673
Foot pain (VAS), 0–10 scale	1.2 ± 2.8	1.5 ± 2.2	.260
Plantar sensation (VPT), volts	25.2 ± 13.4	18.5 ± 12.7	.214
Performance-based tests
Timed Up and Go Test score, s	11.7 ± 4.4	12.1 ± 3.8	.449
Alternate-Step Test score, s	12.32 ± 4.18	10.66 ± 2.80	.155

Note. Values are presented as mean ± SD or number (percentage). P values are given for differences between the intervention group and the control group. Abbreviations: CES-D, Center for Epidemiological Studies-Depression; FES-I, Fall Efficacy Scale-International; VAS, Visual Analog Scale; VPT, vibration perception threshold.

Table 2 shows the outcome measures across time for each group and Time effects within each group. Table 3 summarizes statistical analyses for Group and Group × Time interaction effects for our outcome measures. Figure 2 illustrates the fall incidence rate across the baseline and the 12-month follow-up for both groups. At the baseline, there was no significant difference between the CG and the IG (ie, nonsignificant Group effect) for any of the outcome measures (P > .050). Within the IG, significant reductions were found in the fall incidence rate (65%; P = .039) and the proportion of fallers (33%; P = .036) at the 12-month follow-up compared to the baseline (ie, significant Time effect), while no significant reduction was found in the proportion of multiple fallers (21%; P = .152). Within the CG, while trends of improvement, in particular, for the fall incident rate (48%) and the proportion of multiple fallers (19%) were observed, neither achieved statistical significance in our sample (P > .050). At the 12-month follow-up, there was no significant difference between the CG and the IG (ie, nonsignificant Group effect) for any of the outcome measures (P > .050). The change from the baseline to the 12-month follow-up was not significantly different between groups (ie, no significant Group × Time interaction) for any of the outcome measures (P > .050).

Table 2.

Fall Incidence Rate, Number (Proportion) of Fallers, and Number (Proportion) of Multiple Fallers for Both Groups Across Time (Baseline and 12-Month Follow-Up) and Time Effects (Within-Group Comparison)

	Control Group				Intervention Group
Variable	Baseline	At 12 mo	Change, %	Time Effect	Baseline	At 12 mo	Change, %	Time Effect
Fall incidence rate, mean ± SD, falls/y	1.82 ± 4.19	0.94 ± 1.25	−48.4	.217	2.27 ± 4.17	0.80 ± 1.27	−64.8%	.039a
No. (%) of fallers	12 (54.5)	10 (58.8)	4.3	.757	16 (72.7)	6 (40)	−32.7	.036a
No. (%) of multiple fallers	8 (36.4)	3 (17.6)	−18.8	.156	9 (40.9)	3 (20)	−20.9	.152

Note. Values of Time effect are presented as P values from generalized estimating equation models.

^aStatistically significant.

Table 3.

Intent-To-Treat Comparisons of Fall Incidence Rate, Number (Proportion) of Fallers, and Number (Proportion) of Multiple Fallers, for Group Effects (Between-Group Difference at Each Time Point) and Group × Time Interaction Effects (Between-Group Difference for Change from Baseline to 12-Month Follow-Up)

	Group Effects		Group × Time Interaction Effects
Variable	Baseline	At 12 mo
Fall incidence rate, falls/y	.716	.747	.572
No. (%) of fallers	.214	.291	.080
No. (%) of multiple fallers	.757	.865	.967

Note. Values are presented as P values using generalized estimating equation models.

Figure 2.

Fall incidence rate (mean and SEM) at baseline and 12-month follow-up.

DISCUSSION

To our knowledge, this is the first RCT that directly assessed effects of daily use of bilateral custom-made AFOs to reduce prospective falls in community-dwelling older adults. A few previous studies investigated effects of AFOs on parameters related to fall risk,^8,9,33 such as balance, fear of falling, and gait, but not on actual falls.

The results of within-group comparison (Table 2 and Figure 2) suggest that participants in the IG had lower fall rate and less proportion of fallers compared to the preceding year. The fall prevention effects of AFOs could be attributed to improvements in balance and reduction in fear of falling thanks to ankle stability and enhanced somatosensory feedback provided by AFOs. This hypothesis is supported by our previous publication from this RCT,²³ in which we have demonstrated that the 6-month daily use of bilateral custom-made AFOs is effective to reduce the fear of falling compared to baseline. Also, our previous study revealed a significant reduction in ankle sway (an indicator of ankle stability) and center of mass sway (an indicator of postural control) compared to baseline and compared to the CG in response to daily use of AFOs.²³ Prior studies have suggested that reduction in fear of falling and postural sway are linked to lower fall risk in older adults.^34–36 Additionally, the high adherence and the high acceptability of AFOs reported in our previous study²³ could be an important factor that contributed to converting the improvement in the balance and fear of falling to an actual reduction in falls.

The observed fall incident rate (0.80 per year), proportion of fallers (40%), and proportion of multiple fallers (20%) during the 12-month daily use of custom-made AFOs are in line with the results of the 12-month multifaceted podiatry intervention (consisted of foot orthoses, advice on footwear, subsidy for footwear, a falls prevention education booklet, a home-based program of foot and ankle exercises, and routine podiatry care for 12 months), reported by Spink et al.³⁷ Spink et al³⁷ observed that, in their IG (n = 153), the fall incidence rate for the 12-month podiatry intervention was 0.67 per year, the proportion of fallers was 42%, and the proportion of multiple fallers was 14%; and these numbers were lower than those in their CG. The observed fall incidence rate in the IG in this study (0.80 per year) was higher than the one reported in the study by Spink et al.³⁷ This could be explained by the fact that our study was limited to only AFO intervention without providing other components of podiatry interventions, such as lower-extremity exercise.

We observed nonsignificant trends toward improvement of the fall incidence rate within the CG as well (Table 2 and Figure 2). This could be explained by a potential benefit from the use of fitted walking shoes, which were also provided to all of the participants in the CG. This was supported by previous studies in which it was suggested that poorly fitted shoes are one of the common risk factors for falls in older adults.^38,39 However, there was a trend that the observed effect from fitted walking shoes alone in the CG was lower than that provided by fitted shoes and custom-made AFOs together in the IG.

In Table 3, nonsignificant between-group differences for the change from the baseline to the 12-month follow-up (ie, nonsignificant Group × Time interaction effect) could be explained by the small sample size of this study. Based on the observed results, we did post-hoc power analyses using two-tailed independent-sample comparisons; and we found that a minimum sample size of 111 subjects per group will be required to observe significant between-group differences in the change of fall incidence rate from the baseline to the 12-month follow-up, with a statistical significance of 5% or lower and 80% power.

Another limitation of our study was that we used self-reported fall recall to calculate the outcome measures. Self-reported recall of falls may lead to underreporting due to older persons not recognizing a fall or not remembering a fall.⁴⁰ Additionally, we could not get a deep insight into when and where a fall occurred and whether the participants were wearing AFOs during the fall accident based on their self-reported recall of falls. In the future, we recommend that a daily fall calendar (gold standard) should be used to count the number of falls for each participant per year.

In conclusion, this exploratory study failed to demonstrate fall prevention benefits of bilateral custom-made AFOs compared to fitted walking shoes alone. However, a within-group comparison suggested that the AFO group had a lower fall rate and a lower proportion of fallers compared to the preceding year. The observations of this study should be, however, confirmed in a study with a larger sample size.