Biomechanical Evaluation of Pneumatic Sleeve Orthosis for Lofstrand Crutches

Crutch walking, especially when using a swing-through gait pattern, is associated with high, repetitive joint forces, hyperextension/ulnar deviation of the wrist, and excessive palmar pressure that compresses the median nerve. To reduce these adverse effects, we designed a pneumatic sleeve orthosis that utilized a soft pneumatic actuator and secured to the crutch cuff for long-term Lofstrand crutch users. Eleven non-disabled young adult participants performed both swing-through and reciprocal crutch gait patterns with and without the custom orthosis for comparison. Wrist kinematics, crutch forces, and palmar pressures were analyzed. Significantly different wrist kinematics, crutch kinetics, and palmar pressure distribution were observed in swing-through gait trials with orthosis use (p <0.001, p=0.01, p=0.03, respectively). Reductions in peak and mean wrist extension (7%, 6%), wrist range of motion (23%), and peak and mean ulnar deviation (26%, 32%) indicate improved wrist posture. Significantly increased peak and mean crutch cuff forces suggest increased load sharing between the forearm and cuff. Reduced peak and mean palmar pressures (8%, 11%) and shifted peak palmar pressure location toward the adductor pollicis denote a redirection of pressure away from the median nerve. In reciprocal gait trials, non-significant but similar trends were observed in wrist kinematics and palmar pressure distribution, whereas a significant effect of load sharing was noticed (p=0.01). These results suggest that Lofstrand crutches modified with orthosis may improve wrist posture, reduce wrist and palmar load, redirect palmar pressure away from the median nerve, and thus may reduce or prevent the onset of wrist injuries.

Designers and engineers have explored various crutch designs and attachment designs to help long-term Lofstrand crutch users, such as a spring-loaded crutch shaft, ergonomic crutch grip and shock absorbing crutch tip [17], [18]. However, no scientific studies have evaluated the effectiveness of these devices in terms of wrist posture improvement, load reduction or palmar pressure redistribution. Our group has previously designed a passive orthosis for providing wrist support that attaches to the crutch handle ( Figure 2). This device successfully reduced wrist extension and redistributed load away from the mid palm region to the adductor pollicis [19]. However, all loads still passed through the palm and the wrist.
Hence, we proposed to design a sleeve orthosis that attaches to the crutch cuff, supports the forearm, and minimizes excessive forearm movement during crutch gaits. This new design utilized a lightweight, yet powerful, soft pneumatic actuator to provide the constriction force to support the forearm and position the wrist in a more neutral posture. The actuator,   [20], (b) Illustration of FREE and generation of constriction force in coiled configuration, reproduced with permission [25], (c) Computer Aided Design (CAD) and prototype of the pneumatic sleeve orthosis.
a Fiber Reinforced Elastomeric Enclosure (FREE), is made of an extensible elastomeric tube constrained by families of inextensible fibers coiled around the elastomeric tube (Figuer 3a) [20]. Upon actuation (i.e., inflation), the FREE contracts and generates a tensile force in the longitudinal direction, thus providing constriction force about the forearm (Figuer 3b). Such constriction force creates an interface between the forearm and the crutch that allows load sharing and restricts wrist extension/flexion during crutch-assisted gait [21].
Creating a safe and comfortable interaction between the sleeve orthosis and the forearm is crucial for this design. Previous studies have shown that a normal pressure of 4.0 kPa (30 mmHg) around the forearm is safe for an extended period of time [22]. Compared with other compression products, grade four compression sleeves or stockings that are used to prevent the occurrence of venous disorders usually apply a constant constriction pressure from 4.0 to 5.3 kPa (30 to 40 mmHg) [23]. Intermittent pneumatic compression devices used to improve venous circulation can apply cyclic constriction pressures around 6.7 kPa (50 mmHg) [24]. Therefore, our design should provide a constriction pressure of 4.0 kPa (30 mmHg) or less when the user was not loading the crutch. This paper describes the design and biomechanical testing of a pneumatic sleeve orthosis for use with Lofstrand crutches. The mechanism of forearm interaction of the device is explained, followed by biomechanical testing to evaluate the effectiveness of the pneumatic sleeve orthosis on non-disabled young adults performing swing-through gait and reciprocal gait using the modified crutches. We hypothesized that, with the use of the orthosis, 1) there would be no differences in gait characteristic parameters, but participants would exhibit 2) a more neutral wrist position when loaded would be observed whereas elbow will be the similar, i.e., decreased wrist extension, decreased ulnar deviation angle, and reduced wrist range of motion, along with similar elbow flexion angle; 3) partial wrist force would be off-loaded to the forearm, i.e. increased crutch cuff force; 4) the palmar pressure magnitude would decrease, and the center of pressure location (COP) would shift away from the mid palm region towards the adductor pollicis; and 5) participants would show a reduction in perceived exertion during both swing-through and reciprocal crutch-assisted walking trials with the device compared to standard crutch use.

II. METHODS -DESIGN OF PNEUMATIC SLEEVE ORTHOSIS
To create a physical interface between the cuff and the forearm, we designed the orthosis to apply a moderate constriction force to the forearm, utilizing a contracting FREE coiled around a pair of hinged splints and cushion pads attached to the crutch cuff (Figuer 3c). We designed the splints and cushion pads to follow the contour of a forearm and to redistribute the constriction force over a larger forearm surface area. The FREE (outer diameter: 9.5 mm, inner diameter: 7.9 mm, length: 60 cm) was clamped onto the splints and could be freely adjusted to any other position along the splints. Upon actuation of the FREE, the tensile force generated along its length would decrease the coil diameter and apply a constriction force normal to the forearm (Figuer 3b) [25]. Such constriction force would generate a friction force between the forearm and the splint. Since the splint was attached to the crutch cuff, this force was then transmitted to the cuff. This loading arrangement thus distributed forces to both the handle and the cuff and resulted in reduced wrist force and palmar pressure. The constriction force also restricted the forearm from moving in the distal direction, which restricted movement of the forearm tendons and ligaments and thus restricted wrist extension and deviation. With a more neutral wrist posture, we expected that the peak pressure location would shift away from the carpal tunnel region toward the adductor pollicis area.
To establish a protocol and compare results among users, a pilot study determined that constraining the length of the FREE to equal twice the user's forearm girth plus 9.8 cm would be sufficient for most users. To ensure a safe and comfortable constriction pressure, the pilot study also determined that an actuation pressure of 308.2 kPa (30 psig) of the FREE would produce an overall constriction pressure on the forearm of less than 4.0 kPa when the crutch was not loaded. The constriction pressure would increase temporarily, such as during the crutch stance phase, but return to the lower pressure during the crutch swing phase.

A. Subjects
Eleven non-disabled adults were recruited to assess the effect of the pneumatic sleeve orthosis attached to Lofstrand crutches (5 males and 6 females; 18-32 years old; average height 171.6 ± 7.5 cm; average weight 67.3 ± 9.8 kg). The study was approved by the Institutional Review Boards at the University of Illinois at Urbana-Champaign (UIUC, IRB#15810) and the University of Wisconsin -Milwaukee (UWM, IRB#17.039). Testing was conducted at the Mobility Lab at UWM and informed consent was obtained from all participants. Before testing, anthropometric data including body height and weight, length, width, and girth of the forearms of each participant were recorded.

B. Testing Protocol
Joint kinematics, crutch kinetics, palmar and forearm pressure data, and perceived exertion were collected experimentally from each participant. Kinematic data were collected at 120 fps using a 15-camera motion capture system (T-series, Vicon Motion Systems, Inc., Oxford, UK) and a 39-marker upper extremity (UE) model with additional markers on the heel, toe and crutch on the dominant side (Figuer 4a, Figure 5c) [26]. A pair of previously designed custom instrumented Lofstrand crutches were used to collect crutch force data at 960 Hz (Figuer 4b) [27]. Each crutch was instrumented with two 6-axis load cells (MCW-6-500, AMTI Inc., Watertown, MA) to measure force applied to the crutch shaft and crutch cuff. Palmar pressure was measured at 60 Hz using a flexible 16 × 16 pressure measuring sensor (1 cm × 1 cm unit cell, Pliance Sensor S2129, Novel GmbH., Munich, Germany) folded and wrapped around the dominant side crutch handle. Several rows and columns of the measurement units were turned off to compensate for overlapping of certain measurement areas, resulting in a 13×12 measurement area (Figuer 4c). Four 1 cm diameter low-pressure sensors (Pliance Sensor S2011, Novel GmbH., Munich, Germany) were spaced around the inside of the sleeve, right under the hinge joints of the splint, on the dominant side to measure constriction pressure applied to the subject's forearm (Figuer 4c).
Prior to gait trials, each participant was given 15-30 minutes of practice to acclimate to swing-through and reciprocal crutch-assisted gait patterns following instruction [2].
In addition, a palm landmarks identification test was performed before gait trials to estimate the initial placement of the hand on the mat. In this test, the researcher applied pressure on the pressure mat that was wrapped around the handle at five points (P1: under carpal tunnel, P2: lateral side of the first metacarpal head (M1), P3: abductor pollicis transversus, P4: lateral side of the first metacarpal head (M2), P5: lateral side of the fifth metacarpal head (M5) marker) (Figure 5a,b). The test result was used later in the assessment of palmar pressure distribution for each subject.
Four test conditions, randomized for each subject, were evaluated: swing-through gait with orthosis, swing-through gait without orthosis, reciprocal gait with orthosis, and reciprocal gait without orthosis. Subjects performed the four tasks over a 6-m walkway at a self-selected speed. The pneumatic sleeve orthosis was attached and secured to each cuff of the instrumented crutches when testing the orthosis conditions. The length of the FREE actuator constrained on the sleeve orthosis was adjusted based on the aforementioned protocol. The FREE was then inflated slowly using shop air to 308.2 kPa after the subject inserted the forearm into the orthosis and held the crutch at a neutral wrist posture. A minimum of five trials were collected for each condition.
After completing each trial, subjects were asked to rate their perceived exertion (RPE), i.e. "Perceived exertion is a measure of how hard your body is working physically. What was your perceived exertion level for completing this trial?" using the Borg 6-20 scale [28].

C. Data Processing
Five gait cycles were analyzed per test condition for each subject. In swing-through gait trials, a gait cycle was defined as initial foot contact (0%) to the next foot contact (100%) of the dominant foot. Highly asymmetric swing-through gait cycles were dropped when the take-off and strike time of the left and right foot exhibited a time difference greater than one second. In reciprocal gait trials, a gait cycle was defined from heel strike (0%) to the next heel strike (100%) of the dominant foot. Reciprocal gait cycles were dropped in which the subject did not strictly follow a heel-strike to heel-strike gait pattern.
Kinematics, crutch kinetics, and pressure data for the dominant side in the selected gait cycles were collected and processed using Vicon Nexus (V2.5, Vicon Motion Systems, Inc., Oxford, UK), Novel Pliance (pliance/S, Novel GmbH., Munich, Germany) and MATLAB (R2016a, MathWorks, Inc., Natick, MA) software. Variables of interest were calculated and averaged per test condition for each subject.
1) Gait Characteristic Parameters: The gait characteristic parameters, including stride length, walking speed, crutch stance phase ratio (normalized by gait cycle), and peak and mean value of the norm of shaft forces (normalized by body weight) were averaged over the selected gait cycles for each test condition.
2) Joint Kinematics: Wrist flexion-extension, radial-ulnar deviation, and elbow flexion angles were computed in MATLAB using the aforementioned custom UE model [26]. Peak and mean wrist extension angles, peak and mean ulnar deviation angles, and peak and mean elbow flexion angles (during the crutch stance phase), as well as wrist flexion-extension range of motion (ROM) and elbow flexion-extension ROM (across the whole gait cycle) were reported.
3) Crutch Kinetics: Resultant norm of crutch cuff forces on the dominant side were calculated and normalized by subject body weight. Peak and mean values were reported during the crutch stance phase. 4) Palmar Pressure Distribution: Palmar pressure magnitudes and distances between the center of pressure (COP) to palmar landmarks were calculated. Peak and mean palmar pressures, normalized by subject body weight, were reported during the crutch stance phase.
The distance between palmar pressure landmarks and COP location were calculated to assess the shifting in the palmar pressure concentration location. However, the result of the palmar landmark identification is only valid to determine the relative displacement between the palmar landmarks, as the hand might move relative to the handle during the crutch stance phase due to wrist extension.
In this case, an algorithm was created to correlate palmar landmarks with pressure mat cells using motion capture data: We first identified the handle axis in 3D space using locations of five crutch markers (Figure 5b, c). Then we projected the hand marker for the 5 th metacarpal head (M5) onto the pressure mat wrapped around the crutch handle towards the handle axis using motion data of the M5 marker (Figure 5b, c) and then shifted the previously measured landmark locations from the identification test to match with the projected hand marker M5 location on the pressure mat at each collection sample. The mean and variance palmar distance between COP point and palmar landmark L3 and L4 (denoted as COP-L3 and COP-L4) were calculated during the crutch stance phase. 5) Forearm Pressure: Peak forearm pressure across the four low-pressure sensors was reported across the crutch gait cycles in the orthosis trials for both gait types, with the averaged peak forearm pressure calculated during the crutch stance and swing phases.
6) Rated Perceived Exertion: The average rate of perceived exertion score for the five trials per test condition for each subject was calculated and reported.

D. Statistical Analysis
The following parameters were selected to test our hypotheses: 1. Gait characteristics: stride length, walking speed, crutch stance phase, peak and mean shaft forces. 2. Joint kinematics: peak and mean wrist extension angles, wrist extension ROM, peak and mean ulnar deviations, peak and mean elbow flexion angle. 3. Crutch loading kinetics: peak and mean crutch cuff forces, along with peak and mean crutch shaft forces. 4. Palmar pressure distribution: peak and mean palmar pressure, as well as mean and variance of palmar distance between the COP and the palmar landmarks, L3 and L4: COP-L3 and COP-L4. 5. Perceived exertion: RPE in Borg's scale. Statistical analysis was performed to evaluate each hypothesis for each gait type (swing-through and reciprocal; SPSS Statistics V23.0, IBM Corp., Armonk, NY). Multivariate analyses of variance (MANOVAs) were employed to test for significant differences in wrist kinematics, crutch loading kinetics and palmar pressure distribution between trials with and without the use of orthosis. When significant differences were detected, univariate analyses of variance (ANOVAs) were computed to examine the effect of the orthosis on each individual parameter. In addition, an ANOVA was performed on the rate of perceived exertion for each test condition. P-values below 0.05 were considered statistically significant for all tests.

A. Overall Performance
All kinematic, crutch forces and pressure data from the 11 subjects were included in analyses of the swing-through gait trials, whereas only nine subjects' data were included in the crutch force and pressure data analysis of the reciprocal gait trials due to inconsistent foot striking patterns in two subjects. In addition, only eight subjects' data were included in the kinematics analysis in reciprocal gait trials, due to a missing wrist marker in all reciprocal gait cycles of one subject.

B. Gait Characteristics
The MANOVA test for swing-through and reciprocal gait trials found no significant difference in the selected gait characteristic parameters without and without the use of the orthosis (MANOVA: p = 0.56 and p = 0.4, respectively) (TABLE I). Their values were further compared with longterm crutch users from other literatures, these subjects exhibited a longer crutch stance phase, whereas similarities were observed in other parameters ( [3], [29], [30]).

C. Wrist Kinematics
The MANOVA revealed a significant difference in swingthrough gait wrist kinematics with the use of the orthosis in comparison to without orthosis (p<0.001) (TABLE II, Figuer 6a, b). Univariate ANOVAs identified significant effects of the pneumatic orthosis use for all analyzed parameters: peak and mean wrist extension angles during the crutch stance phase were reduced by 6.8% (p=0.01) and 6.0% (p=0.02), respectively. Wrist extension ROM was further reduced by 22.7% (p=0.008). Moreover, throughout the entire gait cycle, peak ulnar deviation angle was reduced by 26.3% (p=0.03) and the mean ulnar deviation angle was reduced by 32.6% (p=0.02).
The MANOVA test for reciprocal gait trials found no significant differences in the selected kinematic parameters with orthosis use (p=0.4) (TABLE II, Figuer 6i, j). Nevertheless, trends similar to swing-through gait trials were observed among all kinematic parameters in reciprocal gait trials.

D. Elbow Kinematics
No significant differences between elbow extension angles were identified in both gait trials between the conditions with and without the orthosis in swing-through and reciprocal gait trials (MANOVA: p = 0.27 and 0.32, respectively) (TABLE III, Figuer 6c, 6k). However, the range of motion of the elbow flexion angles were increased with the use of the orthosis by 12.2% and 18.8% for swing-through and reciprocal gait trials, respectively.

E. Cuff Kinetics
A significant difference was observed in crutch cuff forces in swing-through gait trials between the conditions with and without the orthosis (MANOVA: p<0.001) (TABLE IV, Figuer 6e). Both peak and mean cuff forces significantly increased with orthosis use. Up to 5.9% body weight (BW) was experienced by the crutch cuff with the use of the orthosis, compared to 2% BW in trials without the orthosis.
In reciprocal gait trials, similar significant results were observed in crutch cuff forces (MANOVA: p=0.01) (TABLE IV, Figuer 6l). Follow-up ANOVAs indicated a

F. Palmar Pressure
In swing-through gait trials, the result of the MANOVA on palmar pressure parameters indicated a significantly different palmar pressure distribution between conditions with and without the orthosis (p=0.03) (TABLE V, Figuer 6f, 6g). Followup ANOVAs identified a significant reduction in the peak palmar pressure by 8.4% (p=0.04) and in the mean palmar pressure by 10.5% (p=0.02) with the orthosis. In addition, the palmar distances COP-L3 and COP-L4 (L3: the adductor pollicis, L4: the second metacarpals) were also significantly decreased by 10.9% (p<0.001) and 13.5% (p<0.001), respectively. In addition, non-statistically significant decreases were observed in the variance of COP-L3 and COP-L4 in trials with orthosis use (−4.0%, −5.0%, respectively). However, no significant difference was observed in palmar pressure parameters from reciprocal gait trials (TABLE V, Figuer 6m, 6n).

G. Forearm Pressure
Similar levels of maximum forearm pressure were observed in both gait types when the crutch was unloaded (Figure 8). In swing-through gait trials, subjects experienced an average maximum forearm pressure of 4.1 ± 0.5 kPa during the crutch swing phase and 5.5 ± 0.7 kPa during crutch stance. In reciprocal trials, subject experienced 4.0 ± 0.1 kPa during the crutch swing phase and 4.1 ± 0.2 kPa during the crutch stance phase.

H. Rated Perceived Exertion
The rated perceived exertion (RPE) scores in Borg's scale were significantly reduced with orthosis use during swing-through gait patterns (p<0.001) (TABLE VI). No statistically significant reduction was observed in RPE scores for reciprocal gait patterns (p=0.2).

V. DISCUSSION
The pneumatic sleeve orthosis was designed to help longterm Lofstrand crutch users to improve wrist posture, reduce wrist force, and reduce and redirect palmar pressure.
In this study, we analyzed gait parameters, wrist and elbow kinematics, crutch forces, palmar pressure, and rated perceived exertion to assess the effectiveness of the pneumatic sleeve orthosis during two types of crutch gaits. Similar gait characteristics with and without the use of the orthosis were observed  IV  CRUTCH CUFF LOAD FOR SWING-THROUGH GAIT TRIALS AND RECIPROCAL GAIT TRIALS WITH AND WIHTOUT ORTHOSIS USE   TABLE V PALMAR PRESSURE DISTRIBUTION FOR SWING-THROUGH GAIT AND RECIPROCAL GAIT TRIALS WITH AND WIHTOUT ORTHOSIS USE in swing-through and reciprocal trials (TABLE I), indicating the use of the orthosis did not significantly alter characteristic gait parameter. However, orthosis use did significantly change hand and arm kinematics and kinetics. During swing-through gait, wrist kinematic data indicated significantly improved posture and a more restricted flexion-extension movement of the wrist, while elbow flexion angles were not significantly affected (TABLE II, TABLE III, Figuer 6a, b, c). Crutch kinetic data demonstrated load sharing between the wrist and forearm as peak and mean crutch cuff forces increased significantly with orthosis use (TABLE II, Figuer 6e). Since the total amount of force going through the crutch shaft was similar with and without the use of the orthosis (Figuer 6d), the increased load in the cuff with the use of orthosis indicated decreased load in the handle in this condition. In addition, palmar pressure results indicated a redistributed palmar pressure with the orthosis use with significantly reduced palmar pressure magnitudes, a shift of the COP location from the mid palm to the adductor pollicis, and an increasing consistency of the palmar pressure distribution (Figuer 6f, 6g, TABLE V). Furthermore, significantly lower level of exertion was perceived during completion of the gait condition with the use of the orthosis (TABLE VI). In reciprocal gait trials, statistical test results found significant effects of the pneumatic sleeve orthosis in crutch cuff forces (TABLE IV). Non-statistically significant trends similar to those in the swing-through crutch gait condition were observed in wrist kinematics, palmar pressure distribution, and rated perceived exertion (TABLE II,  TABLE V, and TABLE VI).
The aforementioned effects of the orthosis including wrist posture improvement, wrist force reduction, and palmar pressure redistribution may be generalized for long-term crutch users. Even though this study was only conducted on abledbodied individuals trained to perform crutch-assisted gait, high similarities in crutch gait characteristics were observed among these abled-bodied individuals and long-term crutch users from prior literatures (TABLE I). In this case, similar benefits might be observed on long-term crutch users if their crutches were modified to include the orthosis device.
In addition, the orthosis device may even have beneficial effects on preventing aggravation or development of CTS. Keir et al. have previously identified a wrist extension beyond 32.7 • and an ulnar deviation beyond 14.5 • would result in a critical carpal tunnel pressure (CTP) of 30 mmHg [16]. Another study found that the lowest CTP occurs within 2 • ± 9 • of wrist extension and 2 • ± 6 • of ulnar deviation [14]. Compared with results from this study, the mean wrist extension angle in reciprocal gait with the use of the orthosis (33.3 • ) was very close to the wrist extension threshold (32.7 • ), and the mean wrist deviation angles with the use of orthosis in both swing-through and reciprocal gait conditions (6.4 • , 7.7 • , respectively) are within the range of wrist postures resulting in the lowest CTP. The reduced peak and mean palmar pressure magnitude during the crutch stance phase, along with the shifted palmar COP location from the carpal tunnel to the abductor pollicis region, also suggest the potential effect of orthosis use to reduce CTP [13]. The non-significant effect of the pneumatic sleeve orthosis in reciprocal gait can be explained by the nature of reciprocal gait as well as potential difficulties in imitation. Reciprocal gait is generally less physiologically and biomechanically demanding than swing-through gait due to the existence of double limb support. In this case, only a small portion of the body weight is supported by the crutch, which results in less extended and deviated wrist joint angles, lower forces experienced by the wrist, as well as lower palmar pressure magnitudes. The wrist extension and extension ROM were less extreme compared with those in swing-through crutch gait when using reciprocal gait patterns (TABLE II). In this case, due to a lower wrist loading and more neutral wrist posture, palmar pressure was drastically decreased and more evenly distributed when using reciprocal gait pattern. In addition, reciprocal gait is harder to imitate compared with swing-through gait, as it requires more coordinated synergistic movements in upper and lower limbs, compared with swing-through gait. In this case, it could be possible that the reciprocal gait pattern performed by abledbodied subjects, even with extended period of training time, were still immature and did not fully imitate that of long-term crutch users, resulting in nonsignificant results.
In this study, we have also demonstrated the feasibility and benefits of the unique load transfer interface utilized in the proposed design. Compared with the passive orthosis developed earlier by our group [19], such active orthosis has shown additional benefits of wrist off-loading. Both studies found significant differences in the mean wrist extension angles with and without the use of orthosis in swing-through crutchassisted gait. However, no significant difference in peak and mean palmar pressure was indicated in the study on the passive wrist orthosis. Such load transfer interface in our pneumatic sleeve orthosis can also be simplified and potentially utilized in other assistive devices. The use of a pneumatic device in this design represents a light-weight mechanism with shapemorphing and stiffness-varying capability. We do realize this particular design with pneumatically actuated soft actuator is too complicated and not practical for commercialization. However, people can easily build such mechanism with cheaper material and simpler design [31].
The constriction pressure applied to the user's forearm could be further optimized. A constriction pressure ≤ 4.00 kPa is considered safe to be applied to the human skin for an extended period of time, without causing reduction in blood flow and discomfort [22]. The average peak constriction pressure during the crutch swing phase in swing-through gait trials was 4.07 kPa, whereas the average pressure during reciprocal trials was 3.97 kPa. The actuation pressure and the effective length of the FREE can be further customized for each individual subject to achieve the best effectiveness of the pneumatic sleeve orthosis and a user-selected constriction pressure during the interaction of the device and the forearm.
It should also be noted that the proposed sleeve orthosis is mainly affecting the wrist and the palm, since the load transfer happened between the forearm and the cuff. Similar elbow kinematics and crutch shaft forces were observed between conditions, indicating little effect of the orthosis on other upper limb joints. In this case, injuries associated with elbow and shoulder joint due to crutch-assisted gait may not be reduced.
Limitations of the current study include the use of nondisabled subjects and a small sample size. Even though all subjects received crutch gait training before starting data collection, their gait pattern could still be different from those of long-term crutch users. Previous researchers had found a significant increase in the crutch stance time for long-term crutch users compared with healthy subjects, due to diminished strength in the lower extremities among crutch users [29]. Slightly shorter crutch stance phases in both swing-through and reciprocal crutch-assisted gait trial were also observed among abled-bodied subjects in this study. In addition, the sample size is relatively small for this study. Data from 11 subjects were collected during swing-through gait trials, whereas data from a fewer number of subjects were used in reciprocal gait trial analyses. Lower sample size may have reduced the statistical power as well as the accuracy of the statistical analysis results.

VI. CONCLUSION
We designed a pneumatic sleeve orthosis utilizing a soft pneumatic actuator and demonstrated its effectiveness for improving wrist posture, off-loading wrist force to the cuff, and reducing and redirecting palmar pressure while safely interacting with users of Lofstrand crutches. The highest effects were observed in the more biomechanically taxing swing-through crutch gait.
The statistically significant findings and non-statistically significant trends in swing-through and reciprocal crutch gait trials suggest potential clinical impact for crutch users. The improved wrist posture and restricted hyperextension could reduce median nerve pressure and development of carpal tunnel syndrome [13], [14], [15], [16]. The significant increase in the crutch cuff force suggests increased load sharing between the wrist and the forearm. In this case, the force experienced by the wrist through the crutch handle was reduced with the use of the orthosis, suggesting a lowered risk of wrist joint injuries. Moreover, the reduced peak palmar pressure magnitude and redirected COP location suggest the potential reduction of carpal tunnel pressure due to reduced externally applied pressure. This promising pneumatic sleeve orthosis may assist long-term crutch users by providing a more neutral wrist posture and lowering wrist and palmar forces to alleviate and potentially prevent the development of wrist pain and CTS.