Abstract

Manual wheelchairs are available with folding or rigid frames to meet the preferences and needs of individual users. Folding styles are commonly regarded as more portable and storable, whereas rigid frames are commonly regarded as more efficient for frequently daily use. To date, there are no studies directly comparing the performances of the frame types. Furthermore, while differences have been reported in the longevity of the frame types, no efforts have been made to relate this durability back to the real-world performance of the frames. This study investigated the propulsion efficiencies of four folding and two rigid ultra-lightweight frames equipped with identical drive tires and casters. A robotic wheelchair tester was used to measure the propulsion costs of each chair over two surfaces: concrete and carpet. A motorized carousel was used to drive the chairs 511 km around a circular track to simulate one year of use for each wheelchair. After simulated use, five of the six wheelchairs showed no decrease in propulsion effort, indicating that the frames were able to withstand the stresses of simulated use without a detrimental impact on performance. In the unused “new” condition, rigid chairs were found to have superior (>5%) performance over folding frames on concrete and carpet, and in the “worn” condition rigid chairs had superior performance over folding chairs on concrete but were comparable on the carpeted surface.

Introduction

Ease of control and maneuverability are crucial factors of manual wheelchairs (MWCs) that influence the mobility and the quality of life of wheelchair users. Unfortunately, MWCs are inherently lossy and inefficient [1,2]. Users exert significant energy to reach desired speeds or to negotiate inclines and can experience fatigue over long bouts of mobility. Expending this propulsion effort results in an accumulation of heightened repetitive loads on the upper extremities. Over time, these forces can develop into injuries related to overuse [36]. Physiological strain on users can be reduced by changing the wheelchair configuration to improve the wheelchair efficiency. Several factors have all been shown to improve wheelchair efficiency and lower the energy costs of self-propulsion, including informed tire selection [79], proper tire maintenance [10], and maximizing the weight distributed over the drive wheels [11,12]. Yet, many other critical aspects of wheelchair design have yet to be investigated.

For ultra-lightweight wheelchair users, selecting a frame style is a fundamental decision. Choosing between a folding or rigid frame is a mandatory choice that comes before a user chooses wheels, tires, or cushions. Folding frames feature a cross-brace folding mechanism with thick-walled frame tubes to reinforce the strength of the structure. The footrest hangers are often removable for ease of storage. Drive wheels are attached separately to an axle housing on each side of the frame. In contrast, rigid frames can afford to be much more compact due to the welded structure. Less material is needed to reach the same structural rigidity, resulting in a lighter frame weight. Additionally, rigid frames have footrests built into the frame and the drive wheels are connected by one solid axle housing that spans the width of the chair. Some users prefer lighter rigid frames for easier lifting and stowing of the chair, whereas others prefer the form factor of the collapsible folding style to facilitate transport and/or storage. Folding and rigid frames are offered at similar price points in the ultra-lightweight category (e.g., often over $1000) so until users know their particular preference, it may be difficult to select a frame type for a new MWC.

Rigid chairs are heavily implied to have a performance advantage over folding chairs for several reasons. First, rigid chairs are found exclusively within the custom-made, high-performance ultra-lightweight category of wheelchairs, whereas folding chairs are much more commonly found in mass-produced depot wheelchairs and lightweight chairs designed for short-term use. Second, manufacturers commonly boast about the efficiency of their folding frames by claiming their folding frame “feels as efficient as a rigid frame.” From a mechanical perspective, energy supplied by the user to a rigid frame will be dissipated primarily through nonconservative forces (e.g., rolling resistance or scrub torque) imposed on the wheels [13]. On a folding frame, additional energy could be exhausted through joint flexion and vibration in the frame as well as relative motion of the removable, loosely attached parts like the footrest. Incremental differences in mass between the frame types are not expected to have any influence on the propulsion efficiency [8,10,14]. However, to our knowledge, folding and rigid ultra-lightweight wheelchair performances have never been directly compared. Performance comparisons could be used as a reference to inform consumers and clinicians when selecting a new frame.

Long-term use of MWCs may be influenced by the durability and longevity of frames. The American National Standards Institute (ANSI) and Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) established several standardized tests to assess the fatigue strength of wheelchairs using double-drum rollers and curb-drops. Gebrosky et al. used these ANSI/RESNA tests to assess the durability or “survivability” of several types of wheelchair frames. Their results suggest that ultra-lightweight frames are more durable than standard or lightweight models [15]. More specifically, folding ultra-lightweight frames were shown to withstand more than three times the amount of fatigue test cycles, on average, than rigid aluminum and 70XX-alloy aluminum frames could survive before failure [16], suggesting that folding ultra-lightweight frames have greater durability than their rigid counterparts.

Double-drum and curb-drop fatigue test protocols do not include any assessment of the impact of structural or mechanical fatigue on tested wheelchairs. Rather, these tests are focused on inducing failure in a structural frame element by imparting stresses that do not reflect those typically incurred during everyday use. The minimum requirements to pass the ANSI/RESNA tests are surviving 200,000 revolutions of the double-drum test with attached metal slats to impact the wheels and 6666 curb-drops. The wheelchair fails to pass if any components are deformed or damaged before these requirements are reached, with the exception of “wear items” such as tires or inner tubes, which may be replaced once during the tests. Ultra-lightweight wheelchairs are exposed to a wide range of environments and stressors during everyday mobility. In real-world use, the user imparts changes in momentum to the chair that manifest as slow, discontinuous motion with frequent stops, starts, and turns [17,18]. Stressors arise during motion due to environmental factors and inertial forces. Rough or uneven surfaces impart vibratory and impact forces to the wheelchair system while turning maneuvers generate cornering forces that apply side-loads on the bearings within the front and rear wheels [19]. These forces and environmental factors are not applied within current failure tests. Therefore, their results may not be generalizable to real-world wheelchair usage.

Propulsion efficiency is a more sensitive parameter to determine how wheelchair performance deteriorates after simulated use testing. Fundamentally, wheelchair propulsion is tied to both the mechanical system of the vehicle as well as the biomechanics of the user, and propulsion efficiency has been studied using human subjects in the past [10,20,21]. This approach is well-suited to represent the propulsion effects of certain factors like pneumatic tire inflation levels on actual wheelchair usage [10], but the measurements can have high variability between pushes [22] and/or between subjects [23,24]. Laboratory experiments using automated or mechanized test methods improve the precision of measurement [9,25,26] and isolate the effect of wheelchair design and components on its mechanical efficiency and performance.

The aims of this study are to (1) compare propulsion costs between folding and rigid ultra-lightweight MWC frames before and after a year of simulated use and (2) assess the changes in propulsion cost for each frame type caused by simulated use. A repeatable robotic wheelchair propulsion device was used to standardize occupant characteristics. Weight distribution (WD) and general geometric measurements were controlled to isolate the impact of the frame style from other prominent confounding factors. The outcome of this study will provide guidance to users and clinicians as they weigh the tradeoffs between frame types in relation to short- and long-term performance.

Hardware and Configurations

Experimental methods involved assessing the propulsion cost of six MWC frames equipped with standardized casters and drive wheels.

Wheelchair Testing.

The Anatomical Model Propulsion System (AMPS), shown in Fig. 1, is a wheelchair-propelling robot used to maneuver the wheelchair in this study. Its propulsion subsystem permits highly repeatable, configurable propulsion patterns to be deployed across a wide range of chair configurations. Its construction mimics a seated person in size, shape, and mass distribution to apply realistic loads to the frame and wheels as per the wheelchair test dummy standard defined by ISO 7176-11 [27], scaled to a total mass of 80 kg. Wheelchair propulsion is controlled by motors attached to custom-made ring gears replacing the push-rims of each drive wheel. A high-powered motor speed controller (HDC 2460, RoboteQ Inc.) utilizes velocity-based feedback system to impart discrete and highly repeatable pushes. Between pushes, the motors are mechanically disconnected from the push-rims, allowing the wheelchair to freely coast. Rotational speed of each drive wheel is collected from axle-mounted optical encoders (EM1-2500-I, US Digital Inc.) fixed to the drive wheel axles. Rotational power is calculated from the voltage and current across the motor armatures. Signals are sampled at 40 Hz by an on-board data acquisition system (USB-6341, National Instruments Corp.). More details of the design and overview can be found in the original design article by Liles et al. [28].

Wheelchair Frame and Component Selection.

Six ultra-lightweight wheelchair frames were used: four folding frames and two rigid frames. These selected wheelchair models were chosen to represent common commercially available ultra-lightweight wheelchairs for full-time users. Mass and WD, reported as percent total mass over the drive wheel axle, are shown for each configuration in Table 1. Aspects of the construction of each frame and other design elements differed between the tested frames, such as the frame mass and the style of folding mechanisms, but are not believed to have significant impacts on measured performance based on existing knowledge [10,20].

Each frame was equipped with standardized 5 in. × 1 in. polyurethane caster wheel assemblies (Primo 5 × 1, Xiamen Lenco Co., Ltd.). 24 in. × 1-3/8 in. pneumatic drive tires (Primo Orion, Xiamen Lenco Co., Ltd.) were fitted onto the OEM spoked wheels for each frame. These standardized components are shown in Fig. 2. The default Drive Medical wheels used 12 mm-diameter axles. The wheels, axle, and mounting hardware were all replaced to equip a standard metal spoked set of wheels on ½ in.-diameter axles.

Maneuver Selection.

A “Straight” maneuver was developed to drive the chair in a straight line. Wheel trajectories (Fig. 3) include two initial pushes to accelerate the wheelchair, five pushes to maintain a steady-state travel speed between 0.8 m/s and 1.0 m/s, and a coast-down period to naturally decelerate the chair to a complete stop. Similar travel speeds were used in prior studies on propulsion effort and vibration exposure [20,29] of human subjects. Travel started at the same position for each tested configuration. Distance for each trial was approximately 10 m on average and varied slightly between trials and configurations.

Tests were conducted on two surface types: smooth-poured concrete and low-pile (high-traffic) carpet, collectively seen in Fig. 4. Concrete is an especially common outdoor surface used for many walkways. The specific concrete flooring utilized for this study was a style of warehouse flooring for forklift paths that is especially flat and wide to minimize vibrations and risks of tipping for forklifts transporting heavy loads. While this surface was beneficial to maintain a straight trajectory and repeatable travel path during robotic propulsion trials, lower propulsion efficiency is expected when users traverse outdoor concrete paths that feature inclines or side-slopes. Low-pile carpet was chosen to represent a surface commonly used indoors within the home or in corporate office spaces.

Simulated Use—Carousel Test Track.

Simulated use of the manual wheelchairs was conducted via motorized carousel tester inside a controlled testing area (Fig. 5). The central column houses a motor that rotates the carousel arms. Wheelchairs were loaded with 80 kg dummies and connected to the carousel arms via metal linkages that grip the wheelchair frame. Obstacles were placed along the track to impart additional shocks and vibrations to the frame. An aluminum door threshold (Frost King CS514/36, Thermwell Products Co. Inc.), an aluminum diamond-pattern plate, and a double-sided steel ramp were secured to the polished concrete floor of the test track.

The three obstacles were placed on the track as shown in Fig. 6. The track was measured to be 3.6 m in diameter, from wheelchair center to wheelchair center. The central motor was set to rotate the system at 5.3 revolutions per minute to drive the chairs at approximately 1 m/s, which is considered a “fast” travel speed for manual wheelchair users. An average full-time MWC user travels approximately 1.4 km daily with an average bout speed of 0.4 m/s, with fast (>1 m/s) bouts comprising less than 4% of recorded bouts [17]. One year of use averages out to 511 km of total travel. The carousel track was used to drive the chairs half of this amount, 255.5 km, in the clockwise direction, or around 71 h of continuous track use. The carousel was stopped at regular 12-h intervals, representing approximately one month of use, to check tire inflation and clear debris from the track. At the end of the 71-h travel, the yokes were reversed and the chairs were driven an additional 255.5 km in the counter-clockwise direction, always facing the direction of travel.

An assessment of stressors induced by the track and obstacles was conducted by measuring the vibration exposure at the wheelchair seat. One triaxial accelerometer (X16-1D, Gulf Coast Data Concepts) was used to measure vibration data along the vertical axis. The sensor was placed on the top surface of the wheelchair cushion, centered underneath the torso of the 80 kg wheelchair dummy, with the primary axes of the accelerometer aligned with the primary axes of the wheelchair (x for antero-posterior, y for medio-lateral, z for vertical), as in ISO 2631-1 [30]. Surface-related vibration magnitudes are most dominant in the global vertical direction of the wheelchair [20,31,32], and the vertical direction is most closely associated with discomfort [33] and physiological injury [3436] out of the three axes. Therefore, vertical vibrations at the seat were measured for the first eight hours of travel as the chairs were driven around the circle track. The frequency-weighted root-mean-square vibration was calculated as per ISO 2631-1 as 0.894 m/s2 which is largely comparable to the daily root-mean-square vibration exposure of wheelchair users (0.83 ± 0.17 m/s2) reported by Garcia-Mendez et al. [37]. This brief vibration assessment would suggest that the circle track tester induced similar vibrations that are reflective of real-world use scenarios.

Methods

Data Collection.

At each stage of simulated use, each wheelchair was tested using the AMPS. To account for any slopes or inconsistencies in each surface, six trials were run in opposing directions along the same path for a total of 12 trials per surface per wheelchair in each stage of simulated use, for a total of 288 over-ground AMPS trials. Data from the motor armature voltage and current sensors, motor encoders, and wheel-mounted encoders were collected at 40 Hz during the over-ground trials. These data were processed in matlab (R2020a, The MathWorks Inc.). Butterworth filters were used to smooth each sensor signal before calculating the input power to the system from each motor.

Propulsion Cost Calculation.

Propulsion cost reports the amount of energy exerted by the AMPS to perform a maneuver, normalized by the distance traveled by the wheelchair. A configuration with lower propulsion cost has greater efficiency. The calculation of propulsion cost is derived from the fundamental theorems of work-energy. Rotational power is found by multiplying the propulsion torques on each wheel (τL, τR) by the corresponding wheel speed (ωL, ωR). Wheel torques are calculated from measurements of the motor armature current, as in Ref. [28], and scaled by the gear ratio of the motor pinion to the custom push-rim. Wheel speeds are measured directly from the hub-mounted optical encoders on each drive wheel. Total work supplied to the chair is then calculated by integrating the sum of the left and right rotational powers over the duration of the maneuver, from the start time (ti) to the end time (tf). Finally, the propulsion cost value is calculated by dividing the total work supplied to the chair, in Joules, by the linear displacement or distance traveled by the center of mass (Δs), in meters
PropulsionCost=titf(τLωL+τRωR)dtΔs
(1)

Statistical Analysis.

Descriptive statistics (mean, standard deviation, coefficient of variation) were calculated for propulsion cost and vibrations across all configurations and surfaces. Analysis consisted of assessing group differences and calculating the magnitude of differences. Equivalence tests were used to assess propulsion cost differences between folding and rigid frames before and after simulated use, and complementary equivalence tests were used to assess differences between worn and new conditions across the other variables. Equivalence testing is used to assess when device performance differs by more than a practically relevant or meaningful amount. Its analysis is often more appropriate than inferring a lack of a difference when assessed by traditional statistical means [38]. Operationally, equivalence tests are a two-sided evaluation of differences using confidence intervals (CI) in relation to equivalence limits. Relevant and meaningful differences were informed by a review of published studies on wheelchair propulsion efforts with human subjects. In these studies, the subjects were asked to propel a wide variety of wheelchair configurations including power-assisted wheels [39], lever-driven wheels [40], sports wheelchairs [41], chairs with weights added to the frame [14,21], and with under-inflated tires [10,42]. Across these studies, the average difference between the biomechanical outcome variables of the studies was calculated to be 9.4%. Therefore, we defined the equivalence interval as ±5% based on the assumption that mechanical testing is more precise than human subject investigation.

The parameter of interest was the ratio of propulsion costs between a “Test” and “Reference” data set. For comparisons between frame types, the rigid frame group was the reference. For comparisons between simulated use conditions, the brand-new frame group was the reference. The respective CIs of the propulsion cost ratio (Test Mean/Reference Mean) are used to define one of three categories: Superior, Comparable, and Inferior. In the superior case, the entire CI must lie below the lower equivalence limit of 0.95 (indicating lower energy loss of the Test condition). Analogously, the inferior classification is defined by the CI lying above the upper equivalence limit (e.g., 1.05). When the CI crosses either equivalence limit, the Test condition is comparable, and when the CI remains between the limits, the Test and Reference conditions are equivalent. Superior and inferior classifications represent situations where the Test condition was significantly different than the Reference, relative to the equivalence limit.

Results

Propulsion cost measurements are summarized for each frame type, surface, and frame condition in Table 2. Lower propulsion costs are associated with better performance. Travel over concrete requires less energy (11–15 J/m) than travel over carpet (17–27 J/m). The propulsion costs tended to decrease after simulated use by up to 10%. Propulsion cost measurements were highly repeatable for each wheelchair. Coefficients of variation (CVs) for each set of frame data on concrete (N = 12 for each individual frame) were ≤ 4.4%. Averages across all folding (N = 48) or rigid (N = 24) frames gave CVs ≤ 4.0%. Carpet trials had more variation with CVs between 7.5% and 11.1%.

The four individual folding frames did not all behave identically. Specifically, chair F2 experienced an increase in propulsion cost after simulated use, unlike the other five frames. Differences between individual frames may require further investigation with broader sample sizes from each frame type. However, the low CVs of the F(all) and R(all) data sets suggest that the average over all folding or all rigid frames were representative of the group as a whole. Several significant differences were found between the propulsion cost data sets using equivalence tests (Fig. 7). Rigid frames had superior performance (i.e., lower propulsion costs) over folding frames when traversing concrete in both new and worn conditions. On carpet, rigid frames only had superior performance over folding frames in the worn condition. The two frame types had comparable propulsion costs over carpet in the new conditions.

Comparisons between new and used conditions only identified one significant difference (Fig. 8). Worn rigid frames had superior performance to brand-new frames over concrete. Folding frames had equivalent performance over carpet between new and worn conditions, and both frame types exhibited comparable performances between new and worn conditions over carpet.

The complete equivalence test results are summarized in Table 3.

The circle test track caused visible deterioration to the components. A collection of images (Fig. 9) shows the extent of wear on the drive wheels and casters including broken spokes, tread wear on drive tires and casters, loosening of threaded components, and wear to the sidewalls of casters and tires.

Additional damage (not pictured) included components (e.g., removable footrests) breaking loose from their normal forward-facing position and the loosening of the locking mechanisms of two different folding frames. These mechanisms were notably difficult to lock into place when the frames were new, but very easily deployed and re-folded after use.

Discussion

Our study indicates that rigid ultra-lightweight frames had superior performance to folding frames. Furthermore, there is no evidence to suggest frame fatigue resulted in increased propulsion cost, and there is no generalizable performance degradation after simulated use for either category. None of the observed damage to the wheelchair frames and components resulting from the carousel track testing had any adverse effects on the propulsion costs of either frame type. With that said, comparisons of these results with existing literature would suggest that other wheelchair configuration differences are more impactful to propulsion costs than the differences observed between folding and rigid frames, or the differences between new and used frames.

While there are many features of wheelchair frames that are left to be investigated, there are no specific features that are anticipated to impact the generalizability of these results. Based on prior studies of manual wheelchair propulsion cost [7,8], as well as other literature assessing manual wheelchair performance with various metrics [10,14,20], the vast majority of differences in propulsion efficiency are caused by wheel and tire selection, and not from factors like frame material, folding mechanism style, or frame mass. Chénier et al. reported no significant effects on weight-normalized work per meter during human propulsion between three frame materials (aluminum, titanium, carbon fiber), nor between four different styles of folding mechanisms within the cohort of the four tested aluminum folding frames [20]. Adding up to 5 kg [14] or 10 kg [10] to the mass of the wheelchair frame had no noticeable impact on the energetic efficiency of human wheelchair propulsion. To put this in perspective, 10 kg of extra mass would nearly double most ultra-lightweight frame masses, and make the mass comparable to many common nonlightweight models. In contrast, simply changing from pneumatic to solid tires increased propulsion costs by 41–54% [7,8] and significantly increased metabolic costs by 17% [10] with other factors held constant. Caster selection was shown to impact propulsion costs by up to 20% as well [7]. For optimal wheelchair performance, it seems most beneficial to select components that have the most ideal performances over common surfaces [8,9] and prioritize proper maintenance including inflation [10] and wheel alignment [25].

Brand-New Frames.

New rigid frames had higher levels of performance than new folding frames over concrete. Folding frames are constructed with many moving joints and removable parts compared to welded rigid frames. Flexion and minute motion in the joints help to absorb and dissipate energy instead of being transmitted directly through rigid tubes and accumulating at welds. This energy dissipation could account for the greater longevity and durability of folding ultra-lightweight frames over rigid frames reported by Gebrosky et al. [16]. However, wheelchair propulsion efficiency is defined by how much of the energy supplied by the user is turned into motion. Energy dissipation within wheelchairs is counter-productive to propulsion efforts, as seen when using certain vibration-attenuating materials for frames [20] and tires [8,9].

New folding and rigid frames had comparable performances over carpet. We attribute this to the high levels of energy loss from the rolling resistance over carpet [8,9], which could easily overshadow the small differences between frame types that were evident from concrete test results.

Effect of Simulated Use.

Simulated use had varying effects over the cohort of wheelchairs. Trends for both frame types suggest that regular use did not inhibit performance. In fact, the performances for both frame types actually improved after simulated use, although the respective average performance was deemed comparable between the new and worn states. Evidently, there is some moderate benefit to “breaking in” some parts of the wheelchair, yet one folding frame did experience performance degradation and showed signs of frame fatigue (chair F2), so these results should not be overly generalized.

We can posit that the degradation to chair F2 and the consequent decrease in its propulsion cost may be due to the choice of components (e.g., bearings, frame articulations) and/or manufacturing quality. In particular, the folding mechanism and rear wheel axle receptacles on chair F2 were loosened over the course of simulated use testing. The rear wheels consequently experienced significant toe-in and ultimately resulted in the observed damage to the sidewalls of the tires seen in Fig. 9(e). The wheels were re-aligned before the second round of AMPS-based propulsion testing, yet some minor misalignment may have remained. As shown by Ott et al., even 1 deg of toe-in or toe-out can vastly increase the rolling resistance of these wheels [25]. Rigid frames may be beneficial at maintaining wheel alignment due to the structural integrity afforded by the solid rear axle, though even rigid frames can experience misalignment due to wear patterns and slop in the bearings or misalignment of camber adjustment in the axle tube. As such, wheelchair users should adhere to a proper maintenance schedule and monitor their wheel alignment, especially if they notice any changes in their propulsion effort. Ultimately, based upon our results, we have no evidence to suggest that frame fatigue led to increased cost of propulsion.

Improved Performance After Simulated Use.

The reduction in propulsion cost after simulated use was an unexpected result, as it was believed general wear and tear would reduce the system efficiency by damaging the bearings, wheels, or tires. Based on the low magnitudes of bearing friction [43] compared with the overall rolling resistance of wheelchair wheels [9,25], we hypothesized that the majority of the propulsion cost reduction came from changes in the tires. Wheel rolling resistance stems from hysteresis losses at the interface between tire and ground. The tire tread material deforms to create a flat contact patch between the tire and the ground. Unequal forces are required for tire deformation and elastic recovery. As a result, tire deformation consumes energy as the wheel rolls along the ground [19]. The tires used in this study were pneumatic tires with moderate tread which was partially worn off (Fig. 9(b)) throughout the carousel testing duration, with wear patterns that suggest scrubbing against the floor during the constant turning was the dominant contributor. Empirical data collection with a Shore-A hardness tester showed no differences in hardness values from new and worn casters and tires. It seems likely that the tires formed a flatter contact patch with the ground after so much of the tread was worn off, which consequently reduced the amount of deformation occurring at the floor-tire interface. Thus, the overall mechanical energy losses due to hysteresis in rolling could be consequently reduced. This behavior has been observed in automotive tire testing [4447]. In Ref. [44], Sandberg et al. reported that the rolling resistance coefficient of automobile tires decreased by up to 20% by wearing away 6 mm of tread depth (i.e., not an insignificant amount of wear). To investigate the possibility of this happening with the wheelchair components used in this study, empirical rolling resistance data were collected on one representative set of new and worn drive wheels. A coast-down cart was loaded to 80 kg with 70% of the weight supported over the drive wheels to represent the loading conditions in this study. Rolling resistance data were collected with the methods presented in Ref. [9]. Rolling resistance forces of the worn tires were 8% lower than the new wheels (new: 1.43 N, worn: 1.31 N). While not a trivial outcome, this trend needs to be corroborated in further studies.

Comparisons to Existing Literature.

Existing long-term wheelchair fatigue and impact tests, such as the standardized ANSI/RESNA tests (ISO 7176-8), are predominately focused on characterizing wheelchair durability and documenting failures of individual components. These tests do not assess any over-ground performance metrics that reflect the mechanical efficiencies of the tested wheelchairs. In contrast, our simulated use testing was uniquely focused on fatigue-induced performance differences using the outcome metric of propulsion cost.

Gebrosky et al. showed that most ultra-lightweight folding frames have higher durability and longevity than rigid frames [16]. In contrast, our findings suggest folding frames experienced greater performance degradation. These results are not necessarily in conflict with those of Gebrosky et al. [16] because “durability” and “performance” hold two very different meanings. Frame durability has been a dominant measure of wheelchair quality (i.e., lightweight versus high-strength lightweight wheelchair categories), and has been tested with curb-drop and drum tests [16,4850]. However, these tests are destructive by intention, and may not be representative of real-world use. Deterioration of components and breakdowns can affect performance but do not reflect the performance of the frame itself. Life-cycle and failure reporting research on wheelchairs show that breakdowns and repairs are likely within the first 12 months of use [5153], but are not typically related to frame failures. Out of 352 MWC users, 129 (36.6%) needed no repairs, 88 (25.0%) needed 1 repair, and 122 (34.7%) needed more than 1 repair. From those needing only 1 repair, a dominant 46.3% of the repairs were caused by wheel and caster failures, whereas only 3.5% were caused by frame failures [51]. Standardized failure testing requires reporting of component breakdowns but continues until frame failure, which is evidently infrequent in real-world use. Our carousel track testing resulted in two flat tires and one set of broken spokes on a rear wheel that had to be replaced. No frame fractures were reported on any of the six tested frames.

Perhaps folding chairs do not suffer as much damage due to the flexion and energy dissipation due to moving parts, but the survivability may be at the expense of mechanical performance. Real-world use includes both considerations of durability and performance; this study investigated performance, and the most feasible explanation for the similarity between pre- and post-simulated use performances in folding frames is that the frames experienced deterioration.

Clinical Implications.

One year of simulated use did not adversely affect the performance of five of the six tested wheelchairs. The primary indication of wear was in the drive wheels. While the loss of tread did not affect propulsion efficiency, it will impact traction and perhaps the ability to withstand puncture. Based upon this study and results from reported failures [51], users should be advised to inspect their wheels annually and consider replacing them when loss of tread is evident. Furthermore, real-world usage of the manual wheelchair can expose the frame and components to dirt, debris, and other environmental factors that were not present in the controlled environments used in this study. It seems possible that exposure to these conditions could interfere with the quality of things like the bearings or other moving parts, despite the majority of these components being sealed or intentionally covered to protect them from these factors. Debris (e.g., dust and hair) caught around the caster and drive wheel axles can harm the mechanical efficiency of the vehicle and may build up over time in unnoticeable amounts for the user. Regular maintenance (e.g., re-inflating pneumatic tires, checking for loose hardware, removing debris from the wheels and other moving parts) is highly recommended to maintain the optimal performance of any wheelchair.

Rigid frames showed a performance advantage over folding frames, but this advantage appears to be less than other influences of propulsion cost or propulsion effort. Prior research suggests that drive wheel tires, casters, and weight distribution are the most impactful influences on performance. Tire replacement is the most effective means of lowering propulsion cost in the existing literature. Swapping solid drive tires to more energy-efficient solid tires improved propulsion costs by up to 25% and up to 41% when the casters were likewise swapped with more energy-efficient casters [7]. Replacing solid tires with pneumatic tires improved propulsion costs by up to 54% [8]. Reducing pneumatic tire pressure from 100% of their recommended inflation value to 75% or 50% significantly increased the rolling resistance by 10% and 18%, respectively, using a coast-down test protocol [11]. In contrast, metabolic cost measurements using human subjects only showed significant differences after reducing the tire pressure from 100% to 50% [42] or 25% [10]. Solid tires had consistently worse performance than pneumatic tires at 25% of their recommended pressure [10]. Additionally, weight distribution can influence propulsion efficiency as well. Shifting the user's center of gravity rearward places more weight over the drive wheels. Changing the weight distribution from 55% or 60% of the total weight over the rear axle to 70% reduced propulsion torques by up to 13.5% [12] and lowered propulsion costs by 5% [8]. These configuration parameters (tires, casters, and weight distribution) have larger and more significant effects on MWC propulsion cost than the differences between rigid and folding frames.

Furthermore, a year's worth of simulated use does not appear to inhibit the mechanical efficiencies of either rigid or folding wheelchairs, when considering group effects. However, one folding chair exhibited fatigue that impacted performance. Clinicians are not provided with information about durability after expected use, but cannot assume that all chairs are equal. Ultimately, the user must make the decision between a folding and rigid frame. This study suggests that both frame types have the capability to yield excellent performance and durability.

Conclusions

Brand-new rigid ultra-lightweight frames were found to have superior performances to folding frames over concrete and carpet. Aging of the wheelchairs (i.e., frames, wheels, and tires) through 511 km of travel on a carousel track did not significantly affect the performance of either frame type. There is no generalizable evidence that one years’ worth of travel on ultra-lightweight wheelchairs has any significant effect on the propulsion cost. Worn rigid wheelchairs were superior to worn folding chairs over concrete. Performances were comparable over carpet, mostly due to the high energetic cost of propulsion over the surface overall. Simulated usage on the carousel track decreased the rolling resistance of the pneumatic drive tires, theorized to be from the decreased tread depth and flatter contact patch with the ground. This was an unexpected outcome and should not be generalized across all wheelchair tires. Neither rigid frame showed signs of deterioration and, in fact, both rigid chairs had slightly improved performances after the carousel track testing. Folding frames had equivalent (concrete) or comparable (carpet) performances before and after simulated use, and the four frames showed indications of repairable deterioration. Any potential performance improvements from the worn tires may have been mitigated by energy losses in the folding frames introduced by damage to the folding mechanism, vibrations in detachable parts, or skew in the frame caused by loosened fasteners. Average propulsion cost values for each frame type are representative of each group. However, individual wheelchairs from each frame type have slight variations in performances that can be studied in more detail. Further studies can expand upon this body of work by assessing a broader range of frames and extending the duration of simulated use testing. Carousel track testing is recommended as it imparts cornering forces and scrub torques to the tires, unlike multi-drum testers. Ultimately, the differences between frame types and between new or worn frame conditions have significance but are small in comparison to other configuration changes. For example, proper tire selection can have a much larger impact on the energy losses of manual wheelchairs than the difference between folding and rigid frames.

Footnote

Acknowledgment

The authors would like to thank Invacare Corporation for access to testing equipment. The authors would also like to thank Ki Mobility and Invacare Corporation for donating several of the wheelchair frames used in this study. This project was supported by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR) through Grant #: 90IFRE0036-01-02. NIDILRR is a Center within the United States Department of Health and Human Services (HHS) Administration for Community Living (ACL). The contents of this article do not necessarily reflect the views of the Department of Health and Human Services.

Conflict of Interest

There are no conflicts of interest. This article does not include research in which human participants were involved. Informed consent not applicable. This article does not include any research in which animal participants were involved.

Data Availability Statement

The data and information that support the findings of this article are freely available.1

References

1.
van der Woude
,
L. H. V.
,
Hendrich
,
K. M.
,
Veeger
,
H. E. J.
,
Van Ingen Schenau
,
G. J.
,
Rozendal
,
R. H.
,
De Groot
,
G.
, and
Hollander
,
A. P.
,
1988
, “
Manual Wheelchair Propulsion: Effects of Power Output on Physiology and Technique
,”
Med. Sci. Sports Exerc.
,
20
(
1
), pp.
70
78
.
2.
van der Woude
,
L. H. V.
, and
Veeger
,
H. E. J.
,
1988
, “
Wheelchair Racing: Effect of Rim Diameter and Speed on Physiology and Technique
,”
Med. Sci. Sports Exerc.
,
20
(
5
), pp.
492
500
.
3.
Boninger
,
M. L.
,
Dicianno
,
B. E.
,
Cooper
,
R. A.
,
Towers
,
J. D.
,
Koontz
,
A. M.
, and
Souza
,
A. L.
,
2003
, “
Shoulder Magnetic Resonance Imaging Abnormalities, Wheelchair Propulsion, and Gender
,”
Arch. Phys. Med. Rehabil.
,
84
(
11
), pp.
1615
1620
.
4.
van der Woude
,
L. H. V.
,
Dallmeijer
,
A. J.
,
Janssen
,
T. W. J.
, and
Veeger
,
D.
,
2001
, “
Alternative Modes of Manual Wheelchair Ambulation: An Overview
,”
Am. J. Phys. Med. Rehabil.
,
80
(
10
), pp.
765
777
.
5.
Curtis
,
K. A.
, and
Drysdale
,
G. A.
,
1999
, “
Shoulder Pain in Wheelchair Users With Tetraplegia and Paraplegia
,”
Arch. Phys. Med. Rehabil.
,
80
(
4
), pp.
453
457
.
6.
Boninger
,
M. L.
,
Impink
,
B. G.
,
Cooper
,
R. A.
, and
Koontz
,
A. M.
,
2004
, “
Relation Between Median and Ulnar Nerve Function and Wrist Kinematics During Wheelchair Propulsion
,”
Arch. Phys. Med. Rehabil.
,
85
(
7
), pp.
1141
1145
.
7.
Misch
,
J.
, and
Sprigle
,
S.
,
2021
, “
Effects of Wheels and Tires on High-Strength Lightweight Wheelchair Propulsion Cost Using a Robotic Wheelchair Tester
,”
Disabil. Rehabil.-Assist. Technol.
, pp.
1
11
.
8.
Sprigle
,
S.
, and
Huang
,
M.
,
2020
, “
Manual Wheelchair Propulsion Cost Across Different Components and Configurations During Straight and Turning Maneuvers
,”
J. Rehabil. Assist. Technol. Eng.
,
7
, pp.
1
14
.
9.
Sprigle
,
S.
,
Huang
,
M.
, and
Misch
,
J.
,
2019
, “
Measurement of Rolling Resistance and Scrub Torque of Manual Wheelchair Drive Wheels and Casters
,”
Assist. Technol.
,
34
(
1
) pp.
1
13
.
10.
de Groot
,
S.
,
Vegter
,
R. J.
, and
van der Woude
,
L. H.
,
2013
, “
Effect of Wheelchair Mass, Tire Type and Tire Pressure on Physical Strain and Wheelchair Propulsion Technique
,”
Med. Eng. Phys.
,
35
(
10
), pp.
1476
1482
.
11.
Lin
,
J. T.
,
Huang
,
M.
, and
Sprigle
,
S.
,
2015
, “
Evaluation of Wheelchair Resistive Forces During Straight and Turning Trajectories Across Different Wheelchair Configurations Using Free-Wheeling Coast-Down Test
,”
J. Rehabil. Res. Dev.
,
52
(
7
), pp.
763
774
.
12.
Sprigle
,
S.
, and
Huang
,
M.
,
2015
, “
Impact of Mass and Weight Distribution on Manual Wheelchair Propulsion Torque
,”
Assist. Technol.
,
27
(
4
), pp.
226
235
; quiz 236-7.
13.
Misch
,
J.
,
Huang
,
M.
, and
Sprigle
,
S.
,
2020
, “
Modeling Manual Wheelchair Propulsion Cost During Straight and Curvilinear Trajectories
,”
PLoS One
,
15
(
6
), p.
e0234742
.
14.
Sagawa
,
Y.
,
Watelain
,
E.
,
Lepoutre
,
F.-X.
, and
Thevenon
,
A.
,
2010
, “
Effects of Wheelchair Mass on the Physiologic Responses, Perception of Exertion, and Performance During Various Simulated Daily Tasks
,”
Arch. Phys. Med. Rehabil.
,
91
(
8
), pp.
1248
1254
.
15.
Fitzgerald
,
S. G.
,
Cooper
,
R. A.
,
Boninger
,
M. L.
, and
Rentschler
,
A. J.
,
2001
, “
Comparison of Fatigue Life for 3 Types of Manual Wheelchairs
,”
Arch. Phys. Med. Rehabil.
,
82
(
10
), pp.
1484
1488
.
16.
Gebrosky
,
B.
,
Bridge
,
A.
,
O’Donnell
,
S.
,
Grindle
,
G. G.
,
Cooper
,
R.
, and
Cooper
,
R. A.
,
2020
, “
Comparing the Performance of Ultralight Folding Manual Wheelchairs Using Standardized Tests
,”
Disabil. Rehabil. Assist Technol.
,
17
(
1
), pp.
40
49
.
17.
Sonenblum
,
S. E.
, and
Sprigle
,
S.
,
2017
, “
Wheelchair use in Ultra-Lightweight Wheelchair Users
,”
Disabil. Rehabil. Assist. Technol.
,
12
(
4
), pp.
396
401
.
18.
Sonenblum
,
S. E.
,
Sprigle
,
S.
, and
Lopez
,
R. A.
,
2012
, “
Manual Wheelchair Use: Bouts of Mobility in Everyday Life
,”
Rehabil. Res. Pract.
,
2012
, pp.
1
7
.
19.
Pacejka
,
H.
,
2005
,
Tire and Vehicle Dynamics
,
Elsevier
,
New York
.
20.
Chénier
,
F.
, and
Aissaoui
,
R.
,
2014
, “
Effect of Wheelchair Frame Material on Users’ Mechanical Work and Transmitted Vibration
,”
Biomed. Res. Int.
,
2014
(
12
), pp.
1
12
.
21.
Cowan
,
R. E.
,
Nash
,
M. S.
,
Collinger
,
J. L.
,
Koontz
,
A. M.
, and
Boninger
,
M. L.
,
2009
, “
Impact of Surface Type, Wheelchair Weight, and Axle Position on Wheelchair Propulsion by Novice Older Adults
,”
Arch. Phys. Med. Rehabil.
,
90
(
7
), pp.
1076
1083
.
22.
Vegter
,
R. J.
,
Lamoth
,
C. J.
,
de Groot
,
S.
,
Veeger
,
D. H.
, and
van der Woude
,
L. H.
,
2013
, “
Variability in Bimanual Wheelchair Propulsion: Consistency of two Instrumented Wheels During Handrim Wheelchair Propulsion on a Motor Driven Treadmill
,”
J. Neuroeng. Rehabil.
,
10
(
1
), pp.
9
.
23.
Hurd
,
W. J.
,
Morrow
,
M. M.
,
Kaufman
,
K. R.
, and
An
,
K.-N.
,
2009
, “
Wheelchair Propulsion Demands During Outdoor Community Ambulation
,”
J. Electromyogr. Kinesiol.
,
19
(
5
), pp.
942
947
.
24.
Koontz
,
A. M.
,
Roche
,
B. M.
,
Collinger
,
J. L.
,
Cooper
,
R. A.
, and
Boninger
,
M. L.
,
2009
, “
Manual Wheelchair Propulsion Patterns on Natural Surfaces During Start-Up Propulsion
,”
Arch. Phys. Med. Rehabil.
,
90
(
11
), pp.
1916
1923
.
25.
Ott
,
J.
,
Wilson-Jene
,
H.
,
Koontz
,
A.
, and
Pearlman
,
J.
,
2020
, “
Evaluation of Rolling Resistance in Manual Wheelchair Wheels and Casters Using Drum-Based Testing
,”
Disabil. Rehabil. Assist. Technol.
,
17
(
6
), pp.
1
12
.
26.
Fallot
,
C.
, and
Bascou
,
J.
,
2019
, “
Manual Wheelchair's Turning Resistance: Swivelling Resistance Parameters of Front and Rear Wheels on Different Surfaces
,”
Disabil. Rehabil. Assist. Technol.
,
16
(
3
), pp.
1
8
.
27.
International Organization for Standardization
,
2012
, “
Wheelchairs—Part 11: Test Dummies
,” ISO 7176-11:2012.
28.
Liles
,
H.
,
Huang
,
M.
,
Caspall
,
J.
, and
Sprigle
,
S.
,
2015
, “
Design of a Robotic System to Measure Propulsion Work of Over-Ground Wheelchair Maneuvers
,”
IEEE Trans. Neural Syst. Rehabil. Eng.
,
23
(
6
), pp.
983
991
.
29.
Wolf
,
E.
,
Pearlman
,
J.
,
Cooper
,
R. A.
,
Fitzgerald
,
S. G.
,
Kelleher
,
A.
,
Collins
,
D. M.
,
Boninger
,
M. L.
, and
Cooper
,
R.
,
2005
, “
Vibration Exposure of Individuals Using Wheelchairs Over Sidewalk Surfaces
,”
Disabil. Rehabil.
,
27
(
23
), pp.
1443
1449
.
30.
International Organization for Standardization
,
1997
,
ISO 2631 Mechanical Vibration and Shock—Section 1: Evaluation of Human Exposure to Whole-Body Vibration, Section 1: Guide for the Evaluation of Human Exposure to Whole-Body Vibration
.
31.
VanSickle
,
D. P.
, and
Cooper
,
R. A.
,
2001
, “
Analysis of Vibrations Induced During Wheelchair Propulsion
,”
J. Rehabil. Res. Dev.
,
38
(
4
), pp.
409
421
.
32.
Cooper
,
R. A.
,
Wolf
,
E.
,
Fitzgerald
,
S. G.
,
Kellerher
,
A.
,
Ammer
,
W.
,
Boninger
,
M. L.
, and
Cooper
,
R.
,
2004
, “
Evaluation of Selected Sidewalk Pavement Surfaces for Vibration Experienced by Users of Manual and Powered Wheelchairs
,”
J. Spinal Cord Med.
,
27
(
5
), pp.
468
475
.
33.
Maeda
,
S.
,
Futatsuka
,
M.
,
Yonesaki
,
J.
, and
Ikeda
,
M.
,
2003
, “
Relationship Between Questionnaire Survey Results of Vibration Complaints of Wheelchair Users and Vibration Transmissibility of Manual Wheelchair
,”
Environ. Health Prev. Med.
,
8
(
3
), pp.
82
89
.
34.
Boninger
,
M. L.
,
Cooper
,
R. A.
,
Fitzgerald
,
S. G.
,
Lin
,
J.
,
Cooper
,
R.
,
Dicianno
,
B.
, and
Liu
,
B.
,
2003
, “
Investigating Neck Pain in Wheelchair Users
,”
Am. J. Phys. Med. Rehabil.
,
82
(
3
), pp.
197
202
.
35.
Hansson
,
T.
,
Magnusson
,
M.
, and
Broman
,
H.
,
1991
, “
Back Muscle Fatigue and Seated Whole Body Vibrations: an Experimental Study in man
,”
Clin. Biomech.
,
6
(
3
), pp.
173
178
.
36.
Pope
,
M. H.
,
Wilder
,
D. G.
, and
Magnusson
,
M. L.
,
1999
, “
A Review of Studies on Seated Whole Body Vibration and low Back Pain
,”
Proc. Inst. Mech. Eng. H
,
213
(
6
), pp.
435
446
.
37.
Garcia-Mendez
,
Y.
,
Pearlman
,
J. L.
, and
Cooper
,
R. A.
,
2013
, “
Health Risks of Vibration Exposure to Wheelchair Users in the Community
,”
J. Spinal Cord Med.
,
36
(
4
), pp.
365
375
.
38.
Wellek
,
S.
,
2010
,
Testing Statistical Hypotheses of Equivalence and Noninferiority
, 2nd ed.,
CRC Press
,
Boca Raton, FL,
xvi, p.
415
.
39.
Lui
,
J.
,
MacGillivray
,
M. K.
,
Sheel
,
A. W.
,
Jeyasurya
,
J.
,
Sadeghi
,
M.
, and
Sawatzky
,
B. J.
,
2013
, “
Mechanical Efficiency of two Commercial Lever-Propulsion Mechanisms for Manual Wheelchair Locomotion
,”
J. Rehabil. Res. Dev.
,
50
(
10
), pp.
1363
1372
.
40.
Pavlidou
,
E.
,
Kloosterman
,
M. G.
,
Buurke
,
J. H.
,
Rietman
,
J. S.
, and
Janssen
,
T. W. J.
,
2015
, “
Rolling Resistance and Propulsion Efficiency of Manual and Power-Assisted Wheelchairs
,”
Med. Eng. Phys.
,
37
(
11
), pp.
1105
1110
.
41.
Hilbers
,
P. A.
, and
White
,
T. P.
,
1987
, “
Effects of Wheelchair Design on Metabolic and Heart Rate Responses During Propulsion by Persons With Paraplegia
,”
Phys. Ther.
,
67
(
9
), pp.
1355
1358
.
42.
Sawatzky
,
B. J.
,
Kim
,
W. O.
, and
Denison
,
I.
,
2004
, “
The Ergonomics of Different Tyres and Tyre Pressure During Wheelchair Propulsion
,”
Ergonomics
,
47
(
14
), pp.
1475
1483
.
43.
Bascou
,
J.
, and
Sauret
,
C.
,
2017
, “
Is Bearing Resistance Negligible During Wheelchair Locomotion? Design and Validation of a Testing Device
,”
Acta Bioeng. Biomech.
,
19
(
3
), pp.
165
176
.
44.
Sandberg
,
U.
, and
Glaeser
,
K.-P.
,
2008
, “
The Influence of Tyre Wear and Ageing on Tyre/Road Noise Emission and Rolling Resistance
,” G.C.E. Sustainable Development, Editor. European Commission, DG Research.
45.
Martini
,
M. E.
,
1983
, “
Passenger Tire Rolling Loss: A Tread Compounding Approach and Its Tradeoffs
,”
Rubber Chem. Technol.
,
56
(
2
), pp.
507
508
.
46.
Schuring
,
D. J.
,
1980
, “
The Rolling Loss of Pneumatic Tires
,”
Rubber Chem. Technol.
,
53
(
3
), pp.
600
727
.
47.
National Research Council (U.S.)
,
2006
, “
Rolling Resistance, Traction, and Wear Performance of Passenger Tires
,” Tires and Passenger Vehicle Fuel Economy : Informing Consumers, Improving Performance—Special Report 286,
Transportation Research Board, Committee for the National Tire Efficiency Study, Transportation Research Board of the National Academies
,
Washington, DC
, pp.
77
103
.
48.
Gebrosky
,
B.
,
Grindle
,
G.
,
Cooper
,
R.
, and
Cooper
,
R.
,
2020
, “
Comparison of Carbon Fibre and Aluminium Materials in the Construction of Ultralight Wheelchairs
,”
Disabil. Rehabil. Assist. Technol.
,
15
(
4
), pp.
432
441
.
49.
Gebrosky
,
B.
,
Pearlman
,
J.
, and
Cooper
,
R.
,
2018
, “
Comparison of High-Strength Aluminum Ultralight Wheelchairs Using ANSI/RESNA Testing Standards
,”
Top Spinal Cord Inj. Rehabil.
,
24
(
1
), pp.
63
77
.
50.
Kwarciak
,
A. M.
,
Cooper
,
R. A.
, and
Fitzgerald
,
S. G.
,
2008
, “
Curb Descent Testing of Suspension Manual Wheelchairs
,”
J. Rehabil. Res. Dev.
,
45
(
1
), pp.
73
84
.
51.
Toro
,
M. L.
,
Worobey
,
L.
,
Boninger
,
M. L.
,
Cooper
,
R. A.
, and
Pearlman
,
J.
,
2016
, “
Type and Frequency of Reported Wheelchair Repairs and Related Adverse Consequences Among People With Spinal Cord Injury
,”
Arch. Phys. Med. Rehabil.
,
97
(
10
), pp.
1753
1760
.
52.
Wang
,
H.
,
Liu
,
H. Y.
,
Pearlman
,
J.
,
Cooper
,
R.
,
Jefferds
,
A.
,
Connor
,
S.
, and
Cooper
,
R. A.
,
2010
, “
Relationship Between Wheelchair Durability and Wheelchair Type and Years of Test
,”
Disabil. Rehabil. Assist. Technol
,
5
(
5
), pp.
318
322
.
53.
Van Oers
,
C. A. J. M.
,
van der Woude
,
L. H. V.
,
Roebroeck
,
M.
, and
Rozendal
,
R.H.
,
1992
, “
A Failure Diary for Hand-Propelled Wheelchairs
,”
J. Rehabil. Sci.
,
4
(
4
), pp.
123
130
.