IT IS essential that the design of electric powered wheelchairs

805 Durability, Value, and Reliability of Selected Electric Powered Wheelchairs Megan V. Fass, MS, Rory A. Cooper, PhD, Shirley G. Fitzgerald, PhD, Mark Schmeler, MS, OTR/L, ATP, Michael L. Boninger, MD, S. David Algood, BS, William A. Ammer, BS, Andrew J. Rentschler, MS, John Duncan, BS ABSTRACT. Fass MV, Cooper RA, Fitzgerald SG, Schmeler M, Boninger ML, Algood SD, Ammer WA, Rentschler AJ, Duncan J. Durability, value, and reliability of selected electric powered wheelchairs. Arch Phys Med Rehabil 2004;85:805-14. Objective: To compare the durability, value, and reliability of selected electric powered wheelchairs (EPWs), purchased in 1998. Design: Engineering standards tests of quality and performance. Setting: A rehabilitation engineering center. Specimens: Fifteen EPWs: 3 each of the Jazzy, Quickie, Lancer, Arrow, and Chairman models. Interventions: Not applicable. Main Outcome Measures: Wheelchairs were evaluated for durability (lifespan), value (durability, cost), and reliability (rate of repairs) using 2-drum and curb-drop machines in accordance with the standards of the American National Standards Institute and Rehabilitation Engineering and Assistive Technology Society of North America. Results: The 5 brands differed significantly (P.05) in durability, value, and reliability, except in terms of reliability of supplier repairs. The Arrow had the highest durability, value, and reliability in terms of the number of consumer failures, supplier failures, repairs, failures, consumer repairs and failures, and supplier repairs and failures. The Lancer had the poorest durability and reliability, and the Chairman had the lowest value. K0014 wheelchairs (Arrow, Permobil) were significantly more durable than K0011 wheelchairs (Jazzy, Quickie, Lancer). No significant differences in durability with respect to rear-wheel drive (Arrow, Lancer, Quickie), midwheel drive (Jazzy), or front-wheel drive (Chairman) wheelchairs were found. Conclusions: The Arrow consistently outperformed the other wheelchairs in nearly every area studied, and K0014 wheelchairs were more durable than K0011 wheelchairs. These From the Departments of Rehabilitation Science & Technology (Fass, Cooper, Fitzgerald, Schmeler, Boninger, Algood, Ammer, Rentschler), Physical Medicine & Rehabilitation (Cooper, Fitzgerald, Schmeler, Boninger, Ammer), and Bioengineering, University of Pittsburgh (Cooper, Boninger, Rentschler); and Human Engineering Research Laboratories, a VA Rehabilitation Research & Development Center, VA Pittsburgh Healthcare System (Fass, Fitzgerald, Schmeler, Boninger, Algood, Ammer, Rentschler, Duncan), Pittsburgh, PA. Supported in part by the Paralyzed Veterans of America, National Institute of Disability and Rehabilitation Research (grant no. H133E990001), Rehabilitation Services Administration (grant no. H129E990004), US Department of Education, VA Rehabilitation Research & Development Service, and US Department of Veterans Affairs (grant no. F2181C). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Rory A. Cooper, PhD, Human Engineering Research Laboratories (151-R1), VA Pittsburgh Healthcare System, 7180 Highland Dr, Pittsburgh, PA 15206, e-mail: rcooper @pitt.edu. 0003-9993/04/8505-8348$30.00/0 doi:10.1016/j.apmr.2003.08.096 results can be used as an objective comparison guide for clinicians and consumers, as long as they are used in conjunction with other important selection criteria. Manufacturers can use these results as a guide for continued efforts to produce higher quality wheelchairs. Care should be taken when making comparisons, however, because the 5 brands had different features. Purchased in 1998, these models may be used for several more years. In addition, problem areas in these models may still be present in newer models. Key Words: Fatigue; Life cycle; Reference standards; Rehabilitation; Wheelchairs. 2004 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation IT IS essential that the design of electric powered wheelchairs (EPWs) reflect the needs and abilities of the consumers who use them. Researchers surveyed a panel of mobility consumer experts using a modified version of the Delphi method to determine what qualities were most important to users of wheeled mobility. 1 The panel first listed factors that were deemed important and then ranked them in order of importance. Fifteen factors were important in powered mobility, with effectiveness, operability, dependability, affordability, and personal acceptance being the top-ranked priorities. Dependable and durable products that can be readily repaired and maintained by either the supplier or consumer are critical to EPW users. Studies indicate that EPWs can be unreliable, which increases the risk to the user. Gaal et al 2 interviewed 109 active wheelchair users who had experienced an injury while in a wheelchair. Fifty-three percent of injuries were associated with powered mobility, with 42% of the injuries dealing with powered mobility affiliated with component failures. Kirby and Ackroyd-Stolarz 3 evaluated the US Food and Drug Administration database between 1975 and 1993 for incidences of injuries among wheelchair users. They found that over 60% of injuries in EPW users were attributed to engineering factors. Standards are among the most important means to minimize risk and to improve the reliability of products. 4 The International Standards Organization (ISO), the American National Standards Institute (ANSI), and the Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Wheelchair Standards Committee have developed standards dealing with performance, safety, and dimensions for wheelchairs. The ANSI/RESNA standards are used primarily in the United States and the ISO standards are used worldwide. Because ISO and ANSI/RESNA have collaborated closely, the standards are virtually identical. 5-7 The voluntary standards are a set of test procedures or instructions that detail how to perform various tests or measurements on the wheelchair. One section of the standards tests the durability of the wheelchair. 5 In the fatigue test, the EPWs are placed on a pair of rotating drums, called a 2-drum machine, for 200,000 cycles. Each drum has a pair of slats attached that are designed to simulate

806 LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass Table 1: Wheelchair Characteristics Feature Jazzy Quickie Lancer Arrow Chairman Price (US $) 4325 5345 5437 7109 13,441 HCPCS K0011 K0011 K0011 K0014 K0014 Drive Mid-wheel Rear-wheel Rear-wheel Rear-wheel Front-wheel Seat Captain Foam Foam Foam Contour Seat pan Integrated Sling Rigid Rigid Rigid Back Captain Sling Sling Sling Contour Armrest Flip-up Removable Removable Removable Flip-up Footrest 1, flip-up 2, swing-away 2, swing-away 2, swing-away 1, with 2 plates, flip-up Motor Reliance Stature Fracmo Stature Levoy-Somer Electronics Penny & Giles Penny & Giles Penny & Giles MKIV Penny & Giles Other Power recline Abbreviation: HCPCS, Health Care Finance Administration Common Procedure Codes. use over uneven terrain. The EPWs are then placed on a curb-drop machine, which lifts the wheelchairs to a certain height and drops them repetitively for 6666 drops. The curbdrop machine is designed to simulate the user driving down curbs. Standards do not require disclosure of most of the results in manufacturer s product literature. 5 Consumers or clinicians can obtain the information if requested, but this is tedious and makes comparisons difficult. In addition, to pass the ANSI/ RESNA standards, the EPW must be tested to 200,000 twodrum cycles and 6666 curb-drop cycles or to the number of cycles that the manufacturer claims that the EPW can exceed. 5 ANSI and RESNA expected that market pressures would elevate the values to which manufacturers test their products, but there is no evidence that this has happened in the more than 13 years since the ANSI/RESNA standards were approved. This may be having a negative impact on the quality of wheelchairs provided within the United States. In addition, testing EPWs until the end of their useful life can provide valuable information about durability (the length of time until replacement is required), reliability (the rate of repairs), and the types of repairs and failures (the number and types of broken components). When factoring in the life of the wheelchair into the initial purchase price, the financial value can also be informative. For example, ultralight manual wheelchairs typically are more expensive to purchase than lightweight and depot manual wheelchairs. When factoring in the durability when tested on the 2-drum and curb-drop machines, ultralight wheelchairs are actually more cost effective than lightweight and depot wheelchairs (in terms of dollars per life cycle). 8,9 The purpose of this study was to determine the durability, value, and reliability of selected EPWs using ANSI/RESNA standards tests. The following hypotheses were evaluated: 1. Durability: The total number of equivalent cycles completed (until the first failure and until no longer operable) differs significantly among brands of wheelchairs. The total number of equivalent cycles completed (until no longer operable) differs significantly between K0011 and K0014 EPWs and between rear-wheel, mid-wheel, and front-wheel drive EPWs. 2. Value: The total number of equivalent cycles (until the first failure and until no longer operable) per average retail purchase price differs significantly among brands. 3. Reliability: The mean equivalent cycles between consumer repairs, supplier repairs, consumer failures, supplier failures, consumer repairs/failures, and supplier repairs/failures are significantly different between the brands. The results of this study may provide a means for objective comparison between the selected EPWs and provide data for assessing durability, value, and reliability for comparing other types of wheelchairs or designs. Clinicians and consumers can use these results for both product comparisons and for funding justification purposes, in conjunction with other components that make up an evaluation. The results can also help manufacturers improve their products, thereby increasing the reliability of EPWs and reducing the risk to users. METHODS Three identical EPWs from 5 different manufacturers (n 15) were purchased in 1998, without the manufacturers knowledge of the study. The wheelchairs were selected based on their high frequency of purchase by the Veteran s Health Administration (VHA). This included the Pride Mobility Jazzy 1100, a Sunrise Medical Quickie P200, b Everest and Jennings Lancer 2000, c Invacare Action Arrow, d and Permobil Chairman. e All wheelchairs purchased were the standard base models, meaning that no extra features were ordered. However, the EPWs had different features (table 1). The Chairman standard model came with a powered seat recline feature, which was not present in the other wheelchairs. In the United States, codes used for billing purposes, called Health Care Finance Administration Common Procedure Codes (HCPCS), are associated with a product type and the allowable amount of money that a supplier can receive for the wheelchair. 10 The Jazzy, Quickie, and Lancer fall under the K0011 code, and the Arrow and Chairman fall under the K0014 code. 10 K0011 wheelchairs are standard weight/frame powered wheelchairs, with programmable control parameters. K0014 wheelchairs are wheelchairs with other power wheelchair bases. 10 The K0011 code is designed for the average wheelchair user, whereas K0014 is designed for users who require special features that are not readily available in other models. 10 Possible examples of users who require K0014 wheelchairs are persons who weigh more than 112.5kg (250lb), are active outdoors, use a ventilator tray, need power tilt or recline, or require other alternative controls. All sections of the ANSI/RESNA standard tests were completed for other research purposes before commencement of this study. These tests include determination of static stability; dynamic stability; energy consumption; overall dimensions, weight, and turning space; maximum speed, acceleration, and retardation; seating and wheel dimensions; static, impact, and fatigue strength; climactic tests; obstacle-climbing ability; and testing of power and control systems. 5 Two-drum and curbdrop tests were then conducted in accordance with section 8 of the ANSI/RESNA standards (fig 1). 5 The 2-drum machine

LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass 807 Fig 1. The 2-drum (left) and curb-drop (right) machines. The 2-drum machine has 2 rollers with slats attached, which are designed to simulate driving over terrain. The curb-drop machine lifts and drops the EPW using 4 chains, which is designed to simulate driving down curbs. consists of 2 drums with four 1-cm (0.4-in) slats attached, with the slats offset at 180 from each other. The front drum turned 7% faster than the rear drum to ensure that the slats hit against the wheels and casters at different rates. Foam was placed on the seats and backs, and the wheelchairs were loaded with a 99-kg (220-lb) test dummy and were secured to the 2-drum machine, as specified in the standards. Tire pressure was monitored and was retained within the manufacturers recommended ratings. The joystick was adjusted so that the wheelchair ran at approximately 1m/s (2.2mph). The wheelchairs were tested on the 2-drum machine for 200,000 cycles, where 1 cycle is equal to 1 revolution of the rear drum. 5 During the 2-drum tests, the EPWs were powered by a direct-current power supply, although the batteries remained in the wheelchair as ballast. The wheelchairs were then attached to the curb-drop machine via chains and repetitively dropped from a height of 5cm (2in). The wheelchairs were tested on the curbdrop machine for 6666 cycles, where 1 cycle is equal to 1 drop from the required height. 5 The combined 200,000 cycles on the 2-drum machine and 6666 cycles on the curb-drop machine approximate 3 to 5 years of typical wheelchair use, 11 the minimum amount required to pass the standards. 5 Equivalent cycles (equation 1) are the combined cycles on the 2-drum and curb-drop machines, based on the formula presented in the ANSI/RESNA standards that 1 curb-drop cycle is comparable to thirty 2-drum cycles. 5,9 Equivalent cycles (EC) (number of 2-drum cycles) 30 (number of curb-drop cycles) (1) The wheelchairs were monitored closely, and any problems were noted. The total number of 2-drum and curb-drop cycles completed when each problem was detected was recorded. Efforts were made to ensure that the testing procedures did not damage the wheelchair or interfere with the normal behavior of the wheelchair. If there were instances in which the testing procedure damaged the wheelchair, these data were excluded (eg, there was an instance in which the chain to the curbdrop machine rubbed against an armrest, which caused some damage). Inoperability Criteria All wheelchairs were alternated between the 2-drum and curb-drop machines until the inoperability criteria were met. The inoperability criteria were developed because the ANSI/ RESNA standards do not stipulate that testing of EPWs be continued after the ANSI/RESNA failure criteria 5 (table 2) have been met. Important information about durability, value, and reliability would be lost. Clinicians, in consultation with wheelchair suppliers, typically make judgment calls about whether to recommend repairs or replacement wheelchairs. Factors that affect these decisions include the age of the wheelchair, repair history, condition of the wheelchair, cost of the repair, and the needs of the wheelchair user. 12 The VHA will approve a replacement wheelchair when the repair costs exceed 50% of the replacement costs. 13 As a matter of consistency, testing was continued until the wheelchairs had motor malfunctions, electronic malfunctions, or frame fractures. These were problems that rendered the wheelchairs inoperable. In addition, problems that met the inoperability criteria also met the ANSI/ RESNA failure criteria. All other malfunctions or maladjustments were relatively minor, and testing continued until the inoperability criteria were met. Engineers who work in our laboratories conducted the repairs and determined when the inoperability criteria were met. Motor and electronic failures were determined by replacing the part with a known working component. Frame fractures were determined by visual inspection. As a matter of practicality, 1 Arrow chair test was terminated at 4,400,100 equivalent cycles, even though the wheelchair was still operable, because it lasted more than twice as long as the second most durable wheelchair. Repairs and Failures Labels were developed for clarity purposes for determining when the ANSI/RESNA failure criteria were met. Problems that met the ANSI/RESNA failure criteria were labeled as failures. Problems that did not meet the ANSI/RESNA failure criteria were labeled as repairs. Consumer Repairs, Supplier Repairs, Consumer Failures, and Supplier Failures Problems were categorized in terms of consumer repairs, supplier repairs, consumer failures, and supplier failures. These labels were developed so that clinicians, consumers, and manufacturers can better understand the level of severity of prob- Table 2: Failure Criteria Based on Current ANSI/RESNA Standards 5 1) No component shall be fractured or have visible cracks. 2) No nut, bolt, screw, locking pin, adjustable component, or similar item shall become detached after having been tightened, adjusted, or refitted once during testing. 3) No electric connector shall be displaced or disconnected. 4) All parts intended to be removable, folding, or adjustable shall operate as described by the manufacturer. 5) All power-operated systems shall operate as described by the manufacturer. 6) Handgrips shall not be displaced. 7) Any multiposition or adjustable component shall not be displaced from the preset position. 8) No component or assembly of parts shall exhibit deformation, free play, or loss of adjustment that could adversely affect the function of the wheelchair.

808 LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass lems that occur in EPWs. These labels are based on earlier ANSI/RESNA standards failure criteria, 14 which classified problems in terms of whether an untrained or trained person is required for repairs. Consumer repairs and consumer failures were problems considered appropriate for wheelchair users or untrained technicians to repair. Supplier repairs and supplier failures were problems classified as appropriate for trained technicians or suppliers to repair and/or required the purchase of parts that could not easily be obtained from a local hardware store. Previous ANSI/RESNA standards 14 categorized failures according to class I, II, and III failures, instead of the current combined ANSI/RESNA failure criteria 5 shown in table 2. For comparison purposes with EPWs that have already been tested according to the older standards, it is helpful to view problems in terms of the class I, II, and III failure criteria. Consumer repairs are analogous to class I failures from the older standards, where the wheelchair required maintenance that would have normally been carried out by a person with no special skill. 14 Examples of consumer repairs include replacing or tightening loose bolts and nuts or adding air to tires. Consumer repairs are considered relatively minor inconveniences, but not having to make such repairs for a long time is preferable. Supplier repairs are analogous to class II failures, which were defined as repairs that would have normally been carried out by an agent, garage, etc. 14 Examples of supplier repairs include fixing flat tires or replacing footrest parts and seat mounting hardware. Supplier repairs can inconvenience or endanger the consumer (eg, flat tires can leave the wheelchair user stranded in inopportune situations). Ordering replacement parts requires the service of the supplier and time to receipt of parts may be lengthy. Consumer and supplier failures are analogous to class III failures. A class III failure by definition is a structural failure, free play or loosening in the frame or its attachments that cannot readily be fixed, or deformation or maladjustment of any part of the wheelchair or its attachments that adversely affects its function. 14 Examples of consumer failures include tightening, adjusting, or refitting the same bolts to footrests, rear bumpers, shrouds more than once and reconnecting an electric connector that was readily apparent and the replacement location was obvious. Examples of supplier failures include motor or gearbox malfunctions, wheelchair veering to the left or right, melted wiring, malfunctioning joystick or controller, brakes intermittently not engaging, cracked seat post, frame fracture, and disconnected electric connectors that cannot easily be located and reconnected. Durability, Value, and Reliability After the EPWs were inoperable, durability, value, and reliability were determined. Durability the number of equivalent cycles completed (until first failure and until no longer operable) was determined first. The number of equivalent cycles completed (until no longer operable) for the K0011 (Jazzy, Quickie, Lancer) and K0014 (Arrow, Chairman) wheelchairs, as well as for the rear-wheel drive (Arrow, Lancer, Quickie), mid-wheel drive (Jazzy), and front-wheel drive (Chairman) wheelchairs was also determined. Value (equations 2, 3) takes into account the number of equivalent cycles completed and the initial cost of the wheelchairs, which was based on the retail purchase price obtained from 3 local wheelchair vendors. Value until first failure Total number of EC completed until first failure (2) Average retail purchase price for wheelchair model Value until no longer operable Total number of EC completed until no longer operable Average retail purchase price for wheelchair model (3) Reliability takes into account both the number of cycles successfully completed between each adjustment and malfunction and the total number of adjustments and malfunctions that occurred. The mean equivalent cycles between consumer repairs, supplier repairs, consumer failures, supplier failures, repairs, failures, consumer repairs and failures, and supplier repairs and failures were determined according to the following formula: Reliability Mean EC between incidences 1 N number of EC completed at nth incidence number of EC completed at n 1th incidence (4) Total number of incidences where N is the total number of incidences and an incidence is an occurrence of consumer repairs, supplier repairs, etc. In some incidences, the wheelchair failed to maintain proper speed, but the problem typically was intermittent. To give the wheelchair the benefit of the doubt, the problem had to exist for at least 10,000 continuous cycles on the 2-drum machine before the wheelchair was failed. For the wheelchairs that had no consumer repairs, supplier repairs, consumer failures, consumer repairs and failures, or repairs, testing could not logistically be continued after the wheelchairs were already considered inoperable. Important information about reliability would be missing and would unfairly classify the wheelchairs with minimal problems as unreliable. In these cases, it was assumed that the corresponding repairs or failures occurred at the time when the wheelchairs were rendered inoperable. Statistical Analysis Nonparametric (Kruskal-Wallis, Mann-Whitney) statistics were used for the number of equivalent cycles completed until the wheelchairs were no longer operable (for all explorations) and the mean equivalent cycles between consumer failures, because the data were not normally distributed. Parametric (1-way analysis of variance, Bonferroni adjustment) statistics were used for all other variables. An level of.05 was used for all statistical analysis. RESULTS Durability Table 3 and figure 2A show the number of equivalent cycles completed until the first occurrence of a failure and the number of equivalent cycles completed until the wheelchairs were no longer operable. When looking at the number of equivalent cycles completed until the first occurrence of a failure, the Arrow lasted 2.0 times longer than the Jazzy, 3.3 times longer than the Quickie, 3.8 times longer than the Chairman, and 5.0 times longer than the Lancer. When looking at the number of equivalent cycles completed until the wheelchairs were no longer operable, the Arrow lasted 2.2 times longer than the Jazzy, 4.1 times longer than the Chairman, 6.5 times longer than the Quickie, and 9.9 times longer than the Lancer.

LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass 809 Table 3: Durability and Value Brand (N 3) Durability* (EC) Durability (EC) Value* (EC/dollar) Value (EC/dollars) Jazzy 714,032 276,590 1,256,061 457,737 165 64 290 106 Quickie 425,472 19,558 425,472 19,558 80 4 80 4 Lancer 276,981 219,337 276,981 219,337 51 40 51 40 Arrow 1,395,222 421,368 2,752,869 1,452,028 196 59 387 204 Chairman 363,987 152,882 677,038 152,864 27 9 50 11 NOTE. Values are mean standard deviation (SD). *Until first failure; until no longer operable. The 5 brands differed significantly (P.01) in the number of equivalent cycles completed until the first failure, with the Arrow being significantly greater (P.01) than the Quickie, Chairman, and Lancer. The 5 brands also differed significantly (P.05) in the number of equivalent cycles completed until the wheelchairs were no longer operable, with the Arrow and Jazzy being significantly greater (P.05) than the Chairman, Quickie, and Lancer and the Chairman being significantly greater (P.05) than the Quickie and Lancer. Further analysis also showed significant differences (P.05) in the durability until no longer operable, based on their HCPCS codes. The K0014 wheelchairs (1,714,953 1,464,727 equivalent cycles) lasted significantly longer (P.05) than the K0011 wheelchairs (652,837 522,799 equivalent cycles). There were no significant differences between the rear-wheel drive, mid-wheel drive, and front-wheel drive wheelchairs in terms of durability. Value The value (until first failure) for the Arrow was 1.2 times higher than that for the Jazzy, 2.5 times higher than that for the Quickie, 3.9 times higher than that for the Lancer, and 7.3 times higher than that for the Chairman. The value (until no longer operable) for the Arrow was 1.3 times higher than that for the Jazzy, 4.9 times higher than that for the Quickie, 7.6 times higher than that for the Lancer, and 7.7 times higher than that for the Chairman. When looking at the number of equivalent cycles completed until the first occurrence of a failure, the 5 brands differed significantly (P.01) in value (fig 2B). The Arrow was significantly higher (P.05) than the Lancer and Chairman, and the Jazzy was significantly higher (P.05) than the Chairman. When looking at the number of equivalent cycles completed until the wheelchairs were no longer operable, the 5 brands also differed significantly (P.01) in value, with the Arrow being significantly higher (P.05) than the Quickie, Lancer, and Chairman. Consumer Repairs, Supplier Repairs, Consumer Failures, and Supplier Failures Table 4 shows where at least 1 instance of consumer repairs, supplier repairs, consumer failures, and supplier failures occurred. The boldface in table 4 indicates an where either a repair or failure occurred within the first 400,000 equivalent cycles (the minimum number of cycles required for successful passing of the ANSI/RESNA standards 5 ). The wheelchairs were rendered inoperable because (1) 3 Quickies failed to maintain proper speed; (2) 1 Lancer had a motor/gearbox failure, and 2 Lancers failed to maintain proper speed; (3) 1 Chairman had a seat post crack, 1 Chairman had a frame fracture located underneath the wheelchair, and 1 Chairman failed to maintain proper speed; (4) 1 Jazzy had a motor/ gearbox failure, 1 Jazzy had a joystick/controller failure, and 1 Jazzy failed to maintain proper speed; and (5) 2 Arrows had motor/gearbox failures. Testing of 1 Arrow was stopped after 4,400,100 equivalent cycles as a matter of practicality. Figure 3 shows the accumulation of consumer and supplier failures as a function of equivalent cycles. Twenty-six percent of the consumer and supplier failures occurred during the first 400,000 equivalent cycles. All but 3 wheelchairs (1 Lancer, 2 Chairmen) completed the required 400,000 equivalent cycles on the 2-drum and curb-drop machines without meeting the ANSI/RESNA failure criteria. One Lancer had a motor/gearbox failure and 2 Chairmen had electric connector problems (table 4). All motor brands except Levoy-Somer experienced motor failures (table 5). Both Arrows had motor failures, compared with none of the Quickie and Chairman failures. Table 5 also shows the number of wheelchairs that experienced electronic problems. A Jazzy had a joystick/controller that stopped operating; all other wheelchairs had controller problems that resulted in the wheelchairs failing to maintain the required speed (1m/s). Reliability Figure 2C and table 6 show the mean equivalent cycles for consumer repairs, supplier repairs, consumer failures, and supplier failures. The 5 brands differed significantly in the mean equivalent cycles for consumer repairs (P.05), consumer failures (P.05), and supplier failures (P.001). In terms of the mean equivalent cycles between consumer failures, the Arrow had significantly more time between failures (P.05) than all other brands, and the Jazzy had significantly more time (P.05) than the Chairman, Quickie, and Lancer. In terms of the mean equivalent cycles between supplier failures, the Arrow had more time between failures (P.001) than all other brands. No significant differences in the mean equivalent cycles between supplier repairs were found. Figure 2D and table 6 show the mean equivalent cycles for repairs, failures, consumer repairs and failures, and supplier repairs and failures. The 5 brands differed significantly in the mean equivalent cycles between repairs (P.05), failures (P.001), consumer repairs and failures (P.01), and supplier repairs and failures (P.01). The Arrow had significantly more (P.001) mean equivalent cycles between failures than the other brands, significantly more (P.05) mean equivalent cycles between consumer repairs and failures than the Chairman and Lancer, and significantly more (P.05) mean equivalent cycles between supplier repairs and failures than the Chairman, Jazzy, and Lancer. DISCUSSION The Arrow consistently outperformed the other 4 brands in nearly every area studied. The Arrow had the highest durability and value, in terms of both first failure and until no longer operable. The Arrow also had the greatest reliability, with respect to consumer failures, supplier failures, repairs, failures,

810 LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass Fig 2. (A) Durability. The total number of equivalent cycles completed until first failure and until no longer operable. (B) Value. Value takes into account both the durability and the retail price of the wheelchairs, so a higher value is more desirable because it indicates that the wheelchair lasted longer for its price. (C) Reliability, in terms of mean equivalent cycles between consumer repairs, supplier repairs, consumer failures, and supplier failures. A higher number is more desirable, because it indicates that the wheelchair completed more equivalent cycles for the given number of repairs or failures. (D) Reliability, in terms of mean equivalent cycles between repairs, failures, consumer repairs and failures, and supplier repairs and failures. consumer repairs and failures, and supplier repairs and failures. The Jazzy had the greatest reliability in terms of consumer repairs and supplier repairs. Further analysis also showed that the K0014 wheelchairs lasted longer (until no longer operable) than the K0011 wheelchairs. All wheelchairs were purchased as the standard base models, meaning that no modifications were made so that the features of the wheelchairs were equal. The goal was to test 5 wheelchair brands in their basic configurations in accordance with the current standards 5 and to compare these results. There was 1 case for which data were excluded to ensure that comparisons between the 5 brands were more objective. Between 2 of the Chairmen, there were 3 instances in which the seat tilt electric connector became disconnected, which met the ANSI/RESNA failure criteria. 5 These data were excluded from further analysis because the other wheelchairs did not have a seat recline feature. Other problems with the backrest and seat in the Chairman were not excluded because the backrest and seat are inherent to the wheelchair and the cause of the problem could not directly be attributable to the seat recline feature. In addition, rear-wheel, mid-wheel, and front-wheel drive wheelchairs were tested slightly differently. Proper performance on the 2-drum test requires that the wheelchairs be placed so that the front wheels are always placed on the front drums. The front drum on the 2-drum machine turns 7% faster than the rear drum, so that the slats hit against the casters and drive wheels asynchronously, as specified in the standards. For the rear-wheel drive wheelchairs, the drive wheels were on the rear drum. For the mid-wheel and front-wheel drive wheelchairs, the drive wheels were on the front drum. This leads to the slats hitting the drive wheels and casters at different rates. However, no significant differences were found in the durability until no longer operable between the rear-wheel drive (Arrow, Lancer, Quickie), mid-wheel drive (Jazzy), and frontwheel drive (Chairman) EPWs. The majority of the wheelchairs were considered inoperable because of similar causes: either motor and gearbox failures or the inability for the wheelchairs to maintain proper speed. Interestingly, although both the Quickie and Arrow had Stature motors, the Arrow experienced motor failures but the Quickie did not. This may be due to the Quickie failing for other reasons before motor damage. In addition, all wheelchair brands, regardless of the electronics used, had a problem maintaining proper speed. This can be a concern, particularly for wheelchair users who travel long distances or outdoors in extreme weather. Three wheelchairs experienced failures besides motor and gearbox or electronic problems. First, 1 Chairman had a cracked seat post. The seat attached to the wheelchair base via a single post, a feature the manufacturer probably used so that tilt-in-space and seat elevation could easily be implemented if needed. However, the Chairman was the only wheelchair that had this single post design and the only wheelchair to exhibit frame fractures. This design may be flawed, but likely is a consequence of designing the wheelchair to have the capacity to achieve simple conversion to specialized seating features. On subsequent designs of the Chairman, the manufacturer changed the design to strengthen this component. One Chairman had a fracture located underneath the wheelchair. The fracture caused the casters to turn inward and prevented the wheels from moving properly. One Jazzy had a joystick/controller failure. The Jazzy was the only brand tested that had an integrated joystick/controller; the other brands had separate electronics for the joystick and the controller. However, the other brands with Penny & Giles electronics did not experience this problem. Our investigation into the cause of this failure was inconclusive. Most of the wheelchairs performed very well when tested until 400,000 equivalent cycles, which is assumed to be equivalent to 3 to 5 years of typical use. 11 Only 3 of the 15 wheelchairs met the ANSI/RESNA failure criteria, 5 with only

LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass 811 Table 4: Description of Repairs and Failures Jazzy Quickie Lancer Arrow Chairman 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Consumer repairs Footrest Caster/tire Joystick Armrest Shroud/bumper/panel Seat Backrest Battery wiring harness Fuse box/controller bolts Supplier repairs Cracked seat tower Unable to maintain proper speed* Flat caster/tire Cracked footrest mount Consumer failures 2nd adjustment Disconnected electric connectors Supplier failures Unable to maintain proper speed Motor/gearbox Melted wiring Moved to left Joystick/controller Brakes Cracked seat post Frame fracture Disconnected electric connectors NOTE. The boldface indicates where either a repair or failure occurred within the first 400,000 equivalent cycles. *Lasted 10,000 equivalent cycles. Lasted 10,000 equivalent cycles. 1 wheelchair, a Lancer, having had a failure serious enough to consider it inoperable. In addition, the majority of these problems were consumer repairs, which typically are simple and inexpensive to repair. Two Chairman wheelchairs had problems with disconnected electric connectors and, in fact, throughout the lifespan, there were many instances of disconnected electric connectors. Some of these connectors were usually easy to reconnect, but these problems still temporarily incapacitated the wheelchair. This could potentially put the user at risk if this problem occurred in unsafe situations, such as when crossing the street or in extreme weather. An earlier study 15 tested EPWs for 200,000 cycles on the 2-drum machine and 6600 cycles on the curb-drop machine using the older class I, II, and III failure criteria and methodology. 14 Three of the 10 wheelchairs in that study met the ANSI/RESNA class III failure criteria. Although care should be taken when analyzing data retrospectively, all of the EPWs appear to have met the current ANSI/RESNA failure criteria. 5 This implies that the quality of EPWs is improving. Looking at how long the wheelchairs continued to operate after the first occurrence of a failure can also be helpful. After Table 5: Motor and Electronic Problems That Met the ANSI/RESNA Failure Criteria 5 Fig 3. Accumulation of consumer and supplier failures. Note that the data point at 4,400,110 equivalent cycles (for the Arrow) is not actually a failure, because the wheelchair was still operable. Reprinted from Fitzgerald et al 17 with permission from the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Brand Motor Brand Motor Failures (N) Electronics Brand Electronics Failures (N) Jazzy Reliance 2 Penny & Giles 2 Quickie Stature 0 Penny & Giles 3 Lancer Fracmo 1 Penny & Giles 2 Arrow Stature 2 MKIV 2 Chairman Levoy-Somer 0 Penny & Giles 1

812 LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass Table 6: Reliability, in Terms of Mean Equivalent Cycles Between Consumer Repairs, Supplier Repairs, Consumer Failures, and Supplier Failures Reliability Jazzy Quickie Lancer Arrow Chairman Consumer repairs 343,523 193,667 270,456 115,245 73,204 45,488 316,674 76,379 124,547 45,202 Supplier repairs 1,034,009 670,169 425,472 19,558 152,891 217,236 1,016,468 1,025,309 677,038 152,864 Consumer failures 944,672 229,665 425,472 19,558 276,981 1,219,337 2,752,869 1,452,028 451,935 281,436 Supplier failures 387,305 35,115 425,472 19,558 209,698 190,033 1,458,821 311,384 327,594 198,135 Repairs 258,654 101,521 270,456 115,245 59,418 47,392 275,006 108,976 124,547 45,202 Failures 354,252 92,205 425,472 19,558 209,698 190,033 1,458,821 311,384 234,812 137,753 Consumer repairs/failures 248,987 30,949 270,456 115,245 73,204 45,488 316,674 76,379 107,828 42,484 Supplier repairs/failures 307,965 36,622 425,472 19,558 120,024 89,947 822,860 280,402 327,594 198,135 NOTE. Values are equivalent cycles SD. the first failure, the Arrow lasted 2.0 times longer, the Chairman 1.9 times longer, and the Jazzy 1.8 times longer. The Quickie and Lancer had no further useful life after the first failure. Consumers and insurers must decide on the practicality of servicing wheelchairs after the first failure. Repairing failures, depending on the cost of the repairs, seems to be a viable option, particularly for the Arrow, Chairman, and Jazzy. Information about the cost of repairs is needed to fully understand the usefulness of making these repairs. Reliable repair cost estimates were unobtainable because of very high variability among insurers and across regions. Therefore, costs of repairs were not analyzed. Converting durability, until no longer operable, into times or distances provides helpful estimates. Assuming that 400,000 equivalent cycles is equivalent to 3 to 5 years of typical use and using the number of equivalent cycles completed until the wheelchairs were no longer operable, estimates about the life span would be 21 to 34 years (Arrow), 9 to 16 years (Jazzy), 5 to 9 years (Chairman), 3 to 5 years (Quickie), and 2 to 4 years (Lancer). It is unlikely that EPWs would last much more than 10 years, because the standards do not account for thermal damage and environmental aging. However, the EPWs first completed all of the other ANSI/RESNA standards before commencement of the study. By using the circumference of the 2-drum, the total estimated distances traveled would be: 2202km (1368mi) for the Arrow, 1005km (624mi) for the Jazzy, 542km (336mi) for the Chairman, 340km (212mi) for the Quickie, and 222km (138mi) for the Lancer. Previous research logged distances traveled over a 5-day period for EPW users and determined that the maximum theoretic distance for 1 day was 8.0km (5mi). 16 Based on this information, the typical wheelchair user would travel an estimated 5 times farther (assuming 3y is equivalent to 400,000 equivalent cycles) to 9 times farther (assuming 5y is equivalent to 400,000 equivalent cycles) than the distances estimated using the circumference of the 2-drum. Repair and failure data from wheelchair users in real-world environments are still needed to validate these estimates. Value takes into account both the durability and the initial retail price. A higher value indicates that the wheelchair lasted longer for a given purchase price. Although the Arrow had the second highest average retail price ($7109) (table 1), it also had the highest value, both in terms of first failure and until no longer operable (table 3). This reiterates the fact that cost alone is not the only factor to consider when selecting a wheelchair. Conversely, although the Chairman had the highest average retail purchase price ($13,441) (table 1), its value was the lowest (table 3). Reliability takes into account both the times between each incidence of problems and the total number of occurrences of problems. A high reliability indicates that the wheelchair lasted a long time between the incidences of problems for the given number of occurrences of problems. Converting reliability into estimated years can also be helpful. Assuming 400,000 equivalent cycles is an estimated 3 to 5 years of typical wheelchair use and using the data for the mean equivalent cycles between consumer repairs, supplier repairs, consumer failures, and supplier failures, table 7 shows the reliability, in estimated years. The majority of the wheelchairs would experience repairs and failures in less than 3 to 5 years (bold type in table 7), which is typical of the replacement timeframe for most wheelchairs. The cost of repair probably plays a notable role in determining the time to replace an EPW. In addition, the risk of the EPW causing a serious injury also tends to increase with time, which would be a consideration when deciding whether to replace an EPW. The number of equivalent cycles completed until first failure for power wheelchairs and ultralight, lightweight, and depot manual wheelchairs can be compared. 7-9 An ultralight manual wheelchair is adjustable and is designed for long-term use. 7,10 Table 7: Reliability, in Estimated Years Jazzy Quickie Lancer Arrow Chairman Consumer repair Consumer repair Consumer repair Consumer repair Consumer repair 2.6 4.3y 2.0 3.4y 0.5 0.9y 2.4 4.0y 0.9 1.6y Supplier failure Supplier repair Supplier repair Supplier repair Supplier failure 2.9 4.8y 3.2 5.3y* 1.1 1.9y 7.6 12.7y 2.5 4.1y Consumer failure Consumer failure Supplier failure Supplier failure Consumer failure 7.1 11.8y 3.2 5.3y* 1.6 2.6y 10.9 18.2y 3.4 5.6y Supplier repair Supplier failure Consumer failure Consumer failure Supplier repair 7.8 12.9y 3.2 5.3y* 2.1 3.5y 20.6 34.4y 5.1 8.5y NOTE. The boldface portions indicate an estimated time less than 3 to 5 years. *Identical values.

LIFE-CYCLE ANALYSIS OF ELECTRIC POWERED WHEELCHAIRS, Fass 813 Fig 4. Survival curves during the first 400,000 equivalent cycles using manual wheelchairs from earlier study. 17 A step in the curve indicates a first failure occurrence. A lightweight manual wheelchair, by definition, weighs less than 15.7kg (35lb), is adjusted minimally or is nonadjustable, and is geared toward individual or institutional use. 7,10 A depot wheelchair is a generic manual wheelchair designed for airport terminals, hospitals, or other temporary environments and is minimally adjustable or nonadjustable. 10,17 The power wheelchairs lasted 1.7 times as long as ultralight manual wheelchairs (n 12), 3.4 times longer than lightweight wheelchairs (n 9), and 13.0 times longer than depot wheelchairs (n 6). Therefore, users switching from ultralight wheelchairs to power wheelchairs may on average see lowered durability, and users switching from depot or lightweights to power wheelchairs will on average see increased durability. Of course, other factors should be considered when deciding whether to switch to power mobility, such as the benefit of traveling greater distances and minimizing overuse injuries. These factors would encourage greater community participation. Figure 4 compares the survival curves of the EPWs (n 15) with manual wheelchairs (n 61) from an earlier study. 17 Steps in the survival curves indicate the occurrences of failures. Eighty percent of power, 74% of ultralight manual, 29% of lightweight, and 21% of depot manual wheelchairs had no failures during the first 400,000 equivalent cycles. In other words, users of power and ultralight wheelchairs should expect fewer first failures than lightweight and depot wheelchair users. Previous research 18 found that depot wheelchairs accounted for most of the repairs reported in the National Prosthetic Patient Database, a database that collects information on all wheelchairs that have been prescribed through the VHA. Although further information about the types of repairs are still needed, this finding is in agreement with an earlier finding that depot wheelchairs require the most repairs when tested on the 2-drum and curb-drop machines. Studies comparing the laboratory results of power wheelchairs with actual reported problems in real-life settings could validate the results of this study. In terms of the type of repairs and failures that occurred, both the manual and power wheelchairs experienced footrest and caster problems, loose bolts, and flat tires. Other repairs and failures were more specific to the type of wheelchair, such as loose pushrims for the manual wheelchairs and electric problems for the power wheelchairs. In addition, the power wheelchairs experienced failures that typically did not notably affect operability during the first 400,000 equivalent cycles. For example, reconnecting disconnected electric connectors, adjusting the speed, and making more than 1 adjustment to the same part may not be difficult to address. Conversely, the manual wheelchairs experienced failures that notably affected the operability or safety of the wheelchair, such as broken caster assemblies, cracked seat, frame fractures, and frame distortions, that could not easily be repaired. Manual wheelchairs may experience more catastrophic failures because of lighter weight materials, designed for easier propulsion. Value, until first failure, of power and manual wheelchairs can also be compared. 7-9 The value for the power wheelchairs was 6.5 times lower than that for the ultralight wheelchairs, 2.0 times lower than that for the lightweight wheelchairs, and 1.1 times lower than that for the depot. When looking at the value until the occurrence of first failures, ultralight and lightweight wheelchairs are more economical than EPWs, and depot wheelchairs are roughly equal to EPWs. It is not entirely surprising that ultralight and lightweight wheelchairs are more economical than EPWs, given the higher cost of EPWs. However, it is interesting that depot manual wheelchairs, which are typically the least expensive wheelchairs on the market, are roughly as economical as EPWs, when factoring in the durability until the occurrence of first failures. As previously mentioned, depot wheelchairs also tend to experience first failures that affect the operability and safety of the user compared with EPWs. It requires considerable time to conduct research of this type; therefore, EPW models and features may change before publication. However, these wheelchair models are still used by consumers, and the study provides indicators for areas that need improvement. We selected EPWs that were commonly provided at the time of the study, but those selected may not represent the performance of all EPWs. Because of the cost of purchasing wheelchairs and of conducting a study of this type, the number of units tested was limited. Our results indicated significant differences. However, more differences might have been identified had a larger number of wheelchairs for each model been tested. This study was based on laboratory testing on machines designed to simulate wheelchair use, and studies are required to determine whether the results reflect actual consumer experiences. Care should be taken when comparing the EPWs, because the 5 brands have different features. Other important criteria to consider when choosing a wheelchair are the seating system; front-, rear-, or mid-wheel drive; appearance; and other features such as the capacity for tilt-in-space, recline, and seat elevation. There have been regular reports on the quality and value of manual wheelchairs, but information has not been available at the same rate for EPWs. Future studies need to continue to test EPWs, including scooters. Studies are needed to determine whether the laboratory testing results are comparable with actual consumer use in the home and community. Comparisons are needed for wheelchairs with specialized seating features, such as elevating seats, recline, and tilt functions. Recently, payers have been scrutinizing EPW recommendations more stringently. Hence, data are required for basic K0010 EPWs to learn how they compare with other wheelchairs. CONCLUSIONS This is the first known study to test EPWs over their entire lifespan; it provides important information about durability, value, and reliability. The 5 brands showed different performance characteristics when tested on the 2-drum and curb-drop machines according to the ANSI/RESNA standards. 5 The 5