- Email: [email protected]
Evaluation of the Safety and Durability of Low-Cost Nonprogrammable Electric Powered Wheelchairs
2361
ORIGINAL ARTICLE
Evaluation of the Safety and Durability of Low-Cost Nonprogrammable Electric Powered Wheelchairs Jonathan L. Pearlman, MSc, Rory A. Cooper, PhD, Jaideep Karnawat, Rosemarie Cooper, MPT, ATP, Michael L. Boninger, MD ABSTRACT. Pearlman JL, Cooper RA, Karnawat J, Cooper R, Boninger ML. Evaluation of the safety and durability of lowcost nonprogrammable electric powered wheelchairs. Arch Phys Med Rehabil 2005;86:2361-70. Objective: To evaluate whether a selection of low-cost, nonprogrammable electric-powered wheelchairs (EPWs) meets the American National Standards Institute (ANSI)/Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Wheelchair Standards requirements. Design: Objective comparison tests of various aspects of power wheelchair design and performance of 4 EPW types. Specimens: Three of each of the following EPWs: Pride Mobility Jet 10 (Pride), Invacare Pronto M50 (Invacare), Electric Mobility Rascal 250PC (Electric Mobility), and the Golden Technologies Alanté GP-201-F (Golden). Setting: Rehabilitation engineering research center. Interventions: Not applicable. Main Outcome Measures: Static tipping angle; dynamic tipping score; braking distance; energy consumption; climatic conditioning; power and control systems integrity and safety; and static, impact, and fatigue life (equivalent cycles). Results: Static tipping angle and dynamic tipping score were significantly different across manufacturers for each tipping direction (range, 6.6°⫺35.6°). Braking distances were significantly different across manufacturers (range, 7.4⫺117.3cm). Significant differences among groups were found with analysis of variance (ANOVA). Energy consumption results show that all EPWs can travel over 17km before the battery is expected to be exhausted under idealized conditions (range, 18.2⫺32.0km). Significant differences among groups were found with ANOVA. All EPWs passed the climatic conditioning tests. Several adverse responses were found during the power and control systems testing, including motors smoking during the stalling condition (Electric Mobility), charger safety issues (Electric Mobility, Invacare), and controller failures (Golden). All EPWs passed static and impact test-
From the Human Engineering Research Laboratories, VA Rehabilitation Research and Development Center, VA Pittsburgh Healthcare Systems, Pittsburgh, PA (Pearlman, RA Cooper, Karnawat, Cooper, Boninger); and the Departments of Rehabilitation Science and Technology (Pearlman, RA Cooper, Cooper, Boninger), Physical Medicine and Rehabilitation (RA Cooper, Cooper, Boninger), and Bioengineering (RA Cooper, Cooper, Boninger), University of Pittsburgh, Pittsburgh, PA. Supported by the VA Rehabilitation Research and Development Center (grant no. F2181C), the National Institute on Disability and Rehabilitation Research, Rehabilitation Engineering Research Center on Wheeled Mobility (grant no. H133E990001), a National Science Foundation Graduate Research Fellowship, and a National Science Foundation Integrative Graduate Education and Research Traineeship (grant no. 0333420). 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: [email protected]. 0003-9993/05/8612-9854$30.00/0 doi:10.1016/j.apmr.2005.07.294
ing; 9 of 12 failed fatigue testing (3 Invacare, 3 Golden, 1 Electric Mobility, 2 Pride). Equivalent cycles did not differ statistically across manufacturers (range, 9759⫺824,628 cycles). Conclusions: Large variability in the results, especially with respect to static tipping, power and control system failures, and fatigue life suggest design improvements must be made to make these low-cost, nonprogrammable EPWs safe and reliable for the consumer. Based on our results, these EPWs do not, in general, meet the ANSI/RESNA Wheelchair Standards requirements. Key Words: Reference standards; Rehabilitation; Safety; Wheelchairs. © 2005 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation HE AMERICAN NATIONAL Standards Institute (ANSI)/ T Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Wheelchair Standards (the 1,2
U.S. version of the International Standards Organization [ISO] standards) allow clinicians and their patients to objectively compare and contrast wheelchairs to identify the most appropriate product. This becomes increasingly important when wheelchair reimbursement by insurance providers (eg, Medicare) is limited; clinicians and their patients aim to identify the best wheelchair for their needs within the allowable reimbursement set by the insurance company. Safety and durability are critical factors when choosing an electric-powered wheelchair (EPW). Safety issues, such as stability, maximum speed, and braking distance are important factors that can either contribute to or prevent the common tipand fall-related injuries.3 Durability of the wheelchair should be sufficient that it does not fail within a 3- to 5-year period, which is the typical time span that Medicare (and thus other insurance companies) expects wheelchairs to last before they will fund a replacement. Factors that determine wheelchair durability include the ability to withstand static, impact, and fatigue loading conditions, and controller and charger wiring malfunctions. While the U.S. Food and Drug Administration (FDA) requires each wheelchair put on the market to pass the ANSI/ RESNA standards, testing by independent laboratories is not required. Thus, when these evaluations are performed by the manufacturer, bias or misinterpretation of the ANSI/RESNA standards may lead to improper outcomes. Previous studies have found that both EPWs4,5 and manual wheelchairs6-8 on the market do not necessarily pass the ANSI/RESNA testing standards when tested independently. The least expensive of the manual wheelchairs (depot-style) has been shown to fare worst overall,7 which makes it especially important to test this class of wheelchairs. Furthermore, neither the FDA nor Medicare currently has the resources (eg, independent laboratory tests) to verify that the manufacturers’ testing methods are correct. Arch Phys Med Rehabil Vol 86, December 2005
2362
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
The low-cost power equivalent to the depot-style manual wheelchair is the low-cost, nonprogrammable EPW. These EPWs have standard “captain’s chair” seating systems, and do not offer programmable controllers (the primary difference among these low-cost EPWs and the next higher grade). In the United States, Medicare reimbursed for over 26,000 of the low-cost nonprogrammable EPWs from 2000 to 2003 and, while they reimburse for several-fold more next higher grade EPWs (more than 600,000 over the same period9-11), no ANSI/ RESNA standards test results on the currently available lowcost nonprogrammable EPWs have been published. Studies published on the higher grade EPWs have shown that durability is generally high,4,5 but there is no evidence that this is the case for these low-cost EPWs. Identifying if and what shortcomings exist in these low-cost EPWs is becoming increasingly important in the United States, in particular, given the recent EPW coding changes proposed by Medicare (which are generally followed by other insurance providers) in combination with the competitive bidding strategies that may be instated soon.12 Proposed coding changes include requiring all EPWs to have programmable controllers, which would require manufacturers to upgrade the low-cost EPWs with a higher quality controller.13 No specific requirements have been made on the upgrades to the underlying frame, drive train, or seating systems. Considering that competitive bidding for EPWs is likely in the near future,12 manufacturers may choose to base future model designs on the low-cost EPWs to limit costs and thus be more competitive against other manufacturers. If safety and/or durability problems exist in these low-cost EPWs then it is important to identify and correct them before they could be transferred to the upcoming models. Furthermore, these issues are not isolated to the United States, because health care costs are rising worldwide and associated cost-cutting measures may negatively affect device quality and consumer safety. To evaluate the safety and durability of low-cost nonprogrammable EPWs independently, we performed ANSI/RESNA standards testing on a selection of these low-cost EPWs. We performed this study in part to present independent informative data comparing and contrasting a selection of EPWs. We also sought to compare these low-cost nonprogrammable EPWs with the higher-cost and higher-grade EPWs previously tested and discussed in the literature4,5; we make this comparison specifically based on the durability measures (equivalent cycles) and related value (cycles per dollar) for these and the higher class EPWs. METHODS We performed ANSI/RESNA Wheelchair Standards tests on 3 identical low-cost, nonprogrammable EPWs from a total of 4 manufacturers (N⫽12). The following wheelchairs were tested: Pride Mobility Jet 10a (Pride), Invacare Pronto M50b (Invacare), Electric Mobility Rascal 250PCc (Electric Mobility), and the Golden Technologies Alanté GP-201-Fd (Golden) (fig 1). Of the several low-cost, nonprogrammable wheelchairs on the market, we chose to test models offered by the largest manufacturers to ensure that the results would be relevant to the largest population of potential and current users. The wheelchairs were purchased through a third-party purchaser to ensure that we received a random sample of the wheelchairs from the manufacturers. We followed the methods required in the ANSI/RESNA Wheelchair Standards explicitly, and performed all tests required for EPWs with the exception of sections 16 and 21 (ignition of upholstery and electromagnetic compatibility, reArch Phys Med Rehabil Vol 86, December 2005
Fig 1. EPWs tested. (A) Electric Mobility, (B) Invacare, (C) Golden, and (D) Pride.
spectively), because our laboratories are not equipped to complete these sections. For convenience, we tested all wheelchairs according to methods required for a specific section before starting another section. Within a given section, we randomized the order a particular wheelchair was tested to wash out any sequence effects. In sections 14 (Power and Control) and 8 (Static, Impact, and Fatigue), we broke from this protocol and performed tests simultaneously to several (randomly ordered) wheelchairs in the interest of saving time. The methods are published in the standards, but we briefly describe each section and any statistical methods beyond descriptive statistics that we used to analyze and present the data (the standards do not specify statistical methods, only testing methods). Static Stability Static Stability (§1) is performed by placing the wheelchair with a 100-kg test dummy on a test ramp, and changing the inclination of the test ramp until the angle is found where the EPW will tip (fig 2). This angle is recorded with the wheelchair set in both the most and least stable configurations for the forward (wheels unlocked and locked), rearward (wheels unlocked and locked), sideways (left and right sides down slope), and on the antitippers (either front or back). In total, 14 measurements were recorded. A 1-way multivariate analysis of variance (MANOVA; main effect: EPW model) was performed on the tip-angles for the forward, sideways, and rearward direction
2363
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 2. Static stability testing (rearward direction).
with the wheelchair in the least stable condition, and using the dependent variables that described when tires would initially lose contact and cause loss of control of the EPW. Dynamic Stability Dynamic Stability (§2) is performed by evaluating the response of the EPW to dynamic tasks while traveling on a 0°, 3°, 6°, and 10° test plane. Responses are coded with a score from 0 to 4, which indicates if the EPW tipped completely (0), became stuck on the antitipper (1), performed a transient tip and the antitippers touched the ground (2), performed a transient tip (3), or did not tip (4). These codes were recorded for 31 tasks, including starting and stopping, traveling upward and downward, while turning, and when traveling up and down a step transition of 12, 25, and 50mm. For most cases, a human test pilot maneuvered the wheelchair unless the expected response was dangerous for the rider. In these cases, a 100-kg test dummy was secured to the EPW while a human operator walked or ran beside the EPW. All trials were performed at maximum speed. To compare dynamic stability scores across manufacturers, a Kruskal-Wallis test was performed and, if significant differences were found, pairwise Mann-Whitney U tests were performed. Effectiveness of Brakes Effectiveness of Brakes (§3) is evaluated by measuring the braking distance of the EPW while traveling on a 0°, 3°, 6°, and 10° test plane in both the forward and rearward direction. Three braking modes were evaluated: normal joystick release, joystick reverse, and power off. In total, 72 data points were collected for each EPW. We used a repeated-measures analysis of variance (ANOVA) and Tukey post hoc (where appropriate) to distinguish if EPW had significantly different braking distances for each condition.
Energy Consumption Energy Consumption (§4) is reported as the theoretical range a particular EPW can travel before it depletes the batteries. By measuring the depletion of a fully charged battery (E amperehours), with a known capacity (C ampere-hours) while traveling a known distance (D meters), the theoretical range can be calculated (R kilometers) by the following equation: R⫽
C⫻D E ⫻ 1000
Climatic Testing Climatic Testing (§9) is performed by exposing the EPW to 5 adverse environmental conditions including long- and shortterm heating and cooling, and water to simulate conditions that may occur during normal use, shipping, or storage. After environmental exposure, each EPW is maneuvered through a test track and adverse behaviors are recorded (if it does not maneuver correctly, it fails this section). The adverse behaviors and other reasons the EPW would fail the test include (1) any behaviors deemed dangerous by the tester, (2) the time taken to drive around the test track is greater than 60 seconds, (3) the wheelchair fails to stop, or (4) the wheelchair moves when not commanded to. Static, Impact, and Fatigue Static, Impact, and Fatigue (§8) testing is performed by applying static and impact loading conditions to parts of the EPW (armrests, footrests, wheels, shrouding) and by testing the fatigue life of the whole wheelchair. Fatigue life (also referred to as durability) is tested using double-drum and curb-drop testing machines. Scores in section 8 are based on whether the EPW passes or fails the given test; for the fatigue testing, the EPW passes the test if it endures 200,000 double-drum and Arch Phys Med Rehabil Vol 86, December 2005
2364
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 3. An Invacare being loaded onto the triple-drum fatigue testing machine.
polarity of the battery wires) or events resulting from wear and tear, such as short circuiting (caused by insulation wear). All adverse behaviors that are potentially dangerous are reported.
6666 curb-drop cycles (which is equivalent to 3⫺5y of use). We mounted an additional drum on the double-drum testing machine when testing the Invacare Pronto M50 because it has 6 wheels simultaneously on the ground (fig 3). This modification has been suggested for future versions of the ISO Wheelchair Standards, but has not been formalized.14 All values are reported with descriptive statistics. Additionally, equivalent fatigue life cycles were calculated using the following equation: equivalent cycles ⫽ double-drum cycles ⫹ 30*(curb-drop cycles)4,6,7 and were compared using a 1-way ANOVA (main effect: EPW model). If main effects were significant at ␣ equal to .05, a Tukey post hoc was performed. Value, defined as the number of equivalent cycles per dollar, was also calculated by normalizing the equivalent cycles by the retail price of the EPW.
RESULTS Static stability results (table 1) are presented for the leaststable EPW setup. The locked forward and rearward condition and antitip condition was not applicable for the Invacare (see figs 2, 3) because it is a mid-wheel drive EPW with front and rear casters that are always in contact with the ground. Thus, tipping angle is insensitive to whether the drive wheels are locked or unlocked because it is measured when the drive wheels initially lose contact with the test plane. The Golden and Pride EPW have rear casters, and the Electric Mobility EPW has front casters, and thus the rearward and forward locked conditions were not applicable, respectively. A MANOVA was performed using dependent variables related to when the user would likely lose control of the EPW in any of the 4 tipping directions. The left and right sideways tipping angles were used for all EPWs. The lock condition was used for the forward direction, and the unlock condition for the rearward direction for both the Golden and Pride EPWs. The
Power and Control Systems Power and Control Systems (§14) are evaluated through a series of tests on the joystick, controller, battery charger, battery wiring, and drive motor wiring. These tests evaluate the response of the power and control system to events due to user or maintenance errors (eg, stalling the wheelchair, reversing the
Table 1: Mean Static Tipping Angles for Each Direction and Condition for the Least Stable Setup Direction
Forward
Rearward
Sideways
Condition
Lock
Unlock
Lock
Unlock
Antitip
Left
Right
Golden Pride Electric Mobility Invacare
20.1⫾0.9b 12.5⫾0.6c * *
25.1⫾1.3 17.9⫾0.3 32.64.5a 13.4⫾2.8b,c
* * 6.6⫾1.0d *
27.6⫾0.6c 35.6⫾0.9a 8.2⫾1.7 30.5⫾0.4b
29.1⫾1.3 26.8⫾0.6 16.9⫾2.3 *
22.3⫾0.3a 23.8⫾1.3a 22.7⫾0.4a 19.0⫾1.7b
23.7⫾0.4a 22.7⫾0.2a,b 23.6⫾1.3a 20.4⫾1.7b
NOTE. Values are mean degrees ⫾ standard deviation (SD). Italicized values were used in the MANOVA and superscripts (a– d) represent significant groupings based on a Tukey post hoc test for each direction (forward, rearward, each sideway direction). Group a is the group with the highest tipping angle (most stable) and b, c, and d are lower tipping angle groups. *Not applicable for that EPW in that configuration.
Arch Phys Med Rehabil Vol 86, December 2005
2365
4 2a 4a 4a 2 1c 4a 1c 2 1.3⫾0.6 4 3.0⫾1.7 2 1.7⫾0.6b 4a 4a 3 1.3⫾0.6 4 4 4* 4 3.7⫾0.6 4 4 4a 1b 1b 4 4a 2.3⫾0.6b 2.7⫾0.6b 4 4a 1b 4a 4 2.7⫾1.2 4 4 4 4a 1c 2.3⫾0.6b Golden Pride Electric Mobility Invacare
4 4a 2b 3.3⫾0.6a
a a
NOTE. Values are mean deg/mm ⫾ SD. When the score was equal for all 3 EPWs for each manufacture, SDs are not included. Only sections that were not all identically scored with a “4” are presented. Underlined values indicate that there were significant differences in that section based on the Kruskal-Wallis test (P⬍.05). Superscripts a, b, and c are groupings found by the pairwise Mann-Whitney U tests, where “a” represents the highest and thus most safe condition. Abbreviations: F(br), forward braking stability when traveling forward; F-TRAN, forward stability when traveling down a step transition; Lat-TRAN, lateral stability when 1 side of the EPW travels down a step transition; LAT-TRN, lateral stability when turning on a downhill slope; RDH(br), rearward braking stability when traveling backward down a slope; RUH(br), rearward stability when braking after traveling forward on an uphill slope; RUH(sta), rearward stability when starting uphill on a slope; R-TRAN, rearward stability when traveling down a step transition. *Two EPWs were unable to climb the inclination.
4 3.0⫾1.0 4 4 4 3.3⫾1.2 4 4 4 2b 4a 4a 4 2b 3.7⫾0.6a 4a
a a a a b a a a
Lat-TRAN 25 Lat-TRN
10 50
F-TRAN
25 10 6 F(br) 3 0 50
R-TRAN
10 6
RDH(br)
10 RUH(br)
3 10
RUH(sta) 6 Chair
Section
Table 2: Dynamic Stability Scores
unlock condition was used for the forward direction and the lock condition for the rearward direction of the Electric Mobility. For the Invacare EPW, the unlock condition was used for both the forward and rearward directions (see table 1, italicized values). The data were tested and confirmed to be multivariate normal with appropriate skewness and kurtosis.15 The multivariate tests suggest that there are overall statistical differences in tipping angles as a function of EPW model. Post hoc analysis results (see table 1, superscripts) showed groupings among the EPWs for each tipping direction, where the groupings (a⫺d) represent the lowest to the highest tipping angle groups, respectively. Dynamic Stability scores (table 2) showed the EPWs’ behavior during maneuvering on sloped surfaces. Scores ranged from 0 to 4: 4 indicates that at least 1 wheel remains on the test plane; 3 indicates all uphill wheels lift and then drop back to the test plane without antitipper contact; 2 indicates the same as previous but antitippers make contact with the test plane; 1 indicates all uphill wheels lift and the EPW gets stuck on the antitippers; and 0 indicates the EPW tips completely. Potentially dangerous events (eg, when both wheels lift from the test plane) are indicated by a 3 or below, and are especially concerning when they occur on the lower inclinations (eg, 0°) such as in §9.2. Two data points were not recorded because the EPW could not climb the step transition. The Kruskal-Wallis test showed significant differences in all sections (see table 2, underlined headings), except sections 8.5, sections 9.3, sections 10.4, and sections 10.5. Significantly different groups found by the pairwise Mann-Whitney U tests (table 3, superscripts a⫺c) show similarities and differences among and between manufacturers. Forward and reverse braking distances on 4 test planes (0°, 3°, 6°, 10°) for 3 braking conditions are presented below (see table 3). We performed a Kruskal-Wallis test for overall differences across manufacturer (because the data were not multivariate normal) and found significant differences in all testing conditions except the 10°/forward condition (likely because of missing data due to some EPWs skidding down the test plane). Manufacture-wise groupings were calculated with multiple Mann-Whitney U comparisons and are presented as superscripts in the table. Energy consumption results (mean ⫾ standard deviation [SD]) showed the theoretical range of all the EPWs, which are all relatively high: Golden, 18.2⫾0.83km; Invacare, 17.2⫾0.78km; Electric Mobility, 22.5⫾0.73km; and Pride, 32.0⫾0.96km. ANOVA results indicated significant differences across manufacturers, and Tukey post hoc results are reported to show that Pride had the highest theoretical range, followed by Electric Mobility, followed by Golden and Invacare (not significantly different from each other). Maximum speeds (table 4) differed significantly across manufacturers, directions, and angle. The general trend (fastest to slowest EPW) was generally similar across direction and angle, suggesting selected EPWs (eg, Golden) have overall higher top speeds. All EPWs passed the climatic testing section without adverse responses. The only noticeable effect of the cold-storage condition was on the front (drive) tires of the Pride EPWs, which had slight flat-spots after removal from storage. These flat spots were noticeable when maneuvering the EPW, especially at high speeds. After several weeks of testing, the tires became completely round again. All EPWs passed the static and impact tests of section 8. Fatigue (durability) tests varied both between and within manufacturers (table 5). All but 3 EPWs (2 Electric Mobility, 1
50
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Arch Phys Med Rehabil Vol 86, December 2005
2366
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman Table 3: Braking Distance
Inclination
Horizontal
Direction EPW/Brake
Golden Pride Electric Mobility Invacare
3°
Forward Rel
Reverse
Rev d
Off c
Rel c
Forward
Rev b
Off
Rel
b
c
d
79.3⫾6.7 72.4⫾4.0 56.6⫾4.8 33.4⫾4.0 26.3⫾5.4 26.1⫾4.4 91.9⫾6.5 52.7⫾1.7c 36.3⫾2.3b 40.4⫾6.6b 38.0⫾2.3b 30.1⫾1.4b 23.7⫾1.5c 65.2⫾3.0c 41.2⫾1.8a 30.4⫾2.0a 28.5⫾2.9a 12.4⫾1.5a 45.7⫾1.8b 38.0⫾3.0b 38.4⫾2.6b 13.2⫾1.1a
Reverse
Rev
Off b
Rel b
Rev b
Off
79.2⫾2.2 62.2⫾3.5 36.2⫾3.5 28.8⫾2.6 28.6⫾3.4b 44.7⫾5.5a 46.3⫾8.1a 40.8⫾3.3b 32.4⫾2.8b 26.8⫾0.2b
7.5⫾0.3a 7.6⫾0.3a 34.3⫾16.3a 34.2⫾4.5a 33.2⫾7.1a 12.8⫾1.0a 8.3⫾0.7a 10.3⫾1.2b 45.8⫾1.4b 38.1⫾2.7a 40.9⫾2.1a 13.7⫾2.0a
b
9.7⫾2.1a 8.8⫾1.1a 7.4⫾0.8a 10.4⫾0.8a
NOTE. Values are mean centimeters ⫾ SD. The Kruskal-Wallis test suggested overall differences in all conditions except the 10°/forward. Superscripts show significant groupings predicted from the Mann-Whitney U tests, where “a” is the lowest braking distance, and “d” the highest (most unsafe). Abbreviations: Off, controller power off; Rel, normal release; Rev, joystick reversal. *Data were not available because the EPW skidded down the slope and never came to a complete stop.
Pride) did not achieve the required 400,000 equivalent cycles required by the standards. Failure modes included drive-train or seat-related failures (Golden). Seats either bent (n⫽1) (fig 4) or bolts attaching the seat post to the frame broke (n⫽2) for the Golden. No significant differences between or within factor (manufacturer) were found with an ANOVA. Value reported as cycles per dollar was calculated by normalizing the equivalent cycles by the retail price of the EPW. One variation in testing occurred with Electric Mobility #3, which initially failed prematurely on the curb-drop test because of a backrest adjustment problem (at nearly 400,000 equivalent cycles). Because we believed this failure was due to a mis-adjusted backrest, we replaced the backrest with the Electric Mobility #2 backrest and continued testing. The response of the power and control systems varied widely for the EPWs from each manufacturer. Because of the number of tests included in section 14, we only report adverse effects: (1) 2 of 3 Electric Mobility EPWs failed the stalled condition testing (§6.14), resulting in smoke coming from the motors when the EPW was stalled against a wall with the controller pushed forward; (2) 2 of 3 Electric Mobility EPW battery chargers were damaged when performing the reverse polarity at the battery test (§6.10); (3) all Golden EPWs failed the controller command signal processing test (§6.12) because the joysticks had electric failures; all joysticks were replaced so testing could continue; (4) all Golden EPWs failed the reversepolarity battery charger connection test (§9.2.5), suggesting that current continues to flow from the battery charger if its polarity is switched; and (5) all Invacare EPWs failed the charging battery safety test (§6.9), which shows that the EPW can be driven while the battery is being charged (ie, it is plugged into the wall outlet). DISCUSSION Overall, we found significant differences among manufacturers of nearly all variables tested. Static stability varied
significantly across manufacturer, but also was highly sensitive to the direction the EPW was facing. Thus, an EPW highly stable in the forward direction may be unstable in the rearward direction (eg, Electric Mobility) or vice-versa (eg, Pride) (see table 1). Stability in all directions is important as shown by Corfman et al16 who demonstrated that adverse events can occur with both descending and ascending obstacles (ramps and curb-cuts, respectively). Overall, Golden EPWs were most consistent across directions, but were not the most stable EPW for any of the 4 directions. During maneuvering and testing, Pride proved to be the least stable in the forward direction, and care should be taken when setting this EPW up for the user (ie, unstable seating setups can exacerbate the problem). Note that our analysis used the lower of the stability measures (where more than 1 was available) by considering the case when the wheels first left the ground, instead of when the antitip devices failed. This allowed us to compare the angles at which the user would likely lose control of the EPW, and these values do not imply the EPW would tip (eg, the tipping angle for the Electric Mobility in the rearward direction doubles when considering the antitip devices). Also, we report the data for the least-stable seat and component setup to be conservative. The static stability of these low-cost, nonprogrammable EPWs was comparable to the higher class of EPWs as reported by Rentchler et al,5 with the exception of the Electric Mobility, which had a much lower rearward stability tipping angle than any of those reported in the other study. There were also significant differences among manufactures in dynamic stability scores; these differences were more pronounced for the higher inclinations, which is consistent with another comparison study of higher class EPWs,5 except for the forward stability when braking (§9.2), which is one of the more important measures (see table 2). As discussed above, the Pride was noticeably unstable in the forward direction, even on a horizontal test plane. Invacare forward instability was also poor, but this was noticeable only on the 10° slope. Results
Table 4: Average Maximum Speed in Forward and Reverse on 3 Test-Plane Inclinations Direction
Forward
Reverse
Angle
0
3
6
0
3
6
Golden Pride Electric Mobility Invacare
2.29⫾0.12c 1.79⫾0.10b 1.39⫾0.03a 1.43⫾0.04a
2.10⫾0.09c 1.75⫾0.07b 1.34⫾0.05a 1.37⫾0.07a
1.91⫾0.08d 1.59⫾0.03c 1.13⫾0.02a 1.27⫾0.05b
1.26⫾0.07c 1.04⫾0.10b 0.68⫾0.03a 0.62⫾0.02a
2.34⫾0.06c 1.94⫾0.01b 1.42⫾0.02a 1.47⫾0.04a
2.64⫾0.09a 2.21⫾0.01a 1.59⫾0.05a 1.61⫾0.07a
NOTE. Values are mean meter per second ⫾ SD. Superscripts indicated groupings of speeds for each inclination and direction (a– d from low to high) found with a Tukey post hoc test.
Arch Phys Med Rehabil Vol 86, December 2005
2367
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman Table 3 (Cont’d): Braking Distance 6°
10°
Forward Rel
Rev
88.2⫾20.5 73.4⫾2.7b 50.7⫾5.3a 50.2⫾7.2a
Reverse
b
73.7⫾18.2 58.8⫾8.3b 40.4⫾7.0a 42.9⫾0.2a
Off b
Rel b
Forward
Rev b
Off
Rel
58.7⫾13.9 40.2⫾7.9 59.8⫾12.0b 51.9⫾1.3c
b
c
39.0⫾5.8 39.9⫾5.8 47.0⫾5.2b 39.2⫾1.8c
39.8⫾7.1a 45.6⫾0.2b
10.6⫾0.6a 9.8⫾1.0a
9.8⫾0.3a 11.6⫾0.9b
14.8⫾2.0a 15.2⫾1.4a
from rearward stability (§§8.2⫺.4) suggest that the Electric Mobility and Invacare are unstable at higher test-plane inclinations. Note that no EPWs scored a zero, indicating that none completely tipped over. Both dynamic and static stability measures are important when choosing an EPW, because tips and falls account for most injuries of wheelchair users.3 Braking distances were significantly different across manufactures (see table 3), but groupings were largely consistent across directions (forward, reverse) and test-plane inclination. Overall, Golden and Pride were the fastest of the EPWs, and Electric Mobility and Invacare were the slowest. Without programmability of the controllers, braking distances are not in the control of the clinician setting up the EPW. Thus, the reported braking distances are important to be aware of for clinicians who are concerned about the user’s ability to control the EPW, especially when stopping. The braking distances of the EPWs we tested are much lower, in general, than those of the higher class EPWs reported in the literature.5 But because the higher class EPW braking distances are reported by testing the EPW with the controller programmed for maximum velocity, this comparison is not entirely valid (because they can be programmed to restrict maximum velocity). Energy consumption results showed a significant difference across manufacturers, but all theoretical ranges were relatively high. In a study investigating the driving distances of EPW users, researchers found that subjects traveled a mean ⫾ SD of 8.35⫾7.07km over a 5-day period.17 This average increased by over 2-fold (17.16⫾8.7km) when users were participating at a sporting event, but still falls below the theoretical range of the EPWs tested in this study. Thus, our results suggest that, on average, all tested EPWs would reliably run for more than 5 days of use without recharging (which is typically done daily17).
Table 5: Equivalent Cycles and Failure Mode EPW
Equivalent Cycles
Value (cycle/$)
Failure Mode
Golden #1 Golden #2 Golden #3 Pride #1 Pride #2 Pride #3 Electric Mobility #1 Electric Mobility #2 Electric Mobility #3 Invacare #1 Invacare #2 Invacare #3
236510 99863 234950 425968 9759 136396 44821 417928 824628 101770 46428 25694
118.55 50.06 117.77 157.53 3.61 50.44 17.23 160.68 317.04 61.57 28.09 15.54
Seat Seat Seat Drive Drive Drive Drive Drive Drive Drive Drive Drive
train train train train train train train train train
NOTE. Bold values indicate which EPWs passed the ANSI/RESNA requirement of 400,000 equivalent cycles.
Rev
Reverse Off
Rel
Rev
Off
91.4⫾16.1 90.6⫾10.3 76.8⫾11.5 * * * 109.3⫾26.9 87.7⫾17.4 117.3⫾9.8 99.8⫾20.5c 91.6⫾13.3c 95.6⫾32.0b * 49.1⫾5.1
* 41.3⫾3.5
* 13.3⫾0.7a 56.0⫾24.6 14.8⫾0.5b
7.5⫾1.7a 10.2⫾0.2b
7.5⫾1.6a 26.4⫾23.1a
However, this finding should be qualified because the test evaluates energy capacity under idealized conditions (horizontal test plane, near constant ampere draw), it does not take into consideration that batteries lose their charge capacity over time, and it does not include the energy consumption due to starts and stops. Maximum speeds varied across manufacturer, direction, and test-plane angle, but there were consistent trends for each condition (ie, the fastest and slowest EPWs were similar for all conditions) (see table 4). In this study, we present the maximum speeds because it can be an important safety issue, since the nonprogrammability of the controller implies that the clinician setting up the EPW does not have control over the maximum speed of the EPW. Thus, a user who may not be able to safely maneuver the EPW at high speeds could raise the manual speed adjustment (located on the joystick for all EPWs tested) and injure themselves. While it is expected that braking distances will be longer for the faster chairs (which was true, in general; see tables 3, 4), we believe it is more informative with the low-cost nonprogrammable EPWs to present the speed and braking distances separately (even though braking distance comparisons typically control for maximum speeds).5 All EPWs successfully passed the climatic testing, which was an improvement over a previous study5 of higher grade EPWs. We noticed only that the Pride had a flat spot on the tires, presumably because some of the airless insert material lost its elasticity during the freezing portion of the tests. Although the tires did recover their shape after several weeks, manufacturers should be careful to use an appropriate material that will not change properties during storage or shipping. Section 8 (Static, Impact, and Fatigue) results were good overall for the static and impact tests, and poor for the fatigue tests (see table 5). The fact that only 3 (25%) of the 12 EPWs tested successfully completed 400,000 equivalent cycles show the distinct difference between the low-cost EPWs we tested and the higher-cost EPWs which had a success rate of 13 of 15 (86.7%)5 (fig 5). Most failures were in the drive train (motor or gearbox failures), but the Golden consistently failed because of seat failures; all types of failures pose a concern because they may lead to injury of the users.3 Although the rates of failure were higher for our EPWs, the low cost of the EPW resulted in relatively high values (cycles/$) for these EPWs. Value ranges spanned from 3.6 to 317 cycles/$ for the EPWs tested here and from 27 to 196 cycles/$ for the higher classified EPWs.4 These results suggest high variability in value, and also imply that overall the low-cost nonprogrammable EPWs provide more value to the user. The consequence of this low-durability but high value could be 2-fold: either insurance companies would not provide a replacement EPW within the 3 to 5 years (the life expectancy for an EPW); or the insurance company would purchase replacements, requiring the typical paperwork and Arch Phys Med Rehabil Vol 86, December 2005
2368
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 4. Close-up of bent seat failure on the Golden during curb-drop testing.
processing time, which would inflate the price (and thus lower the value) of these EPWs for the user and the insurance company. Furthermore, the wide variability of the equivalent cycles before failure within manufacturers suggests that estimates on the life of any of these EPWs is not reliable and may harm the user (because it may strand the user), which would negate the cost-benefits of using these lower quality EPWs. Clearly, our results show that the quality of the EPWs we tested needs to be improved so that users can have more reliable mobility. An important concern is that manufacturers may use the
Fig 5. Survival curve comparing rates of K10 with those of the higher classified EPWs. Survival of the low-cost EPWs we tested at the 400,000 equivalent cycle mark (vertical gray line) is 25 versus the 0.8 survival of the other EPWs.
Arch Phys Med Rehabil Vol 86, December 2005
basic design of the low-cost nonprogrammable EPWs we tested and upgrade only necessary portions to fit within the proposed Medicare coding system,13 and to compete in the competitive bidding for durable medical equipment that Medicare intends to begin.12 Power and control system testing (§14) results showed some serious safety concerns for all but the Pride EPWs. Of particular concern was the failure of 2 of the Electric Mobility wheelchairs on the stalled condition testing, which simulates a user driving against an obstacle and not releasing the joystick. In practice, this
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
behavior could happen by accident (eg, a user is for some reason unable to release the joystick when stalled) or intentionally if a user attempts to use their EPW to negotiate an obstacle. In this situation, a resetting circuit-breaker (automatic or manual) is supposed to trip to stop current flow to the drive system. With 2 of the Electric Mobility wheelchairs this breaker failed to trip and caused the motors to heat to the point that smoke came out of them. If the EPW were stalled for longer then the 2 minutes the test required, it could have caught fire. Another potentially hazardous behavior we found was with the Invacare during charging. Because the EPW can be maneuvered while being charged, if a user does not unplug the EPW, they may pull the cord out of the wall or the EPW, causing them or others harm. Other adverse behaviors during this testing were related to wiring short circuits. These can occur at several places on the EPW due to wear and tear, or user or technician error. The results of these tests caused failures of the battery chargers (Electric Mobility) and controllers (Golden). Our results provide additional evidence that EPWs (and wheelchairs in general) on the market still do not meet the recommended standards when tested independently. As our results show, all the low-cost, nonprogrammable EPWs we tested have low durability overall and thus are not likely to last the expected 3 to 5 years of use. Results from the power and control system, stability (static and dynamic), braking distance, and maximum speed varied across manufactures. These data will help clinicians to identify the best EPW for their client. Although we followed the ANSI/RESNA standards testing methods explicitly, this study by no means describes the behavior of all low-cost nonprogrammable EPWs. We made an effort to minimize bias by randomizing testing order, and chose a selection of low-cost, nonprogrammable EPWs that represented the largest manufacturers of these wheelchairs. Because of the expense and the time required for testing, we were unable to test more than 3 wheelchairs of the same model and manufacturer; the FDA typically requires testing results for 1 to 3 identical devices for certification, which is what we based our sample size on, given the high cost of these devices. The use of a test dummy, while it has been optimized to reflect human weight distribution and kinematics, and reduces variability in the test results, is only an approximation of an average user. Results from other manufacturers may be different from those we present. An overall result from this study is that the outcome of the different tests were not consistent among manufacturers and, in some cases (as with durability), within manufacturers. Thus, we believe that we have shown the wide range of behaviors that are represented by this class of EPW. One informative evaluation that we did not perform was to test the affect of programmability versus nonprogrammability. Our experience in the clinic has been that programmability is nearly always used because it allows the EPWs to be fitted to users’ abilities and their environment; the users’ abilities, adaptive strategies, and environments vary widely, suggesting that no one program would be appropriate for all EPW users. However, this experience is clinical, and does not objectively test the importance of programmability for user maneuverability or safety. The ANSI/RESNA standards do not include such a test, even though it may help convey the importance of the programmability objectively instead of purely clinically, and help drive appropriate and effective EPW design. Furthermore, the value of device comparison studies is maximized when the devices tested are new on the market, and durability and safety issues are not known. Future studies should look at these new devices, such as the new pushrim-activated power-assist wheel-
2369
chairs, and the high-strength ultra-lightweight manual wheelchairs that are widely prescribed. CONCLUSIONS We found large variations in the test results from the EPWs that we tested, raising concern for users’ safety and the longterm durability of these devices. Clinicians and current and future users of these devices should be aware of these shortcomings to avoid injury. In the current flux of the durable medical equipment industry, from Medicare coding changes to competitive bidding contracts, there is concern that quality of the durable medical equipment will decline in order to lower costs. Our results show that simply using the low-cost, nonprogrammable EPWs as a base for higher classified EPWs that fit Medicare’s coding systems would be a mistake, and engineering changes must be made to the current models before they will reliably and safely provide mobility. Acknowledgments: We thank Andrew Rentschler, John Duncan, Jeremy Puhlman, Donald Spaeth, Michael Dvorznak, Mark Schmeler, Annmarie Kelleher, Erik Wolf, and Mark McCartney for their contributions to this study. References 1. American National Standards Institute, Rehabilitation Engineering and Assistive Technology Society of North America. Wheelchair standards: additional requirements for wheelchairs (including scooters) with electrical systems. Vol 2. New York: ANSI/ RESNA; 1998. 2. American National Standards Institute, Rehabilitation Engineering and Assistive Technology Society of North America. Wheelchair standards: requirements and test methods for wheelchairs (including scooters). Vol 1. New York: ANSI/RESNA; 1998. 3. Calder CJ, Kirby RL. Fatal wheelchair-related accidents in the United States. Am J Phys Med Rehabil 1990;69:184-90. 4. Fass MV, Cooper RA, Fitzgerald SG, et al. Durability, value, and reliability of selected electric powered wheelchairs. Arch Phys Med Rehabil 2004;85:805-14. 5. Rentschler AJ, Cooper RA, Fitzgerald SG, et al. Evaluation of selected electric-powered wheelchairs using the ANSI/RESNA standards. Arch Phys Med Rehabil 2004;85:611-9. 6. Cooper RA, Robertson RN, Lawrence B, et al. Life-Cycle analysis of depot versus rehabilitation manual wheelchairs. J Rehabil Res Dev 1996;33:45-55. 7. Fitzgerald SG, Cooper RA, Boninger ML, Rentschler AJ. Comparison of fatigue life for 3 types of manual wheelchairs. Arch Phys Med Rehabil 2001;82:1484-8. 8. Kwarciak AM, Cooper RA, Ammer WA, Boninger ML, Cooper R. Fatigue testing of selected suspension manual wheelchairs using ANSI/RESNA standards. Arch Phys Med Rehabil 2005;86: 123-9. 9. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2003 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges03.asp. Accessed July 20, 2005. 10. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2001 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges01.asp. Accessed July 20, 2005. 11. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2002 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges02.asp. Accessed July 20, 2005. Arch Phys Med Rehabil Vol 86, December 2005
2370
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
12. Centers for Medicare and Medicaid. Medicare announces new initiatives on power wheelchair coverage and payment policy. Sept 17, 2004. Available at: http://www.cms.hhs.gov/media/press/ release.asp?Counter⫽1023. Accessed July 20, 2005. 13. Centers for Medicare and Medicaid. Healthcare Common Procedure Coding System (HCPCS). Sept 17, 2004. Available at: http:// www.cms.hhs.gov/medicare/hcpcs/default.asp?. Accessed July 20, 2005. 14. International Standards Organization. Comments for revisions of ISO 7176-8. Sept 23, 2003. 15. Timm N. Applied multivariate analysis. New York: Springer; 2002. 16. Corfman TA, Cooper RA, Fitzgerald SG, Cooper R. Tips and falls during electric-powered wheelchair driving: effects of seatbelt
Arch Phys Med Rehabil Vol 86, December 2005
use, legrests, and driving speed. Arch Phys Med Rehabil 2003;84:1797-802. 17. Cooper RA, Thorman T, Cooper R, et al. Driving characteristics of electric-powered wheelchair users: how far, fast, and often do people drive? Arch Phys Med Rehabil 2002;83:250-5. Suppliers a. Pride Mobility Products Corp, 182 Susquehanna Ave, Exeter, PA 18643. b. Invacare, One Invacare Way, Elyria, OH 44036. c. Electric Mobility Corp, One Mobility Plz, PO Box 156, Sewell, NJ 08080. d. Golden Technologies, 401 Bridge St, Old Forge, PA 18518.
ORIGINAL ARTICLE
Evaluation of the Safety and Durability of Low-Cost Nonprogrammable Electric Powered Wheelchairs Jonathan L. Pearlman, MSc, Rory A. Cooper, PhD, Jaideep Karnawat, Rosemarie Cooper, MPT, ATP, Michael L. Boninger, MD ABSTRACT. Pearlman JL, Cooper RA, Karnawat J, Cooper R, Boninger ML. Evaluation of the safety and durability of lowcost nonprogrammable electric powered wheelchairs. Arch Phys Med Rehabil 2005;86:2361-70. Objective: To evaluate whether a selection of low-cost, nonprogrammable electric-powered wheelchairs (EPWs) meets the American National Standards Institute (ANSI)/Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Wheelchair Standards requirements. Design: Objective comparison tests of various aspects of power wheelchair design and performance of 4 EPW types. Specimens: Three of each of the following EPWs: Pride Mobility Jet 10 (Pride), Invacare Pronto M50 (Invacare), Electric Mobility Rascal 250PC (Electric Mobility), and the Golden Technologies Alanté GP-201-F (Golden). Setting: Rehabilitation engineering research center. Interventions: Not applicable. Main Outcome Measures: Static tipping angle; dynamic tipping score; braking distance; energy consumption; climatic conditioning; power and control systems integrity and safety; and static, impact, and fatigue life (equivalent cycles). Results: Static tipping angle and dynamic tipping score were significantly different across manufacturers for each tipping direction (range, 6.6°⫺35.6°). Braking distances were significantly different across manufacturers (range, 7.4⫺117.3cm). Significant differences among groups were found with analysis of variance (ANOVA). Energy consumption results show that all EPWs can travel over 17km before the battery is expected to be exhausted under idealized conditions (range, 18.2⫺32.0km). Significant differences among groups were found with ANOVA. All EPWs passed the climatic conditioning tests. Several adverse responses were found during the power and control systems testing, including motors smoking during the stalling condition (Electric Mobility), charger safety issues (Electric Mobility, Invacare), and controller failures (Golden). All EPWs passed static and impact test-
From the Human Engineering Research Laboratories, VA Rehabilitation Research and Development Center, VA Pittsburgh Healthcare Systems, Pittsburgh, PA (Pearlman, RA Cooper, Karnawat, Cooper, Boninger); and the Departments of Rehabilitation Science and Technology (Pearlman, RA Cooper, Cooper, Boninger), Physical Medicine and Rehabilitation (RA Cooper, Cooper, Boninger), and Bioengineering (RA Cooper, Cooper, Boninger), University of Pittsburgh, Pittsburgh, PA. Supported by the VA Rehabilitation Research and Development Center (grant no. F2181C), the National Institute on Disability and Rehabilitation Research, Rehabilitation Engineering Research Center on Wheeled Mobility (grant no. H133E990001), a National Science Foundation Graduate Research Fellowship, and a National Science Foundation Integrative Graduate Education and Research Traineeship (grant no. 0333420). 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: [email protected]. 0003-9993/05/8612-9854$30.00/0 doi:10.1016/j.apmr.2005.07.294
ing; 9 of 12 failed fatigue testing (3 Invacare, 3 Golden, 1 Electric Mobility, 2 Pride). Equivalent cycles did not differ statistically across manufacturers (range, 9759⫺824,628 cycles). Conclusions: Large variability in the results, especially with respect to static tipping, power and control system failures, and fatigue life suggest design improvements must be made to make these low-cost, nonprogrammable EPWs safe and reliable for the consumer. Based on our results, these EPWs do not, in general, meet the ANSI/RESNA Wheelchair Standards requirements. Key Words: Reference standards; Rehabilitation; Safety; Wheelchairs. © 2005 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation HE AMERICAN NATIONAL Standards Institute (ANSI)/ T Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Wheelchair Standards (the 1,2
U.S. version of the International Standards Organization [ISO] standards) allow clinicians and their patients to objectively compare and contrast wheelchairs to identify the most appropriate product. This becomes increasingly important when wheelchair reimbursement by insurance providers (eg, Medicare) is limited; clinicians and their patients aim to identify the best wheelchair for their needs within the allowable reimbursement set by the insurance company. Safety and durability are critical factors when choosing an electric-powered wheelchair (EPW). Safety issues, such as stability, maximum speed, and braking distance are important factors that can either contribute to or prevent the common tipand fall-related injuries.3 Durability of the wheelchair should be sufficient that it does not fail within a 3- to 5-year period, which is the typical time span that Medicare (and thus other insurance companies) expects wheelchairs to last before they will fund a replacement. Factors that determine wheelchair durability include the ability to withstand static, impact, and fatigue loading conditions, and controller and charger wiring malfunctions. While the U.S. Food and Drug Administration (FDA) requires each wheelchair put on the market to pass the ANSI/ RESNA standards, testing by independent laboratories is not required. Thus, when these evaluations are performed by the manufacturer, bias or misinterpretation of the ANSI/RESNA standards may lead to improper outcomes. Previous studies have found that both EPWs4,5 and manual wheelchairs6-8 on the market do not necessarily pass the ANSI/RESNA testing standards when tested independently. The least expensive of the manual wheelchairs (depot-style) has been shown to fare worst overall,7 which makes it especially important to test this class of wheelchairs. Furthermore, neither the FDA nor Medicare currently has the resources (eg, independent laboratory tests) to verify that the manufacturers’ testing methods are correct. Arch Phys Med Rehabil Vol 86, December 2005
2362
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
The low-cost power equivalent to the depot-style manual wheelchair is the low-cost, nonprogrammable EPW. These EPWs have standard “captain’s chair” seating systems, and do not offer programmable controllers (the primary difference among these low-cost EPWs and the next higher grade). In the United States, Medicare reimbursed for over 26,000 of the low-cost nonprogrammable EPWs from 2000 to 2003 and, while they reimburse for several-fold more next higher grade EPWs (more than 600,000 over the same period9-11), no ANSI/ RESNA standards test results on the currently available lowcost nonprogrammable EPWs have been published. Studies published on the higher grade EPWs have shown that durability is generally high,4,5 but there is no evidence that this is the case for these low-cost EPWs. Identifying if and what shortcomings exist in these low-cost EPWs is becoming increasingly important in the United States, in particular, given the recent EPW coding changes proposed by Medicare (which are generally followed by other insurance providers) in combination with the competitive bidding strategies that may be instated soon.12 Proposed coding changes include requiring all EPWs to have programmable controllers, which would require manufacturers to upgrade the low-cost EPWs with a higher quality controller.13 No specific requirements have been made on the upgrades to the underlying frame, drive train, or seating systems. Considering that competitive bidding for EPWs is likely in the near future,12 manufacturers may choose to base future model designs on the low-cost EPWs to limit costs and thus be more competitive against other manufacturers. If safety and/or durability problems exist in these low-cost EPWs then it is important to identify and correct them before they could be transferred to the upcoming models. Furthermore, these issues are not isolated to the United States, because health care costs are rising worldwide and associated cost-cutting measures may negatively affect device quality and consumer safety. To evaluate the safety and durability of low-cost nonprogrammable EPWs independently, we performed ANSI/RESNA standards testing on a selection of these low-cost EPWs. We performed this study in part to present independent informative data comparing and contrasting a selection of EPWs. We also sought to compare these low-cost nonprogrammable EPWs with the higher-cost and higher-grade EPWs previously tested and discussed in the literature4,5; we make this comparison specifically based on the durability measures (equivalent cycles) and related value (cycles per dollar) for these and the higher class EPWs. METHODS We performed ANSI/RESNA Wheelchair Standards tests on 3 identical low-cost, nonprogrammable EPWs from a total of 4 manufacturers (N⫽12). The following wheelchairs were tested: Pride Mobility Jet 10a (Pride), Invacare Pronto M50b (Invacare), Electric Mobility Rascal 250PCc (Electric Mobility), and the Golden Technologies Alanté GP-201-Fd (Golden) (fig 1). Of the several low-cost, nonprogrammable wheelchairs on the market, we chose to test models offered by the largest manufacturers to ensure that the results would be relevant to the largest population of potential and current users. The wheelchairs were purchased through a third-party purchaser to ensure that we received a random sample of the wheelchairs from the manufacturers. We followed the methods required in the ANSI/RESNA Wheelchair Standards explicitly, and performed all tests required for EPWs with the exception of sections 16 and 21 (ignition of upholstery and electromagnetic compatibility, reArch Phys Med Rehabil Vol 86, December 2005
Fig 1. EPWs tested. (A) Electric Mobility, (B) Invacare, (C) Golden, and (D) Pride.
spectively), because our laboratories are not equipped to complete these sections. For convenience, we tested all wheelchairs according to methods required for a specific section before starting another section. Within a given section, we randomized the order a particular wheelchair was tested to wash out any sequence effects. In sections 14 (Power and Control) and 8 (Static, Impact, and Fatigue), we broke from this protocol and performed tests simultaneously to several (randomly ordered) wheelchairs in the interest of saving time. The methods are published in the standards, but we briefly describe each section and any statistical methods beyond descriptive statistics that we used to analyze and present the data (the standards do not specify statistical methods, only testing methods). Static Stability Static Stability (§1) is performed by placing the wheelchair with a 100-kg test dummy on a test ramp, and changing the inclination of the test ramp until the angle is found where the EPW will tip (fig 2). This angle is recorded with the wheelchair set in both the most and least stable configurations for the forward (wheels unlocked and locked), rearward (wheels unlocked and locked), sideways (left and right sides down slope), and on the antitippers (either front or back). In total, 14 measurements were recorded. A 1-way multivariate analysis of variance (MANOVA; main effect: EPW model) was performed on the tip-angles for the forward, sideways, and rearward direction
2363
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 2. Static stability testing (rearward direction).
with the wheelchair in the least stable condition, and using the dependent variables that described when tires would initially lose contact and cause loss of control of the EPW. Dynamic Stability Dynamic Stability (§2) is performed by evaluating the response of the EPW to dynamic tasks while traveling on a 0°, 3°, 6°, and 10° test plane. Responses are coded with a score from 0 to 4, which indicates if the EPW tipped completely (0), became stuck on the antitipper (1), performed a transient tip and the antitippers touched the ground (2), performed a transient tip (3), or did not tip (4). These codes were recorded for 31 tasks, including starting and stopping, traveling upward and downward, while turning, and when traveling up and down a step transition of 12, 25, and 50mm. For most cases, a human test pilot maneuvered the wheelchair unless the expected response was dangerous for the rider. In these cases, a 100-kg test dummy was secured to the EPW while a human operator walked or ran beside the EPW. All trials were performed at maximum speed. To compare dynamic stability scores across manufacturers, a Kruskal-Wallis test was performed and, if significant differences were found, pairwise Mann-Whitney U tests were performed. Effectiveness of Brakes Effectiveness of Brakes (§3) is evaluated by measuring the braking distance of the EPW while traveling on a 0°, 3°, 6°, and 10° test plane in both the forward and rearward direction. Three braking modes were evaluated: normal joystick release, joystick reverse, and power off. In total, 72 data points were collected for each EPW. We used a repeated-measures analysis of variance (ANOVA) and Tukey post hoc (where appropriate) to distinguish if EPW had significantly different braking distances for each condition.
Energy Consumption Energy Consumption (§4) is reported as the theoretical range a particular EPW can travel before it depletes the batteries. By measuring the depletion of a fully charged battery (E amperehours), with a known capacity (C ampere-hours) while traveling a known distance (D meters), the theoretical range can be calculated (R kilometers) by the following equation: R⫽
C⫻D E ⫻ 1000
Climatic Testing Climatic Testing (§9) is performed by exposing the EPW to 5 adverse environmental conditions including long- and shortterm heating and cooling, and water to simulate conditions that may occur during normal use, shipping, or storage. After environmental exposure, each EPW is maneuvered through a test track and adverse behaviors are recorded (if it does not maneuver correctly, it fails this section). The adverse behaviors and other reasons the EPW would fail the test include (1) any behaviors deemed dangerous by the tester, (2) the time taken to drive around the test track is greater than 60 seconds, (3) the wheelchair fails to stop, or (4) the wheelchair moves when not commanded to. Static, Impact, and Fatigue Static, Impact, and Fatigue (§8) testing is performed by applying static and impact loading conditions to parts of the EPW (armrests, footrests, wheels, shrouding) and by testing the fatigue life of the whole wheelchair. Fatigue life (also referred to as durability) is tested using double-drum and curb-drop testing machines. Scores in section 8 are based on whether the EPW passes or fails the given test; for the fatigue testing, the EPW passes the test if it endures 200,000 double-drum and Arch Phys Med Rehabil Vol 86, December 2005
2364
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 3. An Invacare being loaded onto the triple-drum fatigue testing machine.
polarity of the battery wires) or events resulting from wear and tear, such as short circuiting (caused by insulation wear). All adverse behaviors that are potentially dangerous are reported.
6666 curb-drop cycles (which is equivalent to 3⫺5y of use). We mounted an additional drum on the double-drum testing machine when testing the Invacare Pronto M50 because it has 6 wheels simultaneously on the ground (fig 3). This modification has been suggested for future versions of the ISO Wheelchair Standards, but has not been formalized.14 All values are reported with descriptive statistics. Additionally, equivalent fatigue life cycles were calculated using the following equation: equivalent cycles ⫽ double-drum cycles ⫹ 30*(curb-drop cycles)4,6,7 and were compared using a 1-way ANOVA (main effect: EPW model). If main effects were significant at ␣ equal to .05, a Tukey post hoc was performed. Value, defined as the number of equivalent cycles per dollar, was also calculated by normalizing the equivalent cycles by the retail price of the EPW.
RESULTS Static stability results (table 1) are presented for the leaststable EPW setup. The locked forward and rearward condition and antitip condition was not applicable for the Invacare (see figs 2, 3) because it is a mid-wheel drive EPW with front and rear casters that are always in contact with the ground. Thus, tipping angle is insensitive to whether the drive wheels are locked or unlocked because it is measured when the drive wheels initially lose contact with the test plane. The Golden and Pride EPW have rear casters, and the Electric Mobility EPW has front casters, and thus the rearward and forward locked conditions were not applicable, respectively. A MANOVA was performed using dependent variables related to when the user would likely lose control of the EPW in any of the 4 tipping directions. The left and right sideways tipping angles were used for all EPWs. The lock condition was used for the forward direction, and the unlock condition for the rearward direction for both the Golden and Pride EPWs. The
Power and Control Systems Power and Control Systems (§14) are evaluated through a series of tests on the joystick, controller, battery charger, battery wiring, and drive motor wiring. These tests evaluate the response of the power and control system to events due to user or maintenance errors (eg, stalling the wheelchair, reversing the
Table 1: Mean Static Tipping Angles for Each Direction and Condition for the Least Stable Setup Direction
Forward
Rearward
Sideways
Condition
Lock
Unlock
Lock
Unlock
Antitip
Left
Right
Golden Pride Electric Mobility Invacare
20.1⫾0.9b 12.5⫾0.6c * *
25.1⫾1.3 17.9⫾0.3 32.64.5a 13.4⫾2.8b,c
* * 6.6⫾1.0d *
27.6⫾0.6c 35.6⫾0.9a 8.2⫾1.7 30.5⫾0.4b
29.1⫾1.3 26.8⫾0.6 16.9⫾2.3 *
22.3⫾0.3a 23.8⫾1.3a 22.7⫾0.4a 19.0⫾1.7b
23.7⫾0.4a 22.7⫾0.2a,b 23.6⫾1.3a 20.4⫾1.7b
NOTE. Values are mean degrees ⫾ standard deviation (SD). Italicized values were used in the MANOVA and superscripts (a– d) represent significant groupings based on a Tukey post hoc test for each direction (forward, rearward, each sideway direction). Group a is the group with the highest tipping angle (most stable) and b, c, and d are lower tipping angle groups. *Not applicable for that EPW in that configuration.
Arch Phys Med Rehabil Vol 86, December 2005
2365
4 2a 4a 4a 2 1c 4a 1c 2 1.3⫾0.6 4 3.0⫾1.7 2 1.7⫾0.6b 4a 4a 3 1.3⫾0.6 4 4 4* 4 3.7⫾0.6 4 4 4a 1b 1b 4 4a 2.3⫾0.6b 2.7⫾0.6b 4 4a 1b 4a 4 2.7⫾1.2 4 4 4 4a 1c 2.3⫾0.6b Golden Pride Electric Mobility Invacare
4 4a 2b 3.3⫾0.6a
a a
NOTE. Values are mean deg/mm ⫾ SD. When the score was equal for all 3 EPWs for each manufacture, SDs are not included. Only sections that were not all identically scored with a “4” are presented. Underlined values indicate that there were significant differences in that section based on the Kruskal-Wallis test (P⬍.05). Superscripts a, b, and c are groupings found by the pairwise Mann-Whitney U tests, where “a” represents the highest and thus most safe condition. Abbreviations: F(br), forward braking stability when traveling forward; F-TRAN, forward stability when traveling down a step transition; Lat-TRAN, lateral stability when 1 side of the EPW travels down a step transition; LAT-TRN, lateral stability when turning on a downhill slope; RDH(br), rearward braking stability when traveling backward down a slope; RUH(br), rearward stability when braking after traveling forward on an uphill slope; RUH(sta), rearward stability when starting uphill on a slope; R-TRAN, rearward stability when traveling down a step transition. *Two EPWs were unable to climb the inclination.
4 3.0⫾1.0 4 4 4 3.3⫾1.2 4 4 4 2b 4a 4a 4 2b 3.7⫾0.6a 4a
a a a a b a a a
Lat-TRAN 25 Lat-TRN
10 50
F-TRAN
25 10 6 F(br) 3 0 50
R-TRAN
10 6
RDH(br)
10 RUH(br)
3 10
RUH(sta) 6 Chair
Section
Table 2: Dynamic Stability Scores
unlock condition was used for the forward direction and the lock condition for the rearward direction of the Electric Mobility. For the Invacare EPW, the unlock condition was used for both the forward and rearward directions (see table 1, italicized values). The data were tested and confirmed to be multivariate normal with appropriate skewness and kurtosis.15 The multivariate tests suggest that there are overall statistical differences in tipping angles as a function of EPW model. Post hoc analysis results (see table 1, superscripts) showed groupings among the EPWs for each tipping direction, where the groupings (a⫺d) represent the lowest to the highest tipping angle groups, respectively. Dynamic Stability scores (table 2) showed the EPWs’ behavior during maneuvering on sloped surfaces. Scores ranged from 0 to 4: 4 indicates that at least 1 wheel remains on the test plane; 3 indicates all uphill wheels lift and then drop back to the test plane without antitipper contact; 2 indicates the same as previous but antitippers make contact with the test plane; 1 indicates all uphill wheels lift and the EPW gets stuck on the antitippers; and 0 indicates the EPW tips completely. Potentially dangerous events (eg, when both wheels lift from the test plane) are indicated by a 3 or below, and are especially concerning when they occur on the lower inclinations (eg, 0°) such as in §9.2. Two data points were not recorded because the EPW could not climb the step transition. The Kruskal-Wallis test showed significant differences in all sections (see table 2, underlined headings), except sections 8.5, sections 9.3, sections 10.4, and sections 10.5. Significantly different groups found by the pairwise Mann-Whitney U tests (table 3, superscripts a⫺c) show similarities and differences among and between manufacturers. Forward and reverse braking distances on 4 test planes (0°, 3°, 6°, 10°) for 3 braking conditions are presented below (see table 3). We performed a Kruskal-Wallis test for overall differences across manufacturer (because the data were not multivariate normal) and found significant differences in all testing conditions except the 10°/forward condition (likely because of missing data due to some EPWs skidding down the test plane). Manufacture-wise groupings were calculated with multiple Mann-Whitney U comparisons and are presented as superscripts in the table. Energy consumption results (mean ⫾ standard deviation [SD]) showed the theoretical range of all the EPWs, which are all relatively high: Golden, 18.2⫾0.83km; Invacare, 17.2⫾0.78km; Electric Mobility, 22.5⫾0.73km; and Pride, 32.0⫾0.96km. ANOVA results indicated significant differences across manufacturers, and Tukey post hoc results are reported to show that Pride had the highest theoretical range, followed by Electric Mobility, followed by Golden and Invacare (not significantly different from each other). Maximum speeds (table 4) differed significantly across manufacturers, directions, and angle. The general trend (fastest to slowest EPW) was generally similar across direction and angle, suggesting selected EPWs (eg, Golden) have overall higher top speeds. All EPWs passed the climatic testing section without adverse responses. The only noticeable effect of the cold-storage condition was on the front (drive) tires of the Pride EPWs, which had slight flat-spots after removal from storage. These flat spots were noticeable when maneuvering the EPW, especially at high speeds. After several weeks of testing, the tires became completely round again. All EPWs passed the static and impact tests of section 8. Fatigue (durability) tests varied both between and within manufacturers (table 5). All but 3 EPWs (2 Electric Mobility, 1
50
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Arch Phys Med Rehabil Vol 86, December 2005
2366
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman Table 3: Braking Distance
Inclination
Horizontal
Direction EPW/Brake
Golden Pride Electric Mobility Invacare
3°
Forward Rel
Reverse
Rev d
Off c
Rel c
Forward
Rev b
Off
Rel
b
c
d
79.3⫾6.7 72.4⫾4.0 56.6⫾4.8 33.4⫾4.0 26.3⫾5.4 26.1⫾4.4 91.9⫾6.5 52.7⫾1.7c 36.3⫾2.3b 40.4⫾6.6b 38.0⫾2.3b 30.1⫾1.4b 23.7⫾1.5c 65.2⫾3.0c 41.2⫾1.8a 30.4⫾2.0a 28.5⫾2.9a 12.4⫾1.5a 45.7⫾1.8b 38.0⫾3.0b 38.4⫾2.6b 13.2⫾1.1a
Reverse
Rev
Off b
Rel b
Rev b
Off
79.2⫾2.2 62.2⫾3.5 36.2⫾3.5 28.8⫾2.6 28.6⫾3.4b 44.7⫾5.5a 46.3⫾8.1a 40.8⫾3.3b 32.4⫾2.8b 26.8⫾0.2b
7.5⫾0.3a 7.6⫾0.3a 34.3⫾16.3a 34.2⫾4.5a 33.2⫾7.1a 12.8⫾1.0a 8.3⫾0.7a 10.3⫾1.2b 45.8⫾1.4b 38.1⫾2.7a 40.9⫾2.1a 13.7⫾2.0a
b
9.7⫾2.1a 8.8⫾1.1a 7.4⫾0.8a 10.4⫾0.8a
NOTE. Values are mean centimeters ⫾ SD. The Kruskal-Wallis test suggested overall differences in all conditions except the 10°/forward. Superscripts show significant groupings predicted from the Mann-Whitney U tests, where “a” is the lowest braking distance, and “d” the highest (most unsafe). Abbreviations: Off, controller power off; Rel, normal release; Rev, joystick reversal. *Data were not available because the EPW skidded down the slope and never came to a complete stop.
Pride) did not achieve the required 400,000 equivalent cycles required by the standards. Failure modes included drive-train or seat-related failures (Golden). Seats either bent (n⫽1) (fig 4) or bolts attaching the seat post to the frame broke (n⫽2) for the Golden. No significant differences between or within factor (manufacturer) were found with an ANOVA. Value reported as cycles per dollar was calculated by normalizing the equivalent cycles by the retail price of the EPW. One variation in testing occurred with Electric Mobility #3, which initially failed prematurely on the curb-drop test because of a backrest adjustment problem (at nearly 400,000 equivalent cycles). Because we believed this failure was due to a mis-adjusted backrest, we replaced the backrest with the Electric Mobility #2 backrest and continued testing. The response of the power and control systems varied widely for the EPWs from each manufacturer. Because of the number of tests included in section 14, we only report adverse effects: (1) 2 of 3 Electric Mobility EPWs failed the stalled condition testing (§6.14), resulting in smoke coming from the motors when the EPW was stalled against a wall with the controller pushed forward; (2) 2 of 3 Electric Mobility EPW battery chargers were damaged when performing the reverse polarity at the battery test (§6.10); (3) all Golden EPWs failed the controller command signal processing test (§6.12) because the joysticks had electric failures; all joysticks were replaced so testing could continue; (4) all Golden EPWs failed the reversepolarity battery charger connection test (§9.2.5), suggesting that current continues to flow from the battery charger if its polarity is switched; and (5) all Invacare EPWs failed the charging battery safety test (§6.9), which shows that the EPW can be driven while the battery is being charged (ie, it is plugged into the wall outlet). DISCUSSION Overall, we found significant differences among manufacturers of nearly all variables tested. Static stability varied
significantly across manufacturer, but also was highly sensitive to the direction the EPW was facing. Thus, an EPW highly stable in the forward direction may be unstable in the rearward direction (eg, Electric Mobility) or vice-versa (eg, Pride) (see table 1). Stability in all directions is important as shown by Corfman et al16 who demonstrated that adverse events can occur with both descending and ascending obstacles (ramps and curb-cuts, respectively). Overall, Golden EPWs were most consistent across directions, but were not the most stable EPW for any of the 4 directions. During maneuvering and testing, Pride proved to be the least stable in the forward direction, and care should be taken when setting this EPW up for the user (ie, unstable seating setups can exacerbate the problem). Note that our analysis used the lower of the stability measures (where more than 1 was available) by considering the case when the wheels first left the ground, instead of when the antitip devices failed. This allowed us to compare the angles at which the user would likely lose control of the EPW, and these values do not imply the EPW would tip (eg, the tipping angle for the Electric Mobility in the rearward direction doubles when considering the antitip devices). Also, we report the data for the least-stable seat and component setup to be conservative. The static stability of these low-cost, nonprogrammable EPWs was comparable to the higher class of EPWs as reported by Rentchler et al,5 with the exception of the Electric Mobility, which had a much lower rearward stability tipping angle than any of those reported in the other study. There were also significant differences among manufactures in dynamic stability scores; these differences were more pronounced for the higher inclinations, which is consistent with another comparison study of higher class EPWs,5 except for the forward stability when braking (§9.2), which is one of the more important measures (see table 2). As discussed above, the Pride was noticeably unstable in the forward direction, even on a horizontal test plane. Invacare forward instability was also poor, but this was noticeable only on the 10° slope. Results
Table 4: Average Maximum Speed in Forward and Reverse on 3 Test-Plane Inclinations Direction
Forward
Reverse
Angle
0
3
6
0
3
6
Golden Pride Electric Mobility Invacare
2.29⫾0.12c 1.79⫾0.10b 1.39⫾0.03a 1.43⫾0.04a
2.10⫾0.09c 1.75⫾0.07b 1.34⫾0.05a 1.37⫾0.07a
1.91⫾0.08d 1.59⫾0.03c 1.13⫾0.02a 1.27⫾0.05b
1.26⫾0.07c 1.04⫾0.10b 0.68⫾0.03a 0.62⫾0.02a
2.34⫾0.06c 1.94⫾0.01b 1.42⫾0.02a 1.47⫾0.04a
2.64⫾0.09a 2.21⫾0.01a 1.59⫾0.05a 1.61⫾0.07a
NOTE. Values are mean meter per second ⫾ SD. Superscripts indicated groupings of speeds for each inclination and direction (a– d from low to high) found with a Tukey post hoc test.
Arch Phys Med Rehabil Vol 86, December 2005
2367
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman Table 3 (Cont’d): Braking Distance 6°
10°
Forward Rel
Rev
88.2⫾20.5 73.4⫾2.7b 50.7⫾5.3a 50.2⫾7.2a
Reverse
b
73.7⫾18.2 58.8⫾8.3b 40.4⫾7.0a 42.9⫾0.2a
Off b
Rel b
Forward
Rev b
Off
Rel
58.7⫾13.9 40.2⫾7.9 59.8⫾12.0b 51.9⫾1.3c
b
c
39.0⫾5.8 39.9⫾5.8 47.0⫾5.2b 39.2⫾1.8c
39.8⫾7.1a 45.6⫾0.2b
10.6⫾0.6a 9.8⫾1.0a
9.8⫾0.3a 11.6⫾0.9b
14.8⫾2.0a 15.2⫾1.4a
from rearward stability (§§8.2⫺.4) suggest that the Electric Mobility and Invacare are unstable at higher test-plane inclinations. Note that no EPWs scored a zero, indicating that none completely tipped over. Both dynamic and static stability measures are important when choosing an EPW, because tips and falls account for most injuries of wheelchair users.3 Braking distances were significantly different across manufactures (see table 3), but groupings were largely consistent across directions (forward, reverse) and test-plane inclination. Overall, Golden and Pride were the fastest of the EPWs, and Electric Mobility and Invacare were the slowest. Without programmability of the controllers, braking distances are not in the control of the clinician setting up the EPW. Thus, the reported braking distances are important to be aware of for clinicians who are concerned about the user’s ability to control the EPW, especially when stopping. The braking distances of the EPWs we tested are much lower, in general, than those of the higher class EPWs reported in the literature.5 But because the higher class EPW braking distances are reported by testing the EPW with the controller programmed for maximum velocity, this comparison is not entirely valid (because they can be programmed to restrict maximum velocity). Energy consumption results showed a significant difference across manufacturers, but all theoretical ranges were relatively high. In a study investigating the driving distances of EPW users, researchers found that subjects traveled a mean ⫾ SD of 8.35⫾7.07km over a 5-day period.17 This average increased by over 2-fold (17.16⫾8.7km) when users were participating at a sporting event, but still falls below the theoretical range of the EPWs tested in this study. Thus, our results suggest that, on average, all tested EPWs would reliably run for more than 5 days of use without recharging (which is typically done daily17).
Table 5: Equivalent Cycles and Failure Mode EPW
Equivalent Cycles
Value (cycle/$)
Failure Mode
Golden #1 Golden #2 Golden #3 Pride #1 Pride #2 Pride #3 Electric Mobility #1 Electric Mobility #2 Electric Mobility #3 Invacare #1 Invacare #2 Invacare #3
236510 99863 234950 425968 9759 136396 44821 417928 824628 101770 46428 25694
118.55 50.06 117.77 157.53 3.61 50.44 17.23 160.68 317.04 61.57 28.09 15.54
Seat Seat Seat Drive Drive Drive Drive Drive Drive Drive Drive Drive
train train train train train train train train train
NOTE. Bold values indicate which EPWs passed the ANSI/RESNA requirement of 400,000 equivalent cycles.
Rev
Reverse Off
Rel
Rev
Off
91.4⫾16.1 90.6⫾10.3 76.8⫾11.5 * * * 109.3⫾26.9 87.7⫾17.4 117.3⫾9.8 99.8⫾20.5c 91.6⫾13.3c 95.6⫾32.0b * 49.1⫾5.1
* 41.3⫾3.5
* 13.3⫾0.7a 56.0⫾24.6 14.8⫾0.5b
7.5⫾1.7a 10.2⫾0.2b
7.5⫾1.6a 26.4⫾23.1a
However, this finding should be qualified because the test evaluates energy capacity under idealized conditions (horizontal test plane, near constant ampere draw), it does not take into consideration that batteries lose their charge capacity over time, and it does not include the energy consumption due to starts and stops. Maximum speeds varied across manufacturer, direction, and test-plane angle, but there were consistent trends for each condition (ie, the fastest and slowest EPWs were similar for all conditions) (see table 4). In this study, we present the maximum speeds because it can be an important safety issue, since the nonprogrammability of the controller implies that the clinician setting up the EPW does not have control over the maximum speed of the EPW. Thus, a user who may not be able to safely maneuver the EPW at high speeds could raise the manual speed adjustment (located on the joystick for all EPWs tested) and injure themselves. While it is expected that braking distances will be longer for the faster chairs (which was true, in general; see tables 3, 4), we believe it is more informative with the low-cost nonprogrammable EPWs to present the speed and braking distances separately (even though braking distance comparisons typically control for maximum speeds).5 All EPWs successfully passed the climatic testing, which was an improvement over a previous study5 of higher grade EPWs. We noticed only that the Pride had a flat spot on the tires, presumably because some of the airless insert material lost its elasticity during the freezing portion of the tests. Although the tires did recover their shape after several weeks, manufacturers should be careful to use an appropriate material that will not change properties during storage or shipping. Section 8 (Static, Impact, and Fatigue) results were good overall for the static and impact tests, and poor for the fatigue tests (see table 5). The fact that only 3 (25%) of the 12 EPWs tested successfully completed 400,000 equivalent cycles show the distinct difference between the low-cost EPWs we tested and the higher-cost EPWs which had a success rate of 13 of 15 (86.7%)5 (fig 5). Most failures were in the drive train (motor or gearbox failures), but the Golden consistently failed because of seat failures; all types of failures pose a concern because they may lead to injury of the users.3 Although the rates of failure were higher for our EPWs, the low cost of the EPW resulted in relatively high values (cycles/$) for these EPWs. Value ranges spanned from 3.6 to 317 cycles/$ for the EPWs tested here and from 27 to 196 cycles/$ for the higher classified EPWs.4 These results suggest high variability in value, and also imply that overall the low-cost nonprogrammable EPWs provide more value to the user. The consequence of this low-durability but high value could be 2-fold: either insurance companies would not provide a replacement EPW within the 3 to 5 years (the life expectancy for an EPW); or the insurance company would purchase replacements, requiring the typical paperwork and Arch Phys Med Rehabil Vol 86, December 2005
2368
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
Fig 4. Close-up of bent seat failure on the Golden during curb-drop testing.
processing time, which would inflate the price (and thus lower the value) of these EPWs for the user and the insurance company. Furthermore, the wide variability of the equivalent cycles before failure within manufacturers suggests that estimates on the life of any of these EPWs is not reliable and may harm the user (because it may strand the user), which would negate the cost-benefits of using these lower quality EPWs. Clearly, our results show that the quality of the EPWs we tested needs to be improved so that users can have more reliable mobility. An important concern is that manufacturers may use the
Fig 5. Survival curve comparing rates of K10 with those of the higher classified EPWs. Survival of the low-cost EPWs we tested at the 400,000 equivalent cycle mark (vertical gray line) is 25 versus the 0.8 survival of the other EPWs.
Arch Phys Med Rehabil Vol 86, December 2005
basic design of the low-cost nonprogrammable EPWs we tested and upgrade only necessary portions to fit within the proposed Medicare coding system,13 and to compete in the competitive bidding for durable medical equipment that Medicare intends to begin.12 Power and control system testing (§14) results showed some serious safety concerns for all but the Pride EPWs. Of particular concern was the failure of 2 of the Electric Mobility wheelchairs on the stalled condition testing, which simulates a user driving against an obstacle and not releasing the joystick. In practice, this
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
behavior could happen by accident (eg, a user is for some reason unable to release the joystick when stalled) or intentionally if a user attempts to use their EPW to negotiate an obstacle. In this situation, a resetting circuit-breaker (automatic or manual) is supposed to trip to stop current flow to the drive system. With 2 of the Electric Mobility wheelchairs this breaker failed to trip and caused the motors to heat to the point that smoke came out of them. If the EPW were stalled for longer then the 2 minutes the test required, it could have caught fire. Another potentially hazardous behavior we found was with the Invacare during charging. Because the EPW can be maneuvered while being charged, if a user does not unplug the EPW, they may pull the cord out of the wall or the EPW, causing them or others harm. Other adverse behaviors during this testing were related to wiring short circuits. These can occur at several places on the EPW due to wear and tear, or user or technician error. The results of these tests caused failures of the battery chargers (Electric Mobility) and controllers (Golden). Our results provide additional evidence that EPWs (and wheelchairs in general) on the market still do not meet the recommended standards when tested independently. As our results show, all the low-cost, nonprogrammable EPWs we tested have low durability overall and thus are not likely to last the expected 3 to 5 years of use. Results from the power and control system, stability (static and dynamic), braking distance, and maximum speed varied across manufactures. These data will help clinicians to identify the best EPW for their client. Although we followed the ANSI/RESNA standards testing methods explicitly, this study by no means describes the behavior of all low-cost nonprogrammable EPWs. We made an effort to minimize bias by randomizing testing order, and chose a selection of low-cost, nonprogrammable EPWs that represented the largest manufacturers of these wheelchairs. Because of the expense and the time required for testing, we were unable to test more than 3 wheelchairs of the same model and manufacturer; the FDA typically requires testing results for 1 to 3 identical devices for certification, which is what we based our sample size on, given the high cost of these devices. The use of a test dummy, while it has been optimized to reflect human weight distribution and kinematics, and reduces variability in the test results, is only an approximation of an average user. Results from other manufacturers may be different from those we present. An overall result from this study is that the outcome of the different tests were not consistent among manufacturers and, in some cases (as with durability), within manufacturers. Thus, we believe that we have shown the wide range of behaviors that are represented by this class of EPW. One informative evaluation that we did not perform was to test the affect of programmability versus nonprogrammability. Our experience in the clinic has been that programmability is nearly always used because it allows the EPWs to be fitted to users’ abilities and their environment; the users’ abilities, adaptive strategies, and environments vary widely, suggesting that no one program would be appropriate for all EPW users. However, this experience is clinical, and does not objectively test the importance of programmability for user maneuverability or safety. The ANSI/RESNA standards do not include such a test, even though it may help convey the importance of the programmability objectively instead of purely clinically, and help drive appropriate and effective EPW design. Furthermore, the value of device comparison studies is maximized when the devices tested are new on the market, and durability and safety issues are not known. Future studies should look at these new devices, such as the new pushrim-activated power-assist wheel-
2369
chairs, and the high-strength ultra-lightweight manual wheelchairs that are widely prescribed. CONCLUSIONS We found large variations in the test results from the EPWs that we tested, raising concern for users’ safety and the longterm durability of these devices. Clinicians and current and future users of these devices should be aware of these shortcomings to avoid injury. In the current flux of the durable medical equipment industry, from Medicare coding changes to competitive bidding contracts, there is concern that quality of the durable medical equipment will decline in order to lower costs. Our results show that simply using the low-cost, nonprogrammable EPWs as a base for higher classified EPWs that fit Medicare’s coding systems would be a mistake, and engineering changes must be made to the current models before they will reliably and safely provide mobility. Acknowledgments: We thank Andrew Rentschler, John Duncan, Jeremy Puhlman, Donald Spaeth, Michael Dvorznak, Mark Schmeler, Annmarie Kelleher, Erik Wolf, and Mark McCartney for their contributions to this study. References 1. American National Standards Institute, Rehabilitation Engineering and Assistive Technology Society of North America. Wheelchair standards: additional requirements for wheelchairs (including scooters) with electrical systems. Vol 2. New York: ANSI/ RESNA; 1998. 2. American National Standards Institute, Rehabilitation Engineering and Assistive Technology Society of North America. Wheelchair standards: requirements and test methods for wheelchairs (including scooters). Vol 1. New York: ANSI/RESNA; 1998. 3. Calder CJ, Kirby RL. Fatal wheelchair-related accidents in the United States. Am J Phys Med Rehabil 1990;69:184-90. 4. Fass MV, Cooper RA, Fitzgerald SG, et al. Durability, value, and reliability of selected electric powered wheelchairs. Arch Phys Med Rehabil 2004;85:805-14. 5. Rentschler AJ, Cooper RA, Fitzgerald SG, et al. Evaluation of selected electric-powered wheelchairs using the ANSI/RESNA standards. Arch Phys Med Rehabil 2004;85:611-9. 6. Cooper RA, Robertson RN, Lawrence B, et al. Life-Cycle analysis of depot versus rehabilitation manual wheelchairs. J Rehabil Res Dev 1996;33:45-55. 7. Fitzgerald SG, Cooper RA, Boninger ML, Rentschler AJ. Comparison of fatigue life for 3 types of manual wheelchairs. Arch Phys Med Rehabil 2001;82:1484-8. 8. Kwarciak AM, Cooper RA, Ammer WA, Boninger ML, Cooper R. Fatigue testing of selected suspension manual wheelchairs using ANSI/RESNA standards. Arch Phys Med Rehabil 2005;86: 123-9. 9. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2003 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges03.asp. Accessed July 20, 2005. 10. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2001 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges01.asp. Accessed July 20, 2005. 11. Centers for Medicare and Medicaid. Part B Physician/Supplier Nat’l Data, CY 2002 Top 200 Level II Healthcare Common Procedure Coding System. Sept 17, 2004. Available at: http://www.cms.hhs. gov/statistics/feeforservice/top200l2hcpcsbycharges02.asp. Accessed July 20, 2005. Arch Phys Med Rehabil Vol 86, December 2005
2370
EVALUATION OF LOW-COST POWER WHEELCHAIRS, Pearlman
12. Centers for Medicare and Medicaid. Medicare announces new initiatives on power wheelchair coverage and payment policy. Sept 17, 2004. Available at: http://www.cms.hhs.gov/media/press/ release.asp?Counter⫽1023. Accessed July 20, 2005. 13. Centers for Medicare and Medicaid. Healthcare Common Procedure Coding System (HCPCS). Sept 17, 2004. Available at: http:// www.cms.hhs.gov/medicare/hcpcs/default.asp?. Accessed July 20, 2005. 14. International Standards Organization. Comments for revisions of ISO 7176-8. Sept 23, 2003. 15. Timm N. Applied multivariate analysis. New York: Springer; 2002. 16. Corfman TA, Cooper RA, Fitzgerald SG, Cooper R. Tips and falls during electric-powered wheelchair driving: effects of seatbelt
Arch Phys Med Rehabil Vol 86, December 2005
use, legrests, and driving speed. Arch Phys Med Rehabil 2003;84:1797-802. 17. Cooper RA, Thorman T, Cooper R, et al. Driving characteristics of electric-powered wheelchair users: how far, fast, and often do people drive? Arch Phys Med Rehabil 2002;83:250-5. Suppliers a. Pride Mobility Products Corp, 182 Susquehanna Ave, Exeter, PA 18643. b. Invacare, One Invacare Way, Elyria, OH 44036. c. Electric Mobility Corp, One Mobility Plz, PO Box 156, Sewell, NJ 08080. d. Golden Technologies, 401 Bridge St, Old Forge, PA 18518.