Effect of 2-Speed Geared Manual Wheelchair Propulsion on Shoulder Pain and Function

Wheelchair User’s Shoulder Pain Index 

The WUSPI is a 15-item self-report instrument that detects the effect of shoulder pain on the performance of daily activities.25 The items are: (1) transfer from bed to wheelchair, (2) transfer from wheelchair to car, (3) transfer from wheelchair to tub or shower, (4) loading the wheelchair into a car, (5) pushing the wheelchair for 10 minutes or more, (6) pushing up ramps or outdoor inclines, (7) lifting objects to remove them from an overhead shelf, (8) putting on pants, (9) putting on a T-shirt or pullover, (10) putting on a button down shirt, (11) washing one’s back, (12) usual daily activities at work or school, (13) driving, (14) performing household chores; and (15) sleeping. The instrument is based on visual analog scales, with a minimum score of 0 (no pain) and a maximum score of 10cm (worst pain ever experienced) for each of the 15 items. The total score is calculated by summing all 15-item scores resulting in a possible range of 0 to 150cm. The performance-adjusted WUSPI score is calculated by dividing the total score by the number of completed items. The WUSPI has shown high reliability and internal consistency.25

To determine stability of shoulder pain before the intervention, subjects completed WUSPI surveys at enrollment and then for 4 consecutive weeks during the baseline phase (5 surveys total). During the use of the MAGICWheels, subjects completed the WUSPI survey weekly in the initial 4 weeks, then monthly for the remainder of the phase, for a total of 8 surveys completed during the intervention. Weekly WUSPI surveys were completed during the 4-week retention phase.

Wheelchair Users Functional Assessment 

The WUFA26 consists of 13 basic and community activities: (1) tightspace, (2) uneven terrain, (3) door management, (4) street crossing, (5) ramp, (6) curb, (7) bed, (8) transfer, (9) toilet transfer, (10) floor transfer, (11) bathing, (12) upper and lower dressing, and (13) reaching function, picking up objects, and sweeping. Each item is timed with a maximum score of 7. To receive a score of 7, a person must complete each activity before the cutoff time without assistance. In addition, the amount of assistance required and the need for assistive devices are considered when each item is scored. All scores are then summed to obtain a total score. The highest and lowest scores possible are 91 and 13, respectively. The WUFA has good content validity, interrater reliability and stability, and good internal consistency.26 Participants completed the WUFA at initial enrollment and after the 5-month phase in which the MAGICWheels were used.

Activities Log 

The activity log survey required the participant to report the number of times a week they encountered hills, ramps and/or inclines, carpet, gravel, curbs, and grass surfaces. These are the physical environments in which the MAGICWheels’ 2-geared wheelchair system maximally assists the user. To determine stability of activity levels and to verify that inclusion criteria were met prior to the intervention, 4 weekly activity surveys were completed during the baseline phase. To verify the continuation of the level of activity at enrollment, as well as to determine whether they increased their activity when using the MAGICWheels, subjects completed the activity surveys weekly for the initial 4 weeks and then monthly for the remainder of the phase for a total of 8 surveys completed during the intervention. Weekly activity surveys were also completed during the 4-week retention phase.

Incline Test 

The incline test consisted of wheelchair navigation up an incline of 20m, 20° incline (1:2.5 incline) using their personal wheelchairs and the MAGICWheels in 2:1 gearing. The Americans with Disabilities Act of 199027 requires that ramps not exceed 1:12 (1in rise for every 12in of run). The incline, located in a parking garage, was selected to be especially challenging to the manual wheelchair users so that the gearing features of MAGICWheels would be emphasized. Variables measured during the incline tests were: (1) total time (in seconds) from starting point to crossing the end of the 20-m incline and (2) subjective rating of difficulty with the Borg Rate of Perceived Exertion (RPE) scale.28 The incline testing was performed before and after the use of MAGICWheels.

Data Analysis 

Repeated-measures analysis of variance (ANOVA) determined the stability of shoulder pain and activity level during baseline phase. Because there was no change in shoulder pain or activity level without an intervention, we calculated the mean of the 5 baseline scores and defined it as the baseline WUSPI score and baseline activity level. The percentage change in total WUSPI score was calculated (baseline score-current score / baseline score × 100) and compared with the baseline score, using repeated-measures ANOVA at weeks 1 through 4 of MAGICWheels’ use (n=17), and weeks 8 (n=16), 12 (n=16), 16 (n=14), and 20 (n=12). Differences in the pre- and post- MAGICWheels phase WUFA, activity levels, and clinical evaluation variables were determined using repeated-measures ANOVA. We analyzed the incline variables with a 2-way (wheel type and test) repeated-measures ANOVA. To control for family-wise error, significant findings were further analyzed with a Tukey post hoc test.29 An α level of .05 was used in all statistical analyses.

Clinical Evaluation 

Aside from using the 2-geared wheelchair wheel, none of the participants received additional therapeutic interventions, nor had changes in their analgesic or anti-inflammatory medications during the study. There were no differences in shoulder ROM, upper-extremity strength, or the incidence of specific shoulder diagnoses after use of MAGICWheels (P<.05). The mean glenohumeral elevation was greater than or equal to 150°, external rotation was greater than or equal to 80°, and internal rotation was greater than or equal to 45°. Muscle strength was greater than or equal to 4 out of 5 for all upper-extremity muscle groups tested (rotator cuff, serratus anterior, deltoid, upper trapezius, biceps, triceps).19 Clinical findings revealed that 2 participants had bilateral multidirectional instability and bilateral impingement signs. One participant had anterior instability. There were impingement signs in 14 of the 17 participants and were bilateral in 7 participants.

Wheelchair Users Shoulder Pain Index 

Although 17 participants began using MAGICWheels on their personal wheelchairs, only 13 completed all 5 months. Reasons for withdrawal from participation were: “not liking the wheels” (n=1, before week 8), automobile collision causing other injuries (n=1, before week 16), transportation issues (n=1, before week 20), and unrelated personal issues (n=1, before week 20). The mean time of the MAGICWheels phase was 18.1±4.3 weeks for all 17 subjects, with a 91% completion rate for the designated 20 weeks.

Shoulder pain was stable (P=.40) during the baseline phase (table 2). The mean WUSPI score for the baseline phase was 50.5±27.6. Because 4 subjects withdrew before the end of the full 20-week protocol, the data were carried forward and revealed a post-MAGICWheels phase WUSPI of 31.5±33.1. There was a significant reduction in the absolute change score (19±34, P=.035) and in the change expressed as a percentage of baseline 34.2%±64.3% (P=.044). There was a significant reduction (percentage of baseline) in shoulder pain with the MAGICWheels intervention at week 2 (P=.004) and that continued at every week through week 16 (P=.015) (table 3). The difference was not found at week 20 (P=.062), however, 1 participant reported a significant increase in pain during week 20 caused by factors unrelated to the study. This subject had experienced reduced pain throughout all previous weeks of the MAGICWheels phase. Significant reduction in shoulder pain (P=.003) resulted when the percentage change at week 20, compared with baseline, was calculated without this subject’s data (−54.3%±10.4%, n=12) (fig 2). A post hoc correlation analysis determined that there was no significant relation between duration of wheelchair use and pain reduction in any week of the MAGICWheels phase. There was poor correlation among the weeks and the duration of wheelchair use (r2 range, −.05 to 12).

Table 2. WUSPI Scores by Week During Baseline Phase (N=17)

	Week	Mean WUSPI Performance Adjusted Score ± SD	Range of WUSPI Performance Adjusted Score
	Enrollment week 0	53.9±24.3	16.4–82.5
	Baseline week 1	51.4±30.6	13.5–108.0
	Baseline week 2	48.9±28.7	12.8–98.9
	Baseline week 3	48.7±28.2	8.4–116.4
	Baseline week 4	49.5±33.8	7.2–100.6
	Baseline phase	50.5±27.6	15.4–95.9

Abbreviation: SD, standard deviation.

Table 3. WUSPI Absolute Score Change and Percentage Change From Baseline, by Week

	Week	n	Mean Absolute Change From Baseline ± SEM	P	Mean Percentage Change From Baseline ± SEM	P
	MW week 1	17	−8.8±5.9	.154	−24.5±11.7	.053
	MW week 2	17	−12.1±6.1	.063	−34.7±10.3⁎	.004
	MW week 3	17	−10.8±5.7	.077	−33.8±10.4⁎	.005
	MW week 4	17	−13.0±5.6⁎	.032	−37.2±10.7⁎	.003
	MW week 8	16	−16.6±5.3⁎	.007	−39.6±9.1⁎	.001
	MW week 12	16	−17.1±6.1⁎	.013	−40.5±11.3⁎	.003
	MW week 16	14	−22.0±7.8⁎	.015	−42.4±10.7⁎	.002
	MW week 20	12	−30.6±8.2⁎	.003	−54.3±10.4⁎	.003
	Retention week 1	12	−15.5±8.3	.087	−20.1±13.6	.169
	Retention week 2	12	−19.5±7.7⁎	.028	−29.2±13.4	.052
	Retention week 3	12	−7.9±8.2	.352	−5.7±18.5	.763
	Retention week 4	12	−17.6±9.6	.098	−24.9±18.2	.205

Abbreviations: MW, MAGICWheels 2-geared wheelchair wheel; SEM, standard error of mean.

⁎ Significant decrease vs baseline phase mean (P≤.05).

View Large Image Download to PowerPoint

Fig 2. Reduction in shoulder pain during 2-gear wheel phase and retention phase. NOTE. The reduction in WUSPI score is compared with baseline pain level (means and standard error of mean). Baseline is expressed as 100% pain with weeks during phase illustrated as a reduction. *Significant reduction from baseline.

Shoulder pain during the 4-week retention phase (n=12, the participant with increased pain at week 20 did not complete the retention data) showed a trend to increase, as indicated by increased WUSPI scores. There was not, however, a significant percentage increase in comparison with the final week of using the MAGICWheels (P<.05). The final WUSPI score (week 20) during the MAGICWheels phase was 26.3±7.6 and the mean of the 4-week retention phase was 37.1±7.7, which no longer showed a reduction in shoulder pain as compared with the baseline enrollment score (P=.33). There was no difference in percentage change WUSPI scores during the retention phase compared with baseline (see table 3). The mean WUSPI score during the retention phase did not differ (P=.45) from the baseline WUSPI (51.8±7.9).

Activity Level 

All 6 of the self-reported activities were stable throughout the 4-week baseline (range, P=.07 [for carpet] to P=.60 [for inclines and/or ramps]). There was a significant difference in activities among the 3 phases (table 4). Increased encounters with carpeted (P<.01) and grass (P<.001) surfaces occurred during the MAGICWheels phase compared with the baseline phase. Additionally, there were increases in hill encounters (P=.009), gravel (P=.03), and grass (P=.03) surfaces during the retention phase as compared with baseline.

Table 4. Activity Encounters by Phases

	Phase	Mean Baseline ± SEM	n	Mean MW Phase ± SEM	n	Mean Retention Phase ± SEM	n
	Inclines and/or ramps	19.6±1.3	14	19.1±1.1	14	20.7±1.9	9
	Hills	5.9±1.1	12	9.0±1.0	10	11.8±1.8⁎	6
	Carpet	5.5±1.4	8	11.5±1.2⁎	8	10.2±1.8	6
	Gravel	9.5±1.4	11	12.5±1.5	10	17.0±2.7⁎	4
	Curbs	6.7±0.8	9	7.5±0.8	8	7.8±1.2	6
	Grass	7.8±1.5	13	14.3±1.4⁎	10	13.4±2.6⁎	5

⁎ Significant increase vs baseline phase (P≤.05).

Wheelchair Users Functional Assessment 

There was no difference in WUFA scores after the 5-month use of the MAGICWheels; however, there was a strong trend toward improvement in function (pre, 79.2±0.41; post, 80.7±0.55; P=.06).

Incline Test 

There was no interaction between wheel type (personal wheels vs MAGICWheels) in time required to ascend the hill or in the reported RPE. There were no differences pre- and postuse of the MAGICWheels in time required to climb or the RPE. There was a main effect (P<.01), however, for wheel type in the time to climb the hill with the geared wheel taking significantly longer (personal wheels, 21.3±1.7s; MAGICWheels, 35.5±1.8s). The RPE did not differ between the wheel types (P=.13).

Study Limitations 

Internal validity of this study had several potential limitations. The lack of a control group, the use of a cross-sectional sample of convenience, and the specific inclusion criteria may have led to a selection bias of more motivated participants. Because many manual wheelchair users who experience shoulder pain do not seek any interventions, involvement in the study itself (Hawthorne effect36) may have been the reason for the reduction of pain, rather than the specific intervention of the MAGICWheels. Although only 1 participant who withdrew reported it was because of the wheels, the attrition within the study is a potential limitation.