INTRODUCTION

The ability to transfer to and from a wheelchair is crucial to the completion of Activities of Daily Living (ADLs) for wheelchair users [1]. ADLs take place in specific environments that can shape the way in which the person carries out the transfer [2].  Despite this, studies on wheelchair transfers have mainly been carried out within laboratory settings [3,4], which are often not representative of real life conditions [5]. Few researchers have explored the characteristics of wheelchair transfers across different real-life conditions.

The study carried out by [6] describes the technique used by 29 paraplegic participants when transferring and loading the wheelchair into and out of their car. A great variability of leg and hand positioning was reported, resulting in changes to muscular activity and force generation. On the other hand, [7] focused on the kinematic analysis of the movements used during car transfers by 4 individuals with tetraplegia using a transfer board. All participants demonstrated a rotatory head-hip relationship (head moves in the opposite direction of the buttocks) to facilitate the transfer. However, various degrees and combinations of trunk and neck flexion were used to successfully transfer [7].  The study by [8] looked at the effect of two different toilet transfers configurations (front and side) on the transferring technique and associated upper limb load among 26 manual wheelchair users. Results showed that the front transfer setup allowed participants to perform better quality transfers that were associated with reduced upper limb loads. The only study with a broader scope of transfer conditions was a carried out by [2]. This study investigated how factors such as horizontal and vertical gaps, obstacles and grab bars affected the ability of 120 individuals to perform wheelchair transfers. However, the results were aimed at the creation of more accessible environments and the authors do not provide any information on how different conditions affected technique or perceived difficulty of wheelchair transfers.

Our study aimed to explore how wheelchair transfers’ technique changed across different settings and investigate the impact of different environmental conditions on objectively measured transfer quality, subjective perception of difficulty, and the relationship between the two.

METHODS

Participants

Thirteen participants, ten males and three females volunteered to take part in this study. Mean age was 39.5 ± 10.9 years, mean weight was 84.6 ± 20 kg and mean height was 178.9 ± 12.7. Ten participants had a SCI, nine reported a complete SCI and only one an incomplete SCI. Four of the participants with SCI had tetraplegia (C6-C7) and six had paraplegia (T1-T12).  Of the remaining three participants, one had a below the knee amputation (BKA) on the left leg, one had neuromyelitis optica (NMO) and one had transverse myelitis (TM).   All participants were residents in the local community. Eight participants reported using rigid frame manual wheelchairs as a primary means of mobility, three participants reported using foldable manual wheelchairs and two participants used electric wheelchairs. In accordance with the protocol approved by the Georgia Institute of Technology IRB, participants were asked to sign informed consent prior to participation.

Experimental Procedure

Four transfer scenarios were recreated for the study: car, toilet, bed and couch. All participants had to perform transfers in the bed scenario, where the height of the bed was set to match the height of the individual’s wheelchair. Participants where then able to choose two out of the other three scenarios proposed, according to the types of transfers they performed more frequently. The order of the three scenarios was randomized for each participant. Indoor scenarios are shown in Figure 1.  When a participant chose the car transfer scenario, the participant’s own vehicle was used. Four participants drove a Jeep car. However, only 2 transferred directly onto the driver’s seat while the other 2 -transferred onto a car lift. Four participants drove sedan cars, and two an adapted minivan fitted with a swivel seat and performed the transfers with the wheelchair inside the car. Car transfers are shown in Figure 2. Each participant was asked to perform up to four transfers (two from the wheelchair onto the target surface and two from the target surface onto the wheelchair seat) for each scenario. Adequate resting periods were given to avoid fatigue

Transfer strategy, height difference, perceived difficulty and objective quality

As participants used their own wheelchairs throughout the experiment, this resulted in different height gaps between the wheelchair seat and the other transfer surface. Height differences between start and target seat were measured and assigned to one of three categories: less than 5cm, between 5 and 15cm and greater than 15cm. Video recordings were collected using 2 USB Logitech C930e webcams (Logitech Europe S.A., CH) positioned at angles which allowed for all necessary views to be captured. Video recordings were used to analyse the quality of transfers performed by participants. At the end of each transfer, participants were asked to rate its difficulty using a modified CR 10 Borg Scale [9].

To objectively evaluate the quality of transfers, two trained physiotherapists independently evaluated each transfers using the Transfer Assessment Instrument (TAI). The Transfer Assessment Instrument (TAI) is described as a tool that can be used to assess safety, strategies for conservation of the upper limbs and ability to direct caregivers during the performance of independent and assisted wheelchair transfers[10]. Study participants were evaluated using only Part 1 of the TAI as used by [11]. Part 1 of the TAI is composed of 15 items with a dichotomous score concerning different aspects of transfer performance including wheelchair positioning and management, arm and hand position, and quality of body movement during transfer execution [12]. Disagreements over different scores was resolved through consensus meetings. Items 4, 5 and 15 of the TAI were removed from the evaluation as they were not applicable to more than 80% of the transfers performed by participants.

Statistical analysis

Descriptive statistics were calculated for demographic data of participants. Transfers were assigned to three different categories according to the technique used by the individual: standing, transfer board and sitting and only compared within the same category. The different environmental scenarios necessitated different height set ups, and the effect of both environmental scenarios and the height differences on the TAI scores were tested. TAI scores were checked for normality using the Shapiro Wilk test. Due to the non-normality of the data the Kruskal-Walllis test was first used to assess the impact of different transfer scenarios and height gaps on the TAI score and the self-reported perceived difficulty. Where significant differences were found a Dunn’s test with Bonferroni correction was used for post-hoc pairwise comparison. Spearman’s correlation was used to investigate the presence of a linear relationship between TAI score and perceived difficulty. The level of significance for all tests was set at 0.05. The statistical analysis was performed using SPSS 24 software (SPSS Inc., Chicago, IL, USA).

RESULTS

Transfer Characteristics

Data for 153 transfers were collected from 13 participants. Participant 1 only completed 1 of the 4 couch transfers, resulting in a total of 9 transfers. One hundred and twelve transfers were performed using a sitting technique, 26 with a standing technique and 15 with a transfer board.

Transfer technique was consistent across scenarios for 11 participants. However, Participant 2 used a sitting technique for 10 transfers and a standing technique twice. Participant 11 transferred 6 times with a transfer board 6 times without. Eighty-four of the transfers performed (52 bed, 12 toilet, 20 car) had a height gap of less than 5 cm. For 61 of the transfers performed (32 toilet, 16 car, 13 couch), the height difference was between 5 and 15 cm. The remaining 8 transfers (car) featured a height gap greater than 15 cm.

Effect of height gap and scenario on transfer quality and difficulty

Height gaps were only found to have a significant effect on TAI score (χ2(1) = 6.56, p = .01, η2 = .61) for transfers performed with a transfer board. The TAI score was significantly lower for transfers performed between surfaces featuring gaps of 5-15 cm (4.3 ± .88, p = .01) compared to level transfers (7.01 ± 1.23).  No other significant difference of height gaps on TAI score was observed for transfers performed with a sitting (p = .61, η2 = .01) or standing technique (p = .71, η2 = .03). Regardless of the transferring technique used, height gaps did not have any significant effect on the transfer reported difficulty (lowest p = .1).

A significant difference in the reported difficulties among transfers performed in different scenarios was found for sitting (χ2(2) = 7.84, p = .001, η2 = .126) and standing transfers (χ2(3) = 4.19, p = .017, η2 = .489). Among participants who transferred with a sitting technique, toilet transfers were judged to be significantly more difficult (2.17 ± .88) than bed transfers (1.47 ± .65, p = .001) and car transfers (1.63 ± .82, p = .012). When transfers were performed with a standing technique, car transfers were found to be significantly more difficult (3.5 ± .71) than bed (1 ± 0, p = .03) and toilet transfers (1 ± 0, p = .03). Among transfers completed with a transfer board, no significant differences were found in the reported difficulties across the different scenarios (p = .32, η2 = .184). However, the different scenarios had a significant effect on the TAI score attributed to transfers performed with a transfer board (χ2(2) = 6.629, p = .036, η2 = .612). Post hoc analysis showed that couch transfers received significantly lower TAI scores (4.3 ± .88), compared to bed (6.93 ± 1.29, p = .022) and car transfers (7.13 ± 1.32, p = .018).

No significant correlation was found between TAI score and reported difficulty for sitting (rs (110) = .1, p = .3), standing (rs (24) = .07, p = .72) and transfer board transfers (rs (13) = -.21, p = .45).

DISCUSSION

This is the first study to analyze quality and perceived difficulty of wheelchair transfers across various scenarios for participants using different transferring techniques. We found that the environmental scenario impacted both the type and quality of the transfer. Eleven participants exhibited only one technique whereas two participants demonstrated two different technique across various settings. Previous research [13] highlighted how the transferring technique used individuals was primarily dictated by the person’s medical condition, and that the use of assistive technologies (ATs) was linked to greater body weight. Our findings evolve the number of factors which can effect a person’s transferring technique. Specifically, we show that environmental conditions can represent either a barrier or a facilitator of transfer performance with important effects on the individual’s transferring technique and the use of ATs.

Ascending or descending transfers are linked to an increase in the upper limbs’ mechanical load [2,14]. Nevertheless, in the current study, the presence of a greater height gap was only found to have a negative effect on the TAI score attributed to transfers performed with a transfer board. Height gaps are likely to present a greater challenge when participants have limited transferring ability and normally people would then use a transfer board. Due to the design of the study featuring the performance of repeated transfers, recruited wheelchair users potentially had above average transferring ability. This could potentially explain why the reported difficulty and TAI scores attributed to transfers performed with sitting and standing techniques were not affected by the presence of height gaps.

Transferring scenario was found to significantly affect either the perceived difficulty or the TAI score, depending on the transferring technique. Participants performing standing transfers rated car transfers as more difficult. By contrast, when participants used a sitting technique bathroom transfers were found to be more challenging. This was particularly interesting as the accessible toilet used for this study was fitted with grab bars that are supposed to facilitate the performance of wheelchair transfers. However, as noted by [2] and [8] grab bars commonly fitted in accessible toilets are often too high to be useful to people performing sitting transfers and were scarcely used by participants. Overall, indoor scenarios resulted in similar characteristics between participants, whereas car transfers presented strategies that where considerably different, due to the uniqueness of the set up. This reinforce the findings from previous research that highlights the importance of further research into car transfers [6,7].

Interestingly, we were unable to find any significant relationship between the TAI score attributed by clinicians and the perceived difficulty reported by participants. This was surprising as previous studies by [15,16] found a strong correlation between the subjective estimation of effort and the objective evaluation of individual’s capacity across several wheelchair skills. However, [16] found that self-assessment of safety did not correspond to objective evaluation by trained professionals. As transfers and other wheelchair skills become an integral part of the daily routine, individuals could become used to performing these tasks with strategies that are non-efficient or potentially dangerous. Overall, our results show that to gather accurate insights from studies looking at the performance of wheelchair transfers, objective tools should be used to assess transfers’ quality.

Although this study illustrates novel insights concerning the performance of wheelchair transfers across different scenarios, it has limitations that need to be considered when interpreting the results. The small number of participants, although not uncommon among studies focusing on wheelchair users, mean caution should be taken in generalization of the results. Furthermore, the majority of participants had a spinal cord injury (SCI) which creates additional challenges for generalization.  Although the scenarios featured in the study are found in the real world, the indoor scenarios might be different from the one found in participants’ houses. Finally, the physiotherapists evaluating participants’ transfers in the current study expressed concerns about the validity of the TAI with respect to the different scenarios as items related to upper limb positioning were judged to be less applicable for transfers performed to and from the car due to specific environmental constraints.

CONCLUSION

Environmental constraints associated with various wheelchair transfers scenarios can lead to considerable changes in the movement strategy and the quality of the transfer, potentially increasing or reducing the risks of falls and upper limb injuries. This is particularly true for car transfers, as the configuration of different vehicles can create unique set ups that will determine how individuals perform transfers. The lack of a significant relationship between the perceived difficulty and the objective quality of the transfer could be due to the fact that “incorrect movement strategies” become part of the individual technique among participants living in the community.