无监督聚类揭示脊髓损伤患者轮椅推进策略的多种个人和临床决定因素
Unsupervised clustering reveals multiple personal and clinical determinants of wheelchair propulsion strategies in SCI
研究通过无监督聚类分析,基于个人和临床特征对90名脊髓损伤患者进行分类,发现不同组别在年龄、BMI、时间伤后、损伤程度和活动水平等方面存在显著差异。
Introduction
Mobility with a wheelchair is crucial for the quality of life and health of people with spinal cord injuries (SCI) [1]. The way a wheelchair is propelled can have a significant impact on long-term musculoskeletal health [2,3,4,5]. Because wheelchair propulsion is highly repetitive, it places substantial demands on the upper limbs, making the optimization of propulsion biomechanics an important focus in rehabilitation and clinical practice [6, 7].
Wheelchair propulsion can be divided into the push (contact) phase, in which mechanical force is transferred to the wheel, and the recovery phase, in which the hand is repositioned for the next movement. In contrast to the push phase, the hand is significantly less restricted during the recovery phase and can follow various movement paths (e.g., without leaving the handrim or hanging loosely downwards). The resulting hand movement pattern can be classified into four distinct types based on hand movement trajectories [8]: arc (AR), single loop (SL), double loop (DL) and semicircle (SC). The pattern can be categorized into over-rim (AR and SL) and under-rim (SC and DL) propulsion pattern (e.g., [9]).
Despite growing evidence on the biomechanical determinants of wheelchair propulsion, a fundamental methodological challenge persists: the SCI population is inherently heterogeneous. Individuals differ not only in their neurological injury level and ISNCSCI classification (or more commonly referred to as the ASIA Impairment Scale (AIS)), but also in age, body composition, wheelchair use duration, and level of physical activity. Factors that each potentially can influence propulsion biomechanics (e.g., [10,11,12,13]). This multidimensional heterogeneity makes it difficult to form comparable groups based on any single characteristic. Traditional approaches that stratify participants by injury level alone risk masking the true sources of biomechanical variability: two individuals classified as paraplegic may propel their wheelchair in fundamentally different ways depending on their time since injury or habitual activity level, while two individuals with vastly different injury levels may exhibit remarkably similar propulsion strategies due to shared functional capacities. In other words, clinically defined categories do not necessarily correspond to biomechanically meaningful subgroups.
A first step toward addressing this limitation was taken by Kim et al. [14], who applied unsupervised clustering to joint angle time-series data from 28 SCI wheelchair users, identifying three kinematically distinct groups that differed primarily in physical activity level and time since injury rather than injury level. However, demographic and clinical characteristics such as age, gender, or body mass index (BMI) were only examined post-hoc as descriptors of the resulting clusters, and the small sample size (e.g., four participants in cluster 1) limited the generalizability of these exploratory findings.
The present study builds on and extends this work by inverting the logic of classification. Previous studies identified clusters based on kinematic data and then characterized individuals belonging to each group. In the present study, personal and clinical characteristics (age, BMI, arm length, wheelchair use duration, physical activity level, and injury level), together with the observed propulsion pattern, are used directly as input variables for an unsupervised clustering analysis. The aim of this study was to identify distinct subgroups of SCI manual wheelchair users through cluster analysis based on personal, clinical, and biomechanical propulsion characteristics, and to examine whether these subgroups differ systematically in their demographic-clinical profiles and propulsion biomechanics. We hypothesized that subgroups would differ not only in injury-related parameters, but also in personal factors, and that these differences would be reflected in propulsion biomechanics. To this end, a cohort of 90 SCI manual wheelchair users was analyzed.
Methods
Participants
This cross-sectional observational study included 90 individuals with SCI who were recruited between March 2023 and February 2026 in the Department of Orthopaedic Surgery, Klinikum Bayreuth GmbH, Germany (Table 1). Of the 91 participants initially assessed, one was excluded due to incomplete AIS data. The following inclusion criteria applied to study participants: diagnosed with paraplegia or tetraplegia, age ≥ 18 years, dependence on a manual wheelchair for daily mobility, and the ability to independently push the wheelchair at least 100 m. Exclusion criteria were pain and obvious asymmetrical wheelchair propulsion (evaluated by two experienced wheelchair researchers). All participants provided written informed consent to participate in the study. The investigation was approved by the ethics review board of the Friedrich-Alexander-University Erlangen-Nuremberg, Germany (23-20-B; 2023-02-14) and was in accordance to the Declaration of Helsinki.
Neurologic level of injury and extent of paralysis were determined by International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) examination [15] for participants admitted directly to our clinic following spinal cord injury. For participants who transferred to our Medical Center for Adults with Disabilities (MZEB) after initial rehabilitation elsewhere, this information was extracted from clinical records. Physical activity level was assessed via self-report and classified into three categories based on the frequency, duration, and intensity of aerobic and resistance exercise, following previously published criteria for adults with spinal cord injury [16]. Participants were classified as inactive if they reported no moderate-to-vigorous aerobic or resistance activity; as recreational if they reported some activity but did not meet the full guideline criteria; and as athletic if they met the established guideline for both aerobic exercise (≥2 sessions/week, ≥20 min of moderate-to-vigorous intensity aerobic exercise per session) and resistance exercise (≥2 sessions/week).
Measurement
The measurements took place on a test stand in the gait and movement laboratory at Klinikum Bayreuth GmbH, Germany [17, 18]. All participants used their own wheelchairs and were given some time to familiarize themselves with the new sensation of rolling in place. Once the participants were accustomed to the test stand, they were instructed to propel the wheelchair at their preferred speed for 20 s. Data recording only began after several propulsion cycles had been completed and a steady speed had been reached. The wheelchair propulsion itself was recorded with ten infrared cameras (200 Hz) of a 3D marker-based system (Vicon, Oxford, UK). To record the wheelchair propulsion pattern, one reflective marker was applied to the skin projection of the right second metacarpophalangeal joint. To classify the propulsion pattern and to analyze propulsion characteristics (i.e., push frequency, velocity, cycle time, push and recovery time as well as push angle) two further markers were attached to the right wheel, one on the wheel axle and the other on the spokes.
Data processing
Data processing was conducted using Python (version 3.13.4; PyCharm IDE, JetBrains, Prague, Czech Republic). Raw data were filtered using a bidirectional second-order low-pass Butterworth filter with a cutoff frequency of 6 Hz [19]. Push and recovery phases were identified using a wheel acceleration-based approach as previously described [18, 20]: Angular acceleration was derived from wheel marker trajectories. Transitions from positive to negative angular acceleration indicated the start of the push phase and the end of the recovery phase, whereas transitions from negative to positive values indicated the end of the push phase and the start of the recovery phase. Based on the wheel marker trajectories and the identified propulsion phases, all propulsion characteristics were calculated, including mean wheelchair velocity over the entire measurement period, push frequency, cycle time, push and recovery phase durations, and push angle, defined as the angular change of the pushrim during the push phase.
For the classification of propulsion patterns, additional hand marker data were used. The maximum hand distances above (dabove) and below (dbelow) the pushrim during the recovery phase were determined (Fig. 1). Propulsion patterns were then assigned based on these hand–pushrim distances: AR when dabove < 7.5 cm and dbelow < 7.5 cm; SL when dabove > 7.5 cm and dbelow < 7.5 cm; and SC and DL when dbelow > 7.5 cm. Thresholds were established through consensus review of all individual hand trajectories by two experienced researchers. The trajectories were visually inspected, and the 7.5 cm criterion was identified as the value that best distinguished pattern types across the sample. This approach is consistent with the threshold-setting procedure used by Slowik et al. [21], in which cut-off values were determined by optimizing agreement with expert rater classification.
The propulsion patterns were classified based on hand–pushrim distances (dabove, dbelow).
Clustering and statistical analysis
Unsupervised clustering and statistical analyses were performed with SPSS 31 (Chicago, IL, USA). To identify groups of clusters within the collected dataset, an unsupervised two-step cluster analysis was performed using a mix of categorical (i.e., gender, neurologic level of injury, extent of paralysis, physical activity level and propulsion pattern) and continuous (i.e., age, BMI, armlength and time since injury) variables (Table 1). The optimal number of clusters was determined using the Bayesian Information Criterion (BIC). The model with the lowest BIC value was obtained with three clusters, indicating that the optimal number of clusters is three (Fig. 2).
The Bayesian Information Criterion (BIC) values indicate that the optimal number of clusters is three.
Descriptive statistics were obtained for the total sample (Table 1) and derived clusters (Tables 2, 3) for all categorial and continuous variables. To examine the differences between the clusters, an analysis of variance (ANOVA) was performed for continuous variables. Partial Eta Squared was used to calculate the effect size. Values of 0.01 were considered as weak effects, 0.06 as moderate effects, and 0.14 as strong effects [22]. Post-hoc tests (Bonferroni) were conducted to identify specific between-cluster differences. Alternatively, Fisher’s exact test was performed for categorical variables. For Fisher’s exact test effect size was calculated using Cramer’s V. According to Cohen, a Cramer’s V of 0.1 were considered as weak effects, 0.3 as moderate effects, and 0.5 as strong effects [22]. An alpha level of 0.05 was used for all statistical tests.
Results
We identified three distinct groups of SCI manual wheelchair users. Table 2 represents the demographic characteristics of each cluster. Significant differences between the clusters were found in age, BMI, time since injury, location and extent of paralysis, and physical activity (Table 2). A comparison of Cluster 1 and Cluster 2 showed that the participants in Cluster 2 were on average 12 years younger, had been dependent on a wheelchair for significantly longer (an average of 20 years), had a lower BMI and were more physically active than the participants in Cluster 1. Furthermore, 30 of the 31 participants in Cluster 1 had incomplete motor paralysis, while 23 of the 31 participants in Cluster 2 were completely motor-paralyzed. A comparison of Cluster 1 and Cluster 3 showed that the participants in Cluster 3 had been dependent on a wheelchair for significantly longer and were more physically active than the participants in Cluster 1. Furthermore, the levels of paralysis differed among the participants, and the occurrence of complete motor paralysis was significantly higher in the participants of Cluster 3 (28 of 28 participants) than in those of Cluster 1 (one of 31 participants). Cluster 2 differed from Cluster 3 in terms of the location and extent of paralysis as well as the activity level observed (Table 2).
Figure 3 illustrates the distribution of the different propulsion patterns in Clusters 1–3. In Cluster 1, 24 of the participants showed the AR pattern, six the SL pattern and only one an under-rim pattern (SC & DL). In Cluster 2, only five of the participants showed the AR pattern, six the SL pattern and 20 an under-rim pattern 4. In Cluster 3, 17 of the participants showed the AR pattern and 11 the SL pattern. However, the distribution of the patterns differs significantly between Cluster 1 and Cluster 2 and between Cluster 2 and Cluster 3 (Table 3). In addition to the significant cluster differences regarding the propulsion patterns, significant differences in velocity and push angle were also observed. Here too, the both parameters only differed significantly between Cluster 1 and Cluster 2 and between Cluster 2 and Cluster 3 (Table 3). There are no significant differences of propulsion characteristics between Cluster 1 and Cluster 3.
The scatter plot shows the distribution of individual propulsion pattern across clusters.
Discussion
We identified three distinct groups of SCI manual wheelchair users based on personal and clinical characteristics including age, BMI, arm length, wheelchair use duration, physical activity level, and injury level alongside the observed propulsion pattern. The resulting clusters were well balanced. The smallest cluster comprised 28 participants, which is consistent with recent simulation-based recommendations of approximately 20–30 observations per cluster for reliable cluster recovery [23]. Significant differences between the groups were found in age (strong effect), BMI (weak effect), time since injury (weak effect), location (moderate effect) and extent of paralysis (strong effect), and physical activity (strong effect; Table 2). Previous studies have focused, for example, on comparing wheelchair users with different injury levels (paraplegia vs. tetraplegia). However, comparisons based on injury level may overlook the heterogeneity of wheelchair users with spinal cord injury, which can also be influenced by additional factors such as age, BMI, wheelchair use duration, and physical activity level. In a recent bottom-up classification of manual wheelchair users, Kim et al. [14] investigated how propulsion strategies relate to injury level, physical activity level, and neuromuscular coordination. Due to the small sample size, their exploratory findings suggest that wheelchair propulsion strategies are influenced more by time since injury and physical activity level than by injury level. Our study, conducted on a cohort of 90 SCI manual wheelchair users, allowed us to classify groups that differed in terms of wheelchair use duration and physical activity level but also in terms of injury level and extend of paralysis.
Cluster 1 and Cluster 3 most frequently exhibited the over-rim propulsion pattern (AR and SL) and did not differ in their propulsion characteristics. However, both clusters differed in their demographic and clinical profiles. Participants in Cluster 1 were older, had the shortest time since injury, and consisted of a mix of inactive and recreationally active participants. They mainly exhibited incomplete motor paralysis of varying severity. Participants in Cluster 3 showed an intermediate mean age and consisted of recreationally active participants, a longer time since injury, and exclusively included individuals with complete motor paraplegia. Participants in Cluster 2 were the youngest and most physically active group, with the longest time since injury. Despite including the most individuals with tetraplegia, this cluster most frequently exhibited under-rim propulsion patterns (SC and DL) and showed significantly higher velocities and larger push angles compared to the other clusters.
Larger push angles are indicative of the SC and DL patterns compared to the AR and SL patterns [24]. Additionally, initial experiments on wheelchair propulsion have shown that the under-rim patterns results in the lowest joint accelerations during propulsion [25, 26]. Previous studies suggested that under-rim pattern results in a lower cadence, lower peak forces, and better separation between the push and recovery phases (e.g., [8, 24, 27]). Due to these factors, current clinical guidelines recommend the use of the under-rim pattern [6]. In line with this, a simulation study indicated that the SC pattern may reduce the risk of secondary injuries such as rotator cuff tears [28]. Interestingly, under-rim propulsion patterns were also observed in Cluster 2 participants with tetraplegia (n = 2 with complete C4–C8 motor lesions and n = 7 with incomplete C4–C8 motor lesions). These findings suggest that factors such as physical activity level and time since injury may play an important role in shaping propulsion strategies, potentially alongside or beyond injury level and ISNCSCI classification.
It is widely recognized that beginners tend to use the over rim pattern and that Cluster 1, characterized by the shortest time since injury, follows this pattern [26, 29]. Nevertheless, it remains unclear why individuals in Cluster 3 fail to adopt the under-rim pattern, despite having also considerably time since injury than Cluster 1. There are several possible explanations for this deviation. Firstly, the SC pattern may not be intuitive enough to be used even with prolonged wheelchair use. In contrast, individuals in Cluster 2 showed higher overall physical activity levels, which may be associated with greater familiarity with efficient propulsion techniques. Secondly, it is possible that individuals in Cluster 3 do not execute the SC pattern efficiently enough or lack the physical prerequisites to benefit from its biomechanical advantages. This assumption is supported by studies involving physically healthy beginners who report that the AR pattern may be more efficient in the early stages of wheelchair propulsion [26, 29].
Several limitations of the present study should be considered. First, the sample consisted predominantly of male participants. Although this reflects the higher prevalence of male wheelchair users in the SCI population (e.g., [30]), a more balanced gender distribution would improve the generalizability of the findings to female users. Second, this study was conducted exclusively on a stationary test stand, which may not fully reflect real-world overground propulsion conditions, where additional factors such as trunk movement, environmental variability, and surface interactions may influence propulsion strategies. Third, the distance-based thresholds used for propulsion pattern classification have not yet been independently validated. While the 7.5 cm criterion was established through consensus visual inspection of all trajectories by two experienced researchers, future studies should examine the agreement between this approach and other methods such as the NRT/TRT parameters proposed by Slowik et al. [21]. Fourth, push and recovery phases were determined based on sign changes in wheel angular acceleration. This method has not yet been validated for measurements on a roller system and may result in small deviations in phase determination compared to force-based reference methods (e.g., SmartWheel). In addition, potentially relevant factors such as spasticity, pain, sitting position, or patterns of wheelchair use in daily life were not included in the clustering analysis, although they may further influence locomotion strategies [31,32,33,34]. Consequently, the identified clusters depend on the selected variables and should not be considered definitive categorizations, but rather as an initial step toward a more comprehensive characterization of propulsion behavior in SCI wheelchair users.
Data availability
Data will be made available upon reasonable request to the corresponding author.
References
Ferretti EC, Curi HT, Andrade LF, Cooper RA, de Soárez PC. Conceptual mapping proposed to comprehend the effect of wheelchair mobility on social participation and quality of life: a systematic review. Disabil Rehabil Assist Technol. 2024;19:814–30. https://doi.org/10.1080/17483107.2022.2126904
Walford SL, Requejo PS, Mulroy SJ, Neptune RR. Predictors of shoulder pain in manual wheelchair users. Clin Biomech. 2019;65:1–12. https://doi.org/10.1016/j.clinbiomech.2019.03.003
Briley SJ, Vegter RJK, Goosey-Tolfrey VL, Mason BS. The longitudinal relationship between shoulder pain and altered wheelchair propulsion biomechanics of manual wheelchair users. J Biomech. 2021;126:110626. https://doi.org/10.1016/j.jbiomech.2021.110626
Liampas A, Neophytou P, Sokratous M, Varrassi G, Ioannou C, Hadjigeorgiou GM, et al. Musculoskeletal pain due to wheelchair use: a systematic review and meta-analysis. Pain Ther. 2021;10:973–84. https://doi.org/10.1007/s40122-021-00294-5
Salzmann IP, Rietveld T, Togni R, Briley SJ, Goosey-Tolfrey VL, de Vries WHK, et al. Exploring the biomechanical link between wheelchair propulsion, shoulder injury and shoulder pain: a scoping review. J Biomech. 2025;185:112678 https://doi.org/10.1016/j.jbiomech.2025.112678
Paralyzed Veterans of America Consortium for Spinal Cord Medicine. Preservation of upper limb function following spinal cord injury: a clinical practice guideline for health-care professionals. J Spinal Cord Med. 2005;28:434–70. https://doi.org/10.1080/10790268.2005.11753844
Morgan KA, Tucker SM, Klaesner JW, Engsberg JR. A motor learning approach to training wheelchair propulsion biomechanics for new manual wheelchair users: a pilot study. J Spinal Cord Med. 2017;40:304–15. https://doi.org/10.1080/10790268.2015.1120408
Boninger ML, Souza AL, Cooper RA, Fitzgerald SG, Koontz AM, Fay BT. Propulsion patterns and pushrim biomechanics in manual wheelchair propulsion. Arch Phys Med Rehabil. 2002;83:718–23. https://doi.org/10.1053/apmr.2002.32455
Walford SL, Rankin JW, Mulroy SJ, Neptune RR. The relationship between the hand pattern used during fast wheelchair propulsion and shoulder pain development. J Biomech. 2021;116:110202. https://doi.org/10.1016/j.jbiomech.2020.110202
Raina S, McNitt-Gray J, Mulroy S, Requejo P. Effect of choice of recovery patterns on handrim kinetics in manual wheelchair users with paraplegia and tetraplegia. J Spinal Cord Med. 2012;35:148–55. https://doi.org/10.1179/2045772312Y.0000000013
Dysterheft J, Rice I, Learmonth Y, Kinnett-Hopkins D, Motl R. Effects of daily physical activity level on manual wheelchair propulsion technique in full-time manual wheelchair users during steady-state treadmill propulsion. Arch Phys Med Rehabil. 2017;98:1374–81. https://doi.org/10.1016/j.apmr.2017.01.007
Gorce P, Louis N. Wheelchair propulsion kinematics in beginners and expert users: influence of wheelchair settings. Clin Biomech. 2012;27:7–15. https://doi.org/10.1016/j.clinbiomech.2011.07.011
Hong S, Kim H, Ahn J, Park W. Influence of time since injury and physical activity level on upper limb kinematics and muscle activation during wheelchair propulsion in complete T12/L1 spinal cord injury. BMC Musculoskelet Disord. 2025;26:807. https://doi.org/10.1186/s12891-025-08987-0
Kim H, Hong S, Park W, Ahn J. A bottom-up classification of wheelchair users based on propulsion kinematics beyond injury level: exploratory findings. IEEE Trans Neural Syst Rehabil Eng. 2025;33:3878–87. https://doi.org/10.1109/TNSRE.2025.3613726
Rupp R, Biering-Sørensen F, Burns SP, Graves DE, Guest J, Jones L, et al. International standards for neurological classification of spinal cord injury: revised 2019. Top Spinal Cord Inj Rehabil. 2021;27:1–22. https://doi.org/10.46292/sci2702-1
Martin Ginis KA, van der Scheer JW, Latimer-Cheung AE, Barrow A, Bourne C, Carruthers P, et al. Evidence-based scientific exercise guidelines for adults with spinal cord injury: an update and a new guideline. Spinal Cord. 2018;56:308–21. https://doi.org/10.1038/s41393-017-0017-3
Salm L, Schreff L, Benker C, Abel R, Müller R. The effect of a cognitive dual-task on the control of wheelchair propulsion. PLoS One. 2025;20:e0317504 https://doi.org/10.1371/journal.pone.0317504
Karner L, Schreff L, Abel R, Müller R. Validity of IMUs in comparison to a marker-based-motion capture system for spatio-temporal parameters during wheelchair propulsion. Sensors. 2025;25:4676. https://doi.org/10.3390/s25154676
Cooper RA, DiGiovine CP, Boninger ML, Shimada SD, Koontz AM, Baldwin MA. Filter frequency selection for manual wheelchair biomechanics. J Rehabil Res Dev. 2002;39:323–36.
de Vries WHK, van der Slikke RMA, van Dijk MP, Arnet U. Real-life wheelchair mobility metrics from IMUs. Sensors. 2023;23:7174.
Slowik JS, Requejo PS, Mulroy SJ, Neptune RR. The influence of speed and grade on wheelchair propulsion hand pattern. Clin Biomech. 2015;30:927–32. https://doi.org/10.1016/j.clinbiomech.2015.07.007
Cohen, J. Statistical power analysis for the behavioral sciences, 2. ed., reprint; Psychology Press: New York, NY, 2009, ISBN 0805802835.
Dalmaijer ES, Nord CL, Astle DE. Statistical power for cluster analysis. BMC Bioinforma. 2022;23:205. https://doi.org/10.1186/s12859-022-04675-1
Kwarciak AM, Turner JT, Guo L, Richter WM. The effects of four different stroke patterns on manual wheelchair propulsion and upper limb muscle strain. Disabil Rehabil Assist Technol. 2012;7:459–63. https://doi.org/10.3109/17483107.2011.650781
Shimada SD, Robertson RN, Bonninger ML, Cooper RA. Kinematic characterization of wheelchair propulsion. J Rehabil Res Dev. 1998;35:210–8.
de Groot S, Veeger HEJ, Hollander AP, van der Woude LHV. Effect of wheelchair stroke pattern on mechanical efficiency. Am J Phys Med Rehabil. 2004;83:640–9. https://doi.org/10.1097/01.PHM.0000133437.58810.C6
Qi L, Wakeling J, Grange S, Ferguson-Pell M. Patterns of shoulder muscle coordination vary between wheelchair propulsion techniques. IEEE Trans Neural Syst Rehabil Eng. 2014;22:559–66. https://doi.org/10.1109/tnsre.2013.2266136
Walford SL, Rankin JW, Mulroy SJ, Neptune RR. Differences in glenohumeral joint contact forces between recovery hand patterns during wheelchair propulsion with and without shoulder muscle weakness: a simulation study. J Biomech Eng. 2024;146:41005 https://doi.org/10.1115/1.4064590
Bickelhaupt B, Oyama S, Benfield J, Burau K, Lee S, Trbovich M. Effect of wheelchair stroke pattern on upper extremity muscle fatigue. PM R. 2018;10:1004–11. https://doi.org/10.1016/j.pmrj.2018.03.022
Taşoğlu Ö, Koyuncu E, Daylak R, Karacif DY, İnce Z, Yenigün D, et al. Demographic and clinical characteristics of persons with spinal cord injury in Turkey: one-year experience of a primary referral rehabilitation center. J Spinal Cord Med. 2018;41:157–64. https://doi.org/10.1080/10790268.2016.1224215
Arnet U, Bossuyt FM, Beirens BJH, de Vries WHK. Shoulder pain in persons with tetraplegia and the association with force application during manual wheelchair propulsion. Arch Rehabil Res Clin Transl. 2024;6:100310. https://doi.org/10.1016/j.arrct.2023.100310
Richter WM. The effect of seat position on manual wheelchair propulsion biomechanics: a quasi-static model-based approach. Med Eng Phys. 2001;23:707–12. https://doi.org/10.1016/s1350-4533(01)00074-1
Mâsse LC, Lamontagne M, O’Riain MD. Biomechanical analysis of wheelchair propulsion for various seating positions. J Rehabil Res Dev. 1992;29:12–28. https://doi.org/10.1682/jrrd.1992.07.0012
Braaksma J, Vegter RJK, Leving MT, van der Scheer JW, Tepper M, Woldring FAB, et al. Handrim wheelchair propulsion technique in individuals with spinal cord injury with and without shoulder pain: a cross-sectional comparison. Am J Phys Med Rehabil. 2023;102:886–95. https://doi.org/10.1097/PHM.0000000000002240
Acknowledgements
The present work is part of the doctoral thesis of Christian M. Benker and was performed in (partial) fulfilment of the requirements for obtaining the degree “Dr. med.”.
Funding
The author(s) received no specific funding for this work. Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Universitätsklinikum Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany
Christian M. Benker, Rainer Abel & Roy Müller
Department of Orthopedic Surgery, Klinikum Bayreuth GmbH, Bayreuth, Germany
Christian M. Benker, Lucas Schreff, Lisa-Eyleen Koch, Rainer Abel & Roy Müller
Bayreuth Center of Sport Science, University of Bayreuth, Bayreuth, Germany
Roy Müller
Movement and Exercise Science, Friedrich Schiller University Jena, Jena, Germany
Roy Müller
Authors
- Christian M. Benker
- Lucas Schreff
- Lisa-Eyleen Koch
- Rainer Abel
- Roy Müller
Contributions
CB contributed to the initial study design, performed measurements, interpreted the data, and wrote the original draft of the manuscript. LS contributed to the initial study design, performed measurements, calculated propulsion parameters, interpreted the data, and reviewed and edited the manuscript. LK was responsible for project administration, interpreted the data, and reviewed and edited the manuscript. RA provided resources, interpreted the data, and reviewed and edited the manuscript. RM designed the study, performed measurements and statistical analysis, interpreted the data, and wrote the original draft of the manuscript.
Corresponding author
Correspondence to Roy Müller.
Ethics declarations
Competing interests
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Ethical approval
This study was conducted in accordance with the Declaration of Helsinki and approved by the ethics review board of the Friedrich-Alexander-University Erlangen-Nürnberg, Germany (23-20-B).
Informed consent statement
Informed consent was obtained from all participants involved in the study.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Benker, C.M., Schreff, L., Koch, LE. et al. Unsupervised clustering reveals multiple personal and clinical determinants of wheelchair propulsion strategies in SCI. Spinal Cord (2026). https://doi.org/10.1038/s41393-026-01293-z
Received: 16 April 2026
Revised: 30 September 2026
Accepted: 08 October 2026
Published: 26 October 2026
Version of record: 26 October 2026
DOI: https://doi.org/10.1038/s41393-026-01293-z
来源:Nature 健康护理主题 · nature.com