Subjects
Fifteen able-bodied people were recruited to participate in this exploratory study (8 men, 7 women; age, 22.5 ± 1.8 years; weight, 65.2 ± 10.6 kg; height, 168.5 ± 8.9 cm; body mass index, 22.8 ± 2.6 kg/m2). Subjects with preexisting musculoskeletal disorders and spinal pathologies were excluded. All subjects read and signed an informed consent form that explained the study objective and experimental protocol. This study was approved by the Institutional Review Board of National Cheng Kung University Hospital.
Experimental wheelchair
The researcher of this study designed an experimental wheelchair, which was installed with airbags providing adjustable support for lumbar and femur areas. The size of each airbag was 40 × 23 cm2. When filled with air, the airbags were 4 cm thick. A customized microprocessor was used to adjust the extent and cycle period of filling or deflating each airbag to periodically change the sitting postures. A 1-cm-thick foam pad was installed to the backrest and seat to minimize the discomfort caused by the surface discontinuity between the skin contact with the backrest and seat.
Experimental protocol
Before the experiment started, each subject was first asked to rest their upper body on the backrest and relax their arms on both sides. Moreover, they had to keep their thighs parallel to the ground, feet approximately shoulder-width apart and firm on the footrest, and eyes looking straight ahead [16, 17]. Afterward, the sequence of three DSTs was randomly drawn by each subject (illustrated in Fig. 1): (1) LPDS: An airbag providing adjustable support for lumbar area was placed at L3 (on the subject). The airbag configuration was periodically switched between deflated (0-cm-thick) and filled (4-cm-thick), alternating every 5 min, the total period of experiment lasted for 20 min. (2) FUDS: An airbag providing adjustable support for femur area was placed at the middle of the subject’s thighs. The airbag configuration was periodically switched between deflated (0-cm-thick) and filled (4-cm-thick), alternating every 5 min, the total period of experiment lasted for 20 min. (3) LFDS: This technique is a combination of the previous 2 techniques simultaneously. Between each DST experiment, the subject took a 5-min break during which they could stand up and walk around.
Data recordings and analysis
Two pressure mapping mats (Body Pressure Measurement System; Tekscan Inc, South Boston, Massachusetts, USA) were placed over the surface of the backrest and seat to measure the pressure distribution along the human-wheelchair interface. The pressure mapping mats comprised 2016 (48 × 42) measuring cells. Each measuring cell had a dimension of 10.16 × 10.16 mm2. The interface pressure parameters were calculated using the body pressure measurement system research software (BPMS, version 7.02C; Tekscan Inc, South Boston, Massachusetts, USA). Data were sampled at the frequency of 30 Hz. From the interface pressure recordings, the dynamic contact area (DCA), dynamic average pressure (DAP), and dynamic peak pressure (DPP) on the whole backrest (B-DCA, B-DAP, and B-DPP), the entire seat (S-DCA, S-DAP, and S-DPP), the back part of seat (BS-DCA, BS-DAP, and BS-DPP), and the front part of seat (FS-DCA, FS-DAP, and FS-DPP) were calculated. Each interface pressure parameter is given as the averaging value over the entire 20 min sitting trial.
Statistical analysis
The Statistics Package for the Social Sciences (SPSS, version 17; SPSS Institute, Chicago, IL, USA) was used for all statistical analyses. Firstly, the whole backrest (B-DCA, B-DAP, and B-DPP) and the entire seat (S-DCA, S-DAP, and S-DPP) were compared between the different DSTs (LPDS, FUDS and LFDS) by using a Friedman test. A post hoc test, Wilcoxon signed-rank test was used for detecting significant differences in the dependent variables across the tests. Secondly, the back part of seat (BS-DCA, BS-DAP, and BS-DPP) and the front part of seat (FS-DCA, FS-DAP, and FS-DPP) were compared between the three DSTs by using the same statistical methods as above. The statistical significance was set at P < 0.05.