Among the 82 participants, 19 patients with hip OA had fallen at least once in the past year. In addition, 53 patients had experienced a near-miss of a fall in the past year. Thirty-two percent had injuries when they fell, in the following order: fractures (5%), abrasions (12%) and bruises (15%). Among the 19 patients, the most common place of fall was outdoors (52%), followed by indoors (32%) and stairs (16%). Falls occurred during the day (67%), with only a few occurring at night (28%) and in the morning (5%). The most common direction of fall was forward (67%), followed by sideways (29%) and backward (4%). The main causes of falls were tripping (48%), losing balance (33%) and slipping (19%).
Data from 77 of the 82 participants were included in the analysis after incomplete data were excluded. Among these 77 participants, 17 had experienced falls in the past year. In terms of basic information, the faller group was significantly greater in terms of age (t(42) = -2.73, p = 0.009, r = 0.3873, 95% CI [-7.044, -1.056]) and FRI-21 score (t(75) = -2.11, p = 0.038, r = 0.2366, 95% CI [-2.945, -0.084]), was significantly greater, but there were no differences in the other groups (Table 1).
Table 1
Variables | Fallers(n=17) | Nonfallers (n=60) | p value | 95%CI | r |
Age(years) | 69.0 ± 4.6 | 65.0 ± 7.5 | 0.009 | -7.044 -1.056 | 0.387 |
Height(cm) | 153.8 ± 3.2 | 155.3 ± 6.0 | 0.202 | -0.783 3.612 | 0.180 |
Weight(kg) | 52.7 ± 8.7 | 54.3 ± 9.1 | 0.510 | -3.303 6.589 | 0.076 |
BMI(kg/m2) | 20.9(19.5,24.6) | 22.1(19.7,24.6) | 0.759 | | 0.034 |
History of present illness(years) | 5.2(2,10) | 6.0(2,10) | 0.951 | | 0.006 |
Fall risk index | 9.8 ± 2.6 | 8.3 ± 2.6 | 0.038 | -2.945 -0.084 | 0.237 |
mean ± SD, median (interquartile range) |
There were no significant differences in hip ROM, hip abductor muscle strength, or knee extension muscle strength between the faller and nonfaller groups on the affected and unaffected sides. Similarly, there were no significant differences in the 10-meter walking speed or number of steps between the groups (Supplementary Table S1). The overall JOA hip score was significantly lower in the affected (t(75) = 2.61, p = 0.011, r = 0.288, 95% CI [2.113, 15.767]) and unaffected (W = 673.5, p = 0.044, r = 0.228) groups than in the nonfaller group. In addition, in the JOA hip score subcategories, only the activities of daily living (ADL) score was significantly lower in the faller group than in the nonfaller group (W = 675, p = 0.038, r = 0.236) (Table 2).
Table 2
Comparison of physical function between fallers and nonfallers in patients with hip OA. The JOA hip score and MJS depict the Japanese Orthopedic Association hip score and the minimum joint space, respectively.
| Fallers(n=17) | Nonfallers (n=60) | p value | 95%CI | r |
Hip JOA score total Affected Unaffected | 52.2 ± 12.5 81.8(75,89) | 61.1 ± 12.5 86.8(83,93) | 0.011 0.044 | 2.113 15.767 | 0.288 0.228 |
Hip JOA score pain Affected Unaffected | 13.5(10,20) 38.5(35,40) | 17.8(10,20) 37.9(35,40) | 0.140 0.944 | | 0.168 0.007 |
Hip JOA score gait ability | 13.9(10,18) | 14.6(15,15) | 0.856 | | 0.020 |
Hip JOA score ADL | 14.9(12,18) | 16.8(16,20) | 0.038 | | 0.236 |
Minimum joint space(mm) Affected Unaffected | 0 3.3 ± 1.3 | 0 2.8 ± 1.0 | 0.085 | -1.096 0.073 | 0.199 |
mean ± SD, median (interquartile range) |
An example of each group and HC of the three dimensions estimated of the foot trajectory during walking, obtained from the shank, and an example of each group and HC of a graph of the waist trajectory visualizing the temporal changes in displacement of the waist during walking are shown in Fig. 2a-f.
In terms of gait parameters, the stride length of the unaffected side was significantly shorter in the faller group than in the nonfaller group (t(75) = 2.12, p = 0.037, r = 0.2385, 95% CI [0.006, 0.178]). There were no significant differences in maximum foot clearance, stride time, speed, stance, or swing phase times between the affected and unaffected sides (Fig. 3) or in lateral foot distance during the swing phase (Supplementary Table S2). There were also no significant differences in the standard deviations of gait parameters between the affected and unaffected sides (Supplementary Table S2). In terms of symmetry indices, gait asymmetry (GA) for speed was significantly lower in the faller group than in the nonfaller group (p = 0.049, r = 0.223) (Fig. 4a), but there were no significant differences in the other indices (Table 3). There were no significant differences in the coefficients of variation (CVs) for gait parameters between the faller and nonfaller groups (Supplementary Table S3).
Table 3
Comparison of asymmetry in gait parameters between faller and nonfaller patients with hip OA. Lateral distance (LD).
Gait asymmetry | Fallers(n=17) | Nonfallers (n=60) | p value | 95%CI | r |
Stride length GA | 0.99(0.97,1.01) | 0.98(0.97,0.99) | 0.116 | | 0.178 |
Maximum foot clearance GA | 0.98(0.85,1.06) | 0.98(0.88,1.03) | 1.000 | | -0.001 |
Stride time GA | 1.00 ± 0.01 | 1.00 ± 0.00 | 0.809 | -0.004 0.005 | 0.056 |
Speed GA | 0.99(0.98,1.00) | 0.98(0.97,0.99) | 0.049 | | 0.223 |
Stance time GA | 0.91 ± 0.08 | 0.93 ± 0.07 | 0.322 | -0.020 0.061 | 0.114 |
Swing time GA | 1.14 ± 0.12 | 1.11 ± 0.11 | 0.301 | -0.092 0.029 | 0.120 |
LD toe off GA | 0.67(0.34,0.92) | 0.69(0.27,1.41) | 0.632 | | 0.053 |
LD maximum foot clearance GA | 0.57(0.33,0.95) | 0.56(0.19,1.32) | 0.722 | | 0.040 |
LD kick out GA | 0.53(0.28,1.21) | 0.77(0.22,1.41) | 0.921 | | 0.011 |
LD swing down GA | 0.38(-0.49,0.85) | 0.75(0.17,1.54) | 0.067 | | 0.208 |
mean ± SD, median (interquartile range) |
There were no significant differences between the faller and nonfaller groups in terms of the magnitude of lateral sway, asymmetry of lateral swats or asymmetry in the amount of lifting in the hip orbit (Table 4). The amount of waist lifting on the affected side was significantly lower in the faller group than in the nonfaller group (p = 0.019, r = 0.266). The amount of waist lifting on the nonfaller side was also significantly lower in the faller group than in the nonfaller group (p = 0.025, r = 0.257) (Fig. 4b).
Table 4
Comparison of waist trajectory between fallers and nonfallers in patients with hip OA.
Waist trajectory | Fallers(n=17) | Nonfallers (n=60) | p value | 95%CI | r |
Lateral sway of the lumber | 3.16(2.90,4.31) | 3.52(2.78,4.62) | 0.425 | | 0.090 |
Asymmetry lateral sway of the lumber | 10.23(4.76,20.56) | 13.58(4.98,22.49) | 0.469 | | 0.082 |
Asymmetry of the amount of waist lifting | 20.42(7.39,41.78) | 23.87(11.33,44.09) | 0.876 | | 0.017 |
The amount of waist lifting affected unaffected | 2.75 ± 1.01 3.32 ± 1.07 | 3.39 ± 0.95 4.06 ± 1.22 | 0.019 0.025 | 0.106 1.163 0.095 1.399 | 0.266 0.257 |
mean ± SD, median (interquartile range) |