Assessment of risk of tripping before and after crossing obstacles under dimmed lighting conditions

Abstract

BACKGROUND:

Tripping and falling are common at work. Investigating the perceived risk of tripping is important for the safety of workers.

OBJECTIVE:

To test the hypotheses that the perceived risk of tripping is affected by obstacle depth, obstacle height, number of obstacle, and light location under dimmed lighting conditions.

METHODS:

A walkway with one to three obstacles in the middle was prepared. Each obstacle had a height of 0, 5, or 10 cm and a depth of 1 or 10 cm. The laboratory was dimmed with only one light either at the beginning, the midway, or at the end of the walkway. The perceived risk of tripping (PRT) was measured both before and after the participant walked through the walkway. A rating of gait disturbance (RGD) to each participant upon crossing the obstacle was also recorded.

RESULTS:

The PRT measured both before and after the walk were between “almost no” to “medium” risk levels. The RGD was affected significantly by the location of the light, obstacle height, obstacle depth, and number of obstacle.

CONCLUSION:

The location of light significantly affected the PRT both before and after the participants walked. The participants perceived a higher risk of tripping and had a relative high probability of foot-obstacle contact when the light was behind than when the light was in the front.

Keywords

Gait trip and fall obstacle crossing perceived risk of tripping

1 Introduction

Falls are common and are threats to the safety of workers worldwide [1 –3]. The Ministry of Labor [4] has reported more than two thousand fall incidences at work in 2016, which have accounted for 20.8% of all non-fatal occupational incidences, in Taiwan. However, the report did not specify whether those falls were slip- or trip-related. The statistics in the USA in the same year [5] indicated that there were 29,190 same level falls involving days away from work in private sectors due to tripping over an object. In addition, there were 6,090 same level falls due to tripping over uneven surfaces. Tripping over an object and over uneven surface incidences resulted in medians of 12 and 10 days away from work, respectively. In addition to these trip-induced same level falls, there were 16,393 trip-without-fall cases. These incidences resulted in a median of 10 days away from work. Furthermore, falls are common in the older population. Almost 2.8 million emergency treatments for seniors in hospitals in China were reported due to same-level falls [6] each year. The literature has shown that tripping has been found to be responsible for more than half of fall incidences among the older population [7 –10]. Investigating the mechanism and the factors affecting the onset of tripping is crucial for the safety of workers and for the general public.

Tripping occurs when the swinging foot of the walker is trapped by an obstacle or a bump on the ground, and is interrupted suddenly and unexpectedly. The minimum foot clearance (MFC) has been identified as one of the critical parameters affecting the occurrence of tripping [11, 12]. It is the minimum distance between the lowest point of the foot of the swing leg and the ground or the obstacle [10, 13]. A trip may occur when the MFC is reduced to zero, which implies a foot-floor or foot-obstacle contact. The MFC may occur at the toe, heel or the mid-sole [14]. The MFC occurs at the toe is also called the minimum toe clearance (MTC) [9, 15].

The MFC is affected by factors such as footwear, presence of obstacle on the floor, and walking speed. Davis et al. [16] has found that footwear has significant (p < 0.05) impact on the MFC of older women. Their participants walked with a greater MFC when wearing well-fitted footwear as compared with those wearing slippers and those walking with barefoot. Their results supported the hypothesis that wearing footwear with dorsal fixation may assist in reducing the likelihood of tripping. Schulz [9] found that the MTC was significantly affected by walking speed (p < 0.0001). Fast walking speed resulted in significantly (p < 0.001) higher MTCs than those of the preferred walking speed. The participants increased their MTCs by lifting their foot higher when they crossed the obstacle on the walkway. Presence of obstacle on the floor (p < 0.0006) was also found to be significant on the MTC. Presence of hidden obstacle on the floor resulted in significantly (p < 0.05) higher MTCs than those with visible obstacle on the floor. Presence of visible obstacle on the floor resulted in significantly (p < 0.05) higher MTCs than those without obstacle condition. Age was also found to be a significant factor affecting the MTC. Berard and Vallis [17] indicated that children of 7 years old had significantly larger MTC (12.05 cm) than their adult counterparts (10.28 cm) when crossing the obstacles of the same height.

Vision is a walker’s main sensory modality receiving the information of walkway conditions. It provides information of obstacles and unevenness on the floor and to guide locomotion to cross or to avoid the obstacles safely so as to avoid a trip and fall [18 –20]. Lee et al. [21] showed that continuous updating of the information of the distance to an obstacle by vision was essential for proper foot placement when approaching the obstacle to avoid a trip. Patla and Rietdyk [22] have shown that trajectory changes of the foot occurred upon crossing obstacles of different heights and widths.

Even though impaired visual acuity has been found to be a risk factor for falls for seniors, the literature has shown conflicting results with regard to whether impaired visual acuity increases the risk of falls [23]. Both the reduced contrast sensitivity and depth perception have been identified as critical visual risk factors [23, 24]. The depth perception could be damaged when the visual acuities of the two eyes are different. The literature [25] has shown that participants with good vision in both eyes had the lowest rate of falls, whereas those with good vision in one eye but poor vision in the other had elevated falling rates and had equivalent rate of fall to those with poor vision in both eyes.

Lighting affects vision and is obviously one of the important environmental factors affecting the risk of tripping. Haslam and Bently [26] indicated that poor lighting conditions had contributed to 20% of the slip, trip, and fall cases for mail delivery workers. The term “reassurance” has been adopted to describe the feeling of confidence a pedestrian might gain from lighting when walking [27]. Good lighting, i.e. high illuminance, is always recommended at night and in dark environments. Boyce et al. [28] recommended 10 lx as optimum illuminance at night while Fotios et al. [29] suggested a minimum horizontal illuminance of approximately 2.0 lx. It should be noted that the so-called “reassurance” was focused on perceived safety either in general or for “fear of crime”. The risk of tripping and falling has not been explicitly included in the above-mentioned literature.

Perception of risk of tripping (PRT) is the judgement of the likelihood of tripping by a walker. This judgement could affect the gait strategy that a walker adopts. A walker may adopt a conservative gait strategy by reducing the walking speed, stride length, and increasing the foot clearance [11, 25] when he or she feels the likelihood of been tripped is high. Past research studying the risk of tripping has been focused on the kinematics of foot [13 , 30–33]. Investigations studying the factors affecting the PRT especially under dimmed lighting conditions were uncommon.

Traditionally, lights are installed overhead in corridors, walkway, and tunnels. Installing lights intermittently along a long walkway normally results in non-uniform luminance conditions. A walker experiences the approaching, passing, and departing of a light when walking in a long corridor or a tunnel with overhead lighting. The location of a light with respect to the walker may affect the illuminance on the floor and hence affects the vision and the PRT when an obstacle is present. In the current study, human participants were tested under light bulb location conditions to determine the effects of the location of light on the PRT in obstacle crossing trials. In addition, we believe obstacle depth, obstacle height, and number of obstacle on the walkway under dimmed lighting conditions could also affect the PRT both before and after crossing obstacles on a walkway. These hypotheses were also tested and discussed.

2 Methods

The experiment was performed in the laboratory with the temperature at 25°C and relative humidity at 42% approximately.

2.1 Human participants

Twenty four healthy adults, including 12 males and 12 females, without self-reported musculoskeletal disorders and night blindness were recruited. Their age, stature, and body weight were 22.9 (±1.1) yrs, 171.1 (±7.9) cm, and 61.0 (±12.3) kg, respectively. All the participants have passed an Ishihara color test and have normal color vision. The visual acuities of the participants were examined using a Landolt’s E chart. The corrected decimal visual acuities of the left and right eyes of the participants were 0.8 (±0.3) and 0.8 (±0.3), respectively. The participants were categorized into a poor and a normal vision groups. The poor vision group included nine participants (5 males and 4 females) with one or both eyes with corrected visual acuities less than 0.8. The other participants were in the normal group. All participants read and signed an informed consent before the experiment began.

2.2 Walkway and obstacles

A black polyvinyl chloride (PVC) walkway (600×60 cm) with coefficient of friction of 0.34 (±0.09), measured by a pendulum friction tester [34], was prepared.

On the walkway, one, two, or three black obstacles with heights (H in Fig. 1) of 0, 5, or 15 cm and depth (or thickness, D in Fig. 1) of 1 or 10 cm were prepared. For the obstacles with height of 0 cm, black tapes were pasted on the walkway. These obstacles were essentially virtual and were served as a basis for comparisons. The participants were instructed to regard the tapes as obstacles and tried not to step on it. For the obstacles with depth of 1 cm and 10 cm, plywood and high density foam were used, respectively. All the obstacles had the same wide as the walkway and were placed at the locations shown in Fig. 1. All the plywood and foam obstacles were light and moveable when kicked or touched by the foot of the participants. Therefore, the participant would not trip even when the foot of the participant contacted with an obstacle in the trial.

Fig.1

Layout of experiment: (a) one obstacle; (b) two obstacles; (c) three obstacles.

Table 1

Illuminance (lx) on the walkway

Measurement location on the walkway
Location of light	Origin	Midway	End
Origin	0.237(±0.041)	0.028(±0.042)	0.006(±0.005)
Midway	0.028(±0.010)	0.267(±0.077)	0.031(±0.013)
End	0.009(±0.004)	0.050(±0.016)	0.284(±0.086)

2.3 Lighting conditions

The experiment was performed in the laboratory where all the windows were blocked from sunlight. Three light bulbs (3 watt) were suspended 2 m above the walkway (see Fig. 1). They were in the origin, middle, and end of the walkway. Only one of the light bulbs was turned on in each trial. This was to simulate the lighting conditions when a walker was walking toward, under, and walking away from the light. The luminance in the origin, mid-point, and the end on the floor of the walkway for the three locations of light, measured using a light meter (Prokit’s Industries Co., LTD, Shanghai, MT-4617LED-C), are shown in Table 1.

Table 2
Means and standard deviations of the PRT_before

Number of obstacles

Light location^* Obstacle height^ 1 2 3

1 cm 10 cm 1 cm 10 cm 1 cm 10 cm

Origin 0 cm 2.90(1.39) 2.46(1.14) 2.63(1.24) 2.44(1.15) 2.98(1.24) 2.71(1.08)

5 cm 2.94(1.32) 2.60(1.17) 3.00(1.41) 2.22(0.88) 2.75(1.26) 2.38(1.05)

15 cm 2.21(0.99) 2.27(0.97) 2.38(1.07) 2.4(0.87) 2.20(1.19) 2.44(1.01)

Midway 0 cm 1.89(1.22) 2.03(1.27) 2.47(1.50) 1.62(0.87) 2.10(1.27) 1.70(1.04)

5 cm 1.38(0.77) 1.27(0.44) 1.24(0.43) 1.29(0.46) 1.41(0.65) 1.37(0.57)

15 cm 1.37(0.70) 1.33(0.56) 1.59(0.78) 1.5(0.66) 1.59(0.72) 1.76(0.95)

End 0 cm 1.49(0.83) 1.33(0.75) 1.58(0.93) 1.24(0.52) 1.52(0.65) 1.40(0.65)

5 cm 1.18(0.48) 1.33(0.70) 1.17(0.38) 1.27(0.53) 1.24(0.41) 1.50(0.78)

15 cm 1.38(0.49) 1.63(0.88) 1.68(0.77) 1.75(0.90) 2.06(1.00) 2.13(1.12)

Number of obstacles
Origin	0 cm	2.90(1.39)	2.46(1.14)	2.63(1.24)	2.44(1.15)	2.98(1.24)	2.71(1.08)
5 cm	2.94(1.32)	2.60(1.17)	3.00(1.41)	2.22(0.88)	2.75(1.26)	2.38(1.05)
15 cm	2.21(0.99)	2.27(0.97)	2.38(1.07)	2.4(0.87)	2.20(1.19)	2.44(1.01)
Midway	0 cm	1.89(1.22)	2.03(1.27)	2.47(1.50)	1.62(0.87)	2.10(1.27)	1.70(1.04)
5 cm	1.38(0.77)	1.27(0.44)	1.24(0.43)	1.29(0.46)	1.41(0.65)	1.37(0.57)
15 cm	1.37(0.70)	1.33(0.56)	1.59(0.78)	1.5(0.66)	1.59(0.72)	1.76(0.95)
End	0 cm	1.49(0.83)	1.33(0.75)	1.58(0.93)	1.24(0.52)	1.52(0.65)	1.40(0.65)
5 cm	1.18(0.48)	1.33(0.70)	1.17(0.38)	1.27(0.53)	1.24(0.41)	1.50(0.78)
15 cm	1.38(0.49)	1.63(0.88)	1.68(0.77)	1.75(0.90)	2.06(1.00)	2.13(1.12)

^**p < 0.001; ^***p < 0.0001.

2.4 Experimental procedure

In the experiment, the participants wore their own sport shoes and were informed that they would walk on a 6 m walkway with the presence of one, two, or three obstacles under different lighting conditions. The order of the obstacle and lighting condition was randomly determined. The participants would not fall even if their foot runs into an obstacle. Prior to the start of the experiment, only one of the three randomly selected light bulbs was on. The participant waited until he or she could see the walkway environment without difficulties, approximately 10 minutes, and the trail started.

The participant stood at the origin of the walkway, observed and gave a rating of PRT in a five-point scale: 1. no risk, 2. low risk, 3. medium risk, 4. high risk, 5. extremely high risk. This rating was termed PRT_before. The participant then walked at 85 bpm following the sound of a metronome, crossed the obstacle, toward the end of the walkway, and then stopped. The participants were requested to avoid touching the obstacle including stepping on it. They, then, gave a second rating of PRT for the gait just completed using the same five-point scale as in the PRT_before. This rating was termed PRT_after. The PRT_adj was the difference between PRT_before and PRT_after, i.e. PRT_after minus PRT_before. A research assistant determined and recorded a rating of gait disturbance (RGD) for the participant on a five-point scale based on her observation on the participant’s gait while crossing the obstacle:

no disturbance: walked smoothly as if there was no obstacle;

slight disturbance: walked slowly down slightly when approaching the obstacle;

medium disturbance: walked slowly down more obviously when approaching the obstacle;

serious disturbance: stop in front of the obstacle briefly for gait adjustment; and

foot contact with the obstacle.

The time to complete the walk was recorded and the walking speed was calculated by dividing the length of the walkway by the time of the walk. After reporting the PRT_after, the participant walked back to the origin and stood facing away from the walkway. The research personnel prepared the obstacles for the next trial without the witness of the participant. After the obstacles were set up, the participant turned around and was on standby for the next trial.

Table 3
Means and standard deviations of the PRT_after

Number of obstacles^*

1 2 3

Light Obstacle Obstacle Depth^*

location^*** Height^*(cm) 1 cm 10 cm 1 cm 10 cm 1 cm 10 cm

Origin 0 cm 2.42(1.43) 2.13(1.26) 2.80(1.58) 2.10(1.22) 3.03(1.49) 1.82(0.92)

5 cm 2.75(1.42) 1.90(0.86) 2.70(1.33) 2.06(1.05) 2.46(1.41) 1.97(0.75)

15 cm 1.70(0.73) 1.52(0.62) 1.94(0.89) 2.05(0.83) 1.90(0.86) 2.11(0.89)

Midway 0 cm 1.70(1.08) 1.53(0.97) 2.28(1.46) 1.74(1.33) 2.46(1.61) 1.78(1.35)

5 cm 1.32(0.75) 1.13(0.31) 1.15(0.36) 1.22(0.51) 1.24(0.52) 1.28(0.54)

15 cm 1.25(0.51) 1.33(0.69) 1.50(0.78) 1.45(0.65) 1.58(0.71) 1.88(1.16)

End 0 cm 1.45(0.93) 1.24(0.84) 1.38(0.71) 1.24(0.66) 1.68(1.16) 1.40(0.92)

5 cm 1.05(0.20) 1.16(0.36) 1.09(0.28) 1.30(0.91) 1.43(1.10) 1.45(0.77)

15 cm 1.37(0.57) 1.45(0.65) 1.67(0.96) 1.65(0.76) 1.83(0.99) 2.11(1.30)

Number of obstacles^*
Origin	0 cm	2.42(1.43)	2.13(1.26)	2.80(1.58)	2.10(1.22)	3.03(1.49)	1.82(0.92)
5 cm	2.75(1.42)	1.90(0.86)	2.70(1.33)	2.06(1.05)	2.46(1.41)	1.97(0.75)
15 cm	1.70(0.73)	1.52(0.62)	1.94(0.89)	2.05(0.83)	1.90(0.86)	2.11(0.89)
Midway	0 cm	1.70(1.08)	1.53(0.97)	2.28(1.46)	1.74(1.33)	2.46(1.61)	1.78(1.35)
5 cm	1.32(0.75)	1.13(0.31)	1.15(0.36)	1.22(0.51)	1.24(0.52)	1.28(0.54)
15 cm	1.25(0.51)	1.33(0.69)	1.50(0.78)	1.45(0.65)	1.58(0.71)	1.88(1.16)
End	0 cm	1.45(0.93)	1.24(0.84)	1.38(0.71)	1.24(0.66)	1.68(1.16)	1.40(0.92)
5 cm	1.05(0.20)	1.16(0.36)	1.09(0.28)	1.30(0.91)	1.43(1.10)	1.45(0.77)
15 cm	1.37(0.57)	1.45(0.65)	1.67(0.96)	1.65(0.76)	1.83(0.99)	2.11(1.30)

^*p < 0.05; ^***p < 0.0001.

2.5 Experiment design and data analysis

An experiment with factorial block design was performed. Each participant was a block. The orders of obstacle height, obstacle depth, number of obstacle, and light bulb location combinations were randomly determined within each block. A total of 1,296 trails (24 participants ×3 obstacle heights ×2 obstacle depths ×3 numbers of obstacle ×3 light bulb locations) were performed. In addition to descriptive statistics, Wilcoxon rank-sum tests (also called Mann–Whitney U test) were performed to test the significances of obstacle depth on the PRT_before, PRT_after, and RGD. This test was also employed to test the difference between the poor and normal vision groups. Kruskal-Wallis tests were performed to test the significances of the number of obstacle, obstacle height, and location of light. The Spearman’s ρ was calculated for the correlation analyses. Statistical analyses were performed using the SAS^®9.4 (SAS Institute Inc., Cary, NC, USA) software.

3 Results

Tables 2 and 3 show the means and standard deviations of the PRT_before and PRT_after, respectively. The Kruskal-Wallis tests results indicated the significance of light location (χ₂² = 159.24, p < 0.0001), obstacle height (χ₂² = 14.79, p < 0.001) and insignificance (p > 0.05) of number of obstacle on the PRT_before. For light locations, the PRT_before at the origin, midway, and end of walkway were 2.40 (±1.17), 1.67 (±0.98), and 1.57 (±0.85), respectively. The PRT_before in the 0 cm, 5 cm, and 15 cm obstacle height conditions were 2.05 (±4.50), 1.75 (±01.04), and 1.85 (±0.97), respectively. The Wilcoxon rank-sum tests results indicated the insignificance (p > 0.05) of obstacle depth on the PRT_before. The PRT_before between the poor and normal vision groups were also insignificant.

Table 4
Means and standard deviations of the RGD

Number of obstacles^***

1 2 3

Light Obstacle Obstacle Depth^*

location^*** height^*(cm) 1 cm 10 cm 1 cm 10 cm 1 cm 10 cm

Origin 0 cm 2.54(1.41) 2.46(1.11) 3.89(1.43) 2.81(1.40) 4.25(1.29) 2.34(0.96)

5 cm 2.87(1.30) 2.25(0.75) 3.25(1.42) 2.75(0.98) 3.31(1.30) 2.68(0.79)

15 cm 2.67(0.69) 2.95(0.82) 3.35(0.94) 3.34(0.73) 3.36(0.85) 3.51(0.55)

Midway 0 cm 1.94(1.32) 1.63(1.16) 2.85(1.69) 2.37(1.46) 3.25(1.85) 1.96(1.37)

5 cm 1.65(0.79) 1.52(0.50) 1.63(0.59) 1.85(0.54) 1.83(0.64) 2.17(0.83)

15 cm 2.37(0.95) 2.73(0.98) 2.87(0.92) 2.83(1.03) 3.17(0.79) 3.51(0.79)

End 0 cm 1.58(1.01) 1.15(0.34) 2.08(1.50) 1.67(0.94) 2.10(1.27) 1.58(0.83)

5 cm 1.71(0.74) 1.77(0.59) 1.96(0.75) 1.98(0.81) 2.18(0.93) 2.23(0.83)

15 cm 2.44(0.88) 2.71(0.75) 3.20(0.79) 3.03(0.85) 3.37(0.69) 3.54(0.75)

Number of obstacles^***
Origin	0 cm	2.54(1.41)	2.46(1.11)	3.89(1.43)	2.81(1.40)	4.25(1.29)	2.34(0.96)
5 cm	2.87(1.30)	2.25(0.75)	3.25(1.42)	2.75(0.98)	3.31(1.30)	2.68(0.79)
15 cm	2.67(0.69)	2.95(0.82)	3.35(0.94)	3.34(0.73)	3.36(0.85)	3.51(0.55)
Midway	0 cm	1.94(1.32)	1.63(1.16)	2.85(1.69)	2.37(1.46)	3.25(1.85)	1.96(1.37)
5 cm	1.65(0.79)	1.52(0.50)	1.63(0.59)	1.85(0.54)	1.83(0.64)	2.17(0.83)
15 cm	2.37(0.95)	2.73(0.98)	2.87(0.92)	2.83(1.03)	3.17(0.79)	3.51(0.79)
End	0 cm	1.58(1.01)	1.15(0.34)	2.08(1.50)	1.67(0.94)	2.10(1.27)	1.58(0.83)
5 cm	1.71(0.74)	1.77(0.59)	1.96(0.75)	1.98(0.81)	2.18(0.93)	2.23(0.83)
15 cm	2.44(0.88)	2.71(0.75)	3.20(0.79)	3.03(0.85)	3.37(0.69)	3.54(0.75)

^*p < 0.01;^***p < 0.0001.

The Wilcoxon rank-sum tests results indicated the significance of obstacle depth (χ₁² = 5.09, p < 0.05) on the PRT_after. The PRT_after of the 1 cm depth condition (1.81±1.16) was significantly higher than that of the 10 cm condition (1.64±0.96). The Kruskal-Wallis tests results indicated the significance of light location (χ₂² = 99.61, p < 0.0001), obstacle height (χ₂² = 7.53, p < 0.05), and number of obstacle (χ₂² = 8.76, p < 0.05) on the PRT_after. For light locations, the PRT_after for the origin, middle, and end locations were 2.09 (±1.17), 1.57 (±1.01), and 1.50 (±0.90), respectively. The PRT_after for the 0 cm, 5 cm, and 15 cm obstacle height conditions were 1.90 (±1.27), 1.62 (±0.99), and 1.65 (±0.87), respectively. The PRT_after for the 1, 2, and 3 obstacles were 1.61 (±0.98), 1.74 (±1.07), and 1.82 (±1.13), respectively. The PRT_after between the poor (1.63±1.04) and normal (1.78±1.08) vision groups were significantly different (χ₁² = 9.78, p < 0.01). The Wilcoxon rank-sum tests results indicated the insignificance of obstacle depth on the PRT_adj. The Kruskal-Wallis tests results indicated the insignificance of light location, obstacle height, and number of obstacle on the PRT_adj. The PRT_adj between the poor (–0.17±1.01) and normal (–0.15±0.84) vision groups were statistical significant (χ₁² = 9.64, p < 0.01).

The means and standard deviations of the RGD under light location, obstacle height, number of obstacles, and obstacle depth conditions are shown in the Table 4. The Kruskal-Wallis tests results indicated the significance of location of light (χ₂² = 112.9, p < 0.0001), obstacle height (χ₂² = 16.76, p < 0.0001) and number of obstacle (χ₂² = 64.06, p < 0.0001) on the RGD. The RGD for the origin, middle, and end locations were 3.03 (±1.18), 2.30 (±1.22), and 2.23 (±1.09), respectively. The RGD for the 0 cm, 5 cm, and 15 cm obstacle height conditions were 2.36 (±1.49), 2.20 (±1.01), and 3.05 (±0.89), respectively. The RGD for the 1, 2, and 3 obstacles were 2.16 (±1.06), 2.65 (±1.25), and 2.80 (±1.24), respectively. The Wilcoxon rank-sum tests results indicated the significance of obstacle depth (χ₁² = 7.67, p < 0.01) on the RGD. The RGD of the 1 cm depth condition (2.68±1.34) was significantly higher than that of the 10 cm depth condition (2.46±1.18), respectively. The RGD between the poor (2.45±1.19) and normal (2.58±1.24) vision groups were not statistically significant.

A foot-obstacle contact implies tripping of the foot during the swing phase which could possibly result in a fall. Among the 1,296 trials, 113 trials ended up with one or more foot-obstacle contacts. The probability of foot-obstacle contact was 8.7% (113/1,296). The probabilities of foot-obstacle contacts for the 1 cm and 10 cm obstacle depths were 7.7% and 2.0%, respectively. The probabilities of foot-obstacle contact when there were two (3.6%) and three (3.9%) obstacles were higher than that when there was only one obstacle (1.2%). The probabilities of foot-obstacle contacts for the obstacle heights of 0, 5, and 15 cm were 5.9%, 2.0%, and 0.8%, respectively. The probabilities of foot-obstacle contact for the light location at the origin, midway, and end of walkway were 4.7%, 2.7%, 1.3%, respectively. The average number of foot-obstacle contact in one trial for each participant between the poor (4.44) and normal (4.47) vision group were not significant. In the others, a foot-obstacle case was counted so long as there were one or more contacts, and there were 17 trials with two contacts and 11 trials involved three contacts.

Table 5

Spearman’s correlation coefficients

PRT_before	PRT_after	RGD
PRT_before	1	0.712^*	0.348^*
PRT_after	1	0.458^*
PRT_adj	–0.462*	0.182^*	–
Walking speed	–0.342*	–0.332^*	–0.306^*

*p < 0.0001.

Table 5 shows the Spearman’s correlation coefficients for PRT_before, PRT_after, PRT_adj, RGD, and walking speed.

4 Discussion

When there are obstacles on the floor of the walkway, a walker looks and judges the risk of tripping. Depending on the dimensions of the obstacle he or she may choose to cross over, step on, or change path to avoid the obstacle. When environmental constraints and obstacle dimensions limit the choice to change path or to step on the obstacle, the walker needs information concerning the distance to the obstacle for proper foot placement and obstacle height for foot elevation control [19]. In our study, the participants were only allowed to cross over the obstacle. Neither stepping on the obstacle nor changing paths were allowed. Stepping on the obstacle would be counted as a foot-obstacle contact which we assumed would result in a trip. In addition, light and movable obstacles were adopted for safety reasons. These obstacles moved when a foot-obstacle contact occurred. A gait disturbance would occur upon foot-obstacle contact if the obstacle was not movable. This could lead to a trip and possibly a fall.

The PRT_before was the visual judgement of the risk of tripping when the obstacle was approximately 3 m in front of the participant. Such a judgement might be dependent on the capabilities of the participants to visually catch the details of the presence of the obstacle. The mean PRT_before across all experimental conditions ranged from 1.17 to 3.0. This implies that the participants perceived the walkway under various obstacle conditions with almost no to medium risk of tripping. The participants tended to give a relatively high PRT value when they could not see the obstacle(s) clearly. The illuminance (see Table 1) in the middle on the walkway, where the obstacle was present, was the lowest when the light was at the origin (0.028 lx) as compared with those in the other two locations (0.267 lx and 0.05 lx for the midway and end locations, respectively). Installing the light at the origin of the walkway was, therefore, the most disadvantageous for the participants to observe the obstacle(s) on the walkway and hence resulted in the highest PRT_before among the locations of the light. The participants gave the highest PRT_before for the 0 cm obstacle height conditions as compared to those of the 5 cm and 10 cm conditions. This was also because the taped-obstacle was almost invisible as compared to those of the other two obstacle-height conditions. When the participants were aware of the presence of obstacle but they could not see the details, they tended to give relative high PRT.

The PRT_after was the PRT based on the gait experience of crossing the obstacle(s) in a trial. In addition to vision, the PRT_after was believed to be affected by how difficult it was to pass the obstacles without touching them. The range of the PRT_after across all experimental conditions (1.05 to 3.03) was similar to that of the PRT_before. This implies that, in general, the experience of crossing the obstacle(s) did not change the PRT significantly. This was confirmed by the hypotheses testing results of the obstacle depth, obstacle height, number of obstacle, and light location on the PRT_adj (p > 0.05 for all four factors).

It was hypothesized that the 10 cm deep (or thick) obstacle was more difficult to pass and should be associated with higher PRT_after than the 1 cm ones. Our results were, however, contradictory to our expectation that the participants gave significantly higher ratings for the 1 cm deep obstacle(s) as compared to those of the 10 cm conditions. The explanation was that the participants might have more difficult in passing the 1 cm deep obstacle(s) than those of the 10 cm ones because they had more difficult in catching the spatial characteristics (such as the distance and height information) of the former ones visually than the latter ones. In other words, the participants could see the 10 cm deep obstacle(s) better and hence had less difficult to passing them. This was consistent with the results of the RGD that the RDG of the 1 cm depth condition (2.68) was significantly (p < 0.01) higher than that of the 10 cm conditions (2.46). This was also consistent with the results of the probabilities of foot-obstacle contact that the participant had 3.85 times (7.7% versus 2%) higher probability of foot-obstacle contact in the 1 cm condition than in that of the 10 cm condition.

It was anticipated that increasing the number of obstacle resulted in higher PRT_after value as this could make it more difficult for the participants to cross the multiple obstacles. This was confirmed by the significance of number of obstacle on the PRT_after. This was also consistent with the results of the RGD that increasing the number of obstacle resulted in higher RGD values. The same results were also consistent with those of the probability of foot-obstacle contact that the probabilities in the two (3.6%) and three (3.9%) obstacles conditions were 3 and 3.25 times higher than that of the one obstacle condition (1.2%), respectively.

The RGD was the rating of gait disturbance of the participant because of obstacle crossing. The mean RGD across all the experimental conditions ranged 1.15 to 4.25, which corresponds to almost no to more than serious disturbance. Location of light affected the RGD significantly (p < 0.0001). The mean RGD was the highest when the light location was at the origin (3.03) of the walkway, followed by that at the midway (2.3), and finally at the end (2.23). When the light was at the origin, the light was block by the participant when he or she was approaching the obstacle(s), which made it difficult for the participant to visually catch the details of the obstacle(s) and hence resulted in more gait disturbance and higher probability of foot-obstacle contact. The light might also be blocked partially when the light was overhead of the participant when he or she was crossing the obstacle(s). Therefore, the gait disturbance was the most serious when the light was on the back and was the least when the light was in the front at the time of obstacle crossing.

For obstacle height, the 15 cm high obstacle conditions resulted in the highest RGD (3.05). This was consistent with our anticipation that high obstacle required more leg raise to cross and hence were associated with more gait disturbance. The RGD of the 0 cm high obstacle conditions ranked second (2.36). This might be attributed to the phenomenon that the participants needed to be more cautious not to step on the virtual obstacle(s) due to the less visibility of the 0 cm obstacle(s). The 5 cm obstacle height conditions had the lowest RGD (2.20) among all obstacle height conditions.

Vision is believed to have an influence on the risk of trip and fall [23, 25]; however, the effects of visual acuity on the PRT have not been discussed in the literature. Our results indicated that both the PRT_before and RGD were not significantly affected by the vision group, but the PRT_after between the two vision groups were significantly different. This implies that vision did not affect the participants’ rating of the PRT when they were standing 3 m in front of the obstacle. The participants updated their PRT after they have crossed the obstacle based on both their experiences of obstacle crossing and their follow-up observations of the obstacle details on the ground. The participants in the normal vision group could probably see more clearly on the details of the obstacle on the ground. That was why they gave significantly higher PRT_after than their poor vision counterparts.

There were 113 foot-obstacle contacts, and more than half of them (61 or 54%) occurred when the light was at the origin, which was the location at the back of the participant when approaching the obstacle. The counts (probabilities) of the trials with one or more foot-obstacle contacts when the light was at the midway and the end of the walkway were 36 (31.9%) and 17 (15%), respectively. For the 28 trials ended up with two or three foot-obstacle contacts, 19 (67.9%) of them occurred when the light was at the origin. Only one of them (3.6%) occurred when the light was at the end of the walkway, i.e. in front of the participant. This implies that the participants had relative high risk of tripping when the light was behind in a dimmed environment, and relatively low risk when the light was in the front. For all the foot-obstacle contacts, 77% (87) of them occurred when the obstacle(s) had a depth of 1 cm and 23% (26) of them had a depth of 10 cm. This might because the participants were more cautious when crossing the 10 cm thick obstacle(s) as these obstacles were larger and hence provided more spatial and location information than those of the 1 cm ones. For the height of the obstacles, 68.1% (77) of the foot-obstacle contacts occurred when there were actual no obstacle (height of 0 cm), 23% (26) and 8.9% (10) of them occurred when the obstacle height was 5 cm and 10 cm, respectively.

5 Conclusion

Under dimmed lighting conditions, the participants look at the walkway with obstacles in the walkway and perceived almost no to medium risk of tripping. After they walked through the walkway, they did not change their PRT significantly as compared to their perception before they walked. The levels of gait disturbance of the participants for obstacle crossing under the experimental conditions were between almost no to more than serious disturbance. Gait disturbance was negatively correlated with both PRT_before and PRT_after. The location of a light significantly affected the PRT both before and after a walk. The participants perceived higher risk of tripping and had relative high probability of foot-obstacle contact when the light was behind them than when the light was in front of them. Installing spots of light evenly along long walkways and tunnels, where road surface unevenness and presence of obstacle are likely, is not recommended. Continuously lighting, which allows easy access of road surface and obstacle information, is amore preferable way of lighting design.

Conflict of interest

None to report.

Footnotes

Acknowledgments

This research was partially supported the Priority Academic Program Development of Jiangsu Higher Education Institutions, and a funding from the Fundamental Research Funds for the Central Universities under Grant 2017XKQY045.

References

, Zhao

, Peng

, et al. Subjective assessments of floor slipperiness before and after walk under two lighting conditions. Int J Occup Saf Ergon. 2018;24(2):294–302.

, Chen

C-Y

, Chen

, Liu

. Assessment of slip resistance under footwear materials, tread designs, floor contamination, and floor inclination conditions. Work. 2012;41:3349–51.

, Li

. Perceived Floor Slipperiness and Floor Roughness in a Gait Experiment. Work. 2015;50:649–57.

Ministry of Labor Affairs (MAL). 2017. Annual Labor Inspection Report, Taipei, Taiwan. In Chinese.

Bureau of Labor Statistics (BLA). 2017. Case and Demographic Characteristics for Work-related Injuries and Illnesses Involving Days Away From Work: R31. Detailed event or exposure by selected natures. https://www.bls.gov/iif/oshcdnewhtm. Accessed May 22, 2018

Chai

. How can the elderly avoid falling? Yellow River Yellow Earth Yellow Race. 2018;(11):46.

Berg

, Alessio

, Mills

, et al. Circumstances and consequences of falls in independent community dwelling older adults. Age and Ageing. 1997;26:261–8.

Blake

, Morgan

, Bendall

, et al. Falls by elderly people at home: prevalence and associated factors. Age and Ageing. 1988;17(6):365–72.

Schulz

. Minimum toe clearance adaptations to floor surface irregularity and gait speed. J Biomech. 2011;44:1277–84.

10.

Lai

, Taylor

, Begg

. Prediction of foot clearance parameters as a precursor to forecasting the risk of tripping and falling. Human Movement Science. 2012;31:271–83.

11.

Chen

H-C

, Ashton-Miller

, Alexander

, et al. Stepping Over Obstacles: Gait Patterns of Healthy Young and Old Adults. Journal of Gerontology: Medical Sciences. 1991;46(6):196–203.

12.

Barrett

, Mills

, Begg

. A systematic review of the effect of ageing and falls history on minimum foot clearance characteristics during level walking. Gait Posture. 2010;32(4):429–35.

13.

Loverro

, Mueske

, Hamel

. Location of minimum foot clearance on the shoe and with respect to the obstacle changes with locomotor task. J Biomech. 2013;46(11):1842.

14.

Kunzlera

, da Rochaa

, dos Santosa

, Ceccona

, Priarioa

, Carpes

. Should we consider steps with variable height for a safer stair negotiation in older adults? Work 2018;59:15–21.

15.

Mills

, Barrett

, Morrison

. Toe clearance variability during walking in young and elderly men. Gait Posture. 2008;28(1):101–7.

16.

Davis

, Galna

, Murphy

, et al. Effect of footwear on minimum foot clearance, heel slippage and spatiotemporal measures of gait in older women. Gait Posture. 2016;44:43–7.

17.

Berard

, Vallis

. Characteristics of single and double obstacle avoidance strategies: a comparison between adults and children. Exp Brain Res. 2006;175:21–31.

18.

Patla

, Prentice

, Robinson

, et al. Visual control of locomotion: strategies for changing direction and for going over obstacles. Journal of Experimental Psychology: Human Perception and Performance. 1991;17(3):603–34.

19.

Patla

, Tomescu

, Ishac

MGA

. What visual information is used for navigation around obstacles in a cluttered environment? Can J Physiol Pharmacol 2004;82:682–92.

20.

Patla

, Greig

. Any way you look at it, successful obstacle negotiation needs visually guided on-line foot placement regulation during the approach phase. Neuroscience Letters. 2006;397:110–4.

21.

Lee

, Lishman

, Thomson

. Regulation of gait in long jumping. J. ExPsychol. Hum. Percept. Perform. 1982;8:448–59.

22.

Patla

, Rietdyk

. Visual control of limb trajectory over obstacles during locomotion: effect of obstacle height and width. Gait Posture. 1993;1:45–60.

23.

Lord

. Visual risk factors for falls in older people. Age and Ageing. 2006;35-S2:42–5.

24.

Felson

, Anderson

, Annan

. Impaired vision and hip fracture. The Framingham Study. Journal of American Geriatrics Society. 1989;37:495–500.

25.

Heasley

, Buckley

, Scally

, et al. Stepping up to a New level: effects of blurring vision in the elderly. Investigative Ophthalmology & Visual Science. 2004;45(7):2122–8.

26.

Haslam

, Bentley

. Follow-up investigations of slip, trip and fall accidents among postal delivery workers. Saf Sci. 1999;32:3–47.

27.

Knight

Field surveys investigating the effect of lamp spectrum on the perception of safety and comfort at night. Lighting Res Technol. 2010;42(3):313–30.

28.

Boyce

, Gutkowski

The if, why and what of street lighting and crime: A review. Lighting Res Technol. 1995;27:103–12.

29.

Fotios

, Liachenko

, Uttley

Evaluation of pedestrian reassurance gained by higher illuminances in residential streets using the day–dark approach. Lighting Res Technol. 2018;0:1–19.

30.

Begg

, Best

, Dell’Oro

, et al. Minimum foot clearance during walking: Strategies for the minimisation of trip-related falls. Gait Posture. 2007;25(2):191–8.

31.

, Chen

. Comparisons of the lower limb kinematics between young and older adults when crossing obstacles of different heights. Gait Posture. 2006;23(4):471–9.

32.

Paterson

, Hill

, Lythgo

. Stride dynamics, gait variability and prospective falls risk in active community dwelling older women. Gait Posture. 2010;33(2):251–5.

33.

Draganich

, Kuo

. The effects of walking speed on obstacle crossing in healthy young and healthy older adults. J Biomech. 2004;37(6):889–96.

34.

ASTM E303–93. 2018. Standard test method for measuring surface frictional properties using the British Pendulum Tester. Annual Book of ASTM Standards. vol. 04.03. West Conshohochen, PA, American Society forTesting and Materials.

Number of obstacles
Light location^***	Obstacle height^**	1		2		3
1 cm	10 cm	1 cm	10 cm	1 cm	10 cm
Origin	0 cm	2.90(1.39)	2.46(1.14)	2.63(1.24)	2.44(1.15)	2.98(1.24)	2.71(1.08)
5 cm	2.94(1.32)	2.60(1.17)	3.00(1.41)	2.22(0.88)	2.75(1.26)	2.38(1.05)
15 cm	2.21(0.99)	2.27(0.97)	2.38(1.07)	2.4(0.87)	2.20(1.19)	2.44(1.01)
Midway	0 cm	1.89(1.22)	2.03(1.27)	2.47(1.50)	1.62(0.87)	2.10(1.27)	1.70(1.04)
5 cm	1.38(0.77)	1.27(0.44)	1.24(0.43)	1.29(0.46)	1.41(0.65)	1.37(0.57)
15 cm	1.37(0.70)	1.33(0.56)	1.59(0.78)	1.5(0.66)	1.59(0.72)	1.76(0.95)
End	0 cm	1.49(0.83)	1.33(0.75)	1.58(0.93)	1.24(0.52)	1.52(0.65)	1.40(0.65)
5 cm	1.18(0.48)	1.33(0.70)	1.17(0.38)	1.27(0.53)	1.24(0.41)	1.50(0.78)
15 cm	1.38(0.49)	1.63(0.88)	1.68(0.77)	1.75(0.90)	2.06(1.00)	2.13(1.12)

Number of obstacles^*
1		2		3
Light	Obstacle	Obstacle	Depth^*
location^***	Height^*(cm)	1 cm	10 cm	1 cm	10 cm	1 cm	10 cm
Origin	0 cm	2.42(1.43)	2.13(1.26)	2.80(1.58)	2.10(1.22)	3.03(1.49)	1.82(0.92)
5 cm	2.75(1.42)	1.90(0.86)	2.70(1.33)	2.06(1.05)	2.46(1.41)	1.97(0.75)
15 cm	1.70(0.73)	1.52(0.62)	1.94(0.89)	2.05(0.83)	1.90(0.86)	2.11(0.89)
Midway	0 cm	1.70(1.08)	1.53(0.97)	2.28(1.46)	1.74(1.33)	2.46(1.61)	1.78(1.35)
5 cm	1.32(0.75)	1.13(0.31)	1.15(0.36)	1.22(0.51)	1.24(0.52)	1.28(0.54)
15 cm	1.25(0.51)	1.33(0.69)	1.50(0.78)	1.45(0.65)	1.58(0.71)	1.88(1.16)
End	0 cm	1.45(0.93)	1.24(0.84)	1.38(0.71)	1.24(0.66)	1.68(1.16)	1.40(0.92)
5 cm	1.05(0.20)	1.16(0.36)	1.09(0.28)	1.30(0.91)	1.43(1.10)	1.45(0.77)
15 cm	1.37(0.57)	1.45(0.65)	1.67(0.96)	1.65(0.76)	1.83(0.99)	2.11(1.30)