Journal Logo

Applied Sciences: Biodynamics

Lower-Extremity Gait Kinematics on Slippery Surfaces in Construction Worksites

FONG, DANIEL TIK-PUI1; HONG, YOULIAN2; LI, JING XIAN3

Author Information
Medicine & Science in Sports & Exercise: March 2005 - Volume 37 - Issue 3 - p 447-454
doi: 10.1249/01.MSS.0000155390.41572.DE
  • Free

Abstract

Slips and falls are among the most serious causes of morbidity and mortality (7). In the United States, slips and falls are associated with disability, fractures, and deaths in occupational areas (16). In the United Kingdom, about 20% of the occupational injuries are reportedly due to slips and falls each year (7). In Hong Kong, data from local hospitals in 1999 showed that the industrial section ranked top (30.1%) in the causes of traumatic injuries, and accidental falls were the main cause (41.1%) of hospitalized injuries (11). In 2000, slips and falls were the most frequent cause of occupational accidents, contributing to 25% of the total cases (12).

Slips and falls are involved by complex extrinsic environmental factors and intrinsic human factors (4). In a normal nonslippery environment, the extrinsic and intrinsic factors are in balance, resulting in an average low slipping potential. When the extrinsic environmental factors become more likely to introduce a slip, humans could modify the intrinsic human factors to restore a low slipping potential and finally reduce the overall slipping likeliness. The cumulative effects of the risk factors mentioned above can be illuminated by kinetic and kinematic measurement. In kinetics, the dynamic coefficient of friction (DCOF) was commonly investigated because the heel horizontal velocity was not zero at the moment of heel strike (5) and thus the DCOF instead of the static coefficient of friction (SCOF) was believed to be more relevant to slip events (15). Various mechanical slip-resistance tests were conducted to investigate the slipperiness of walking surfaces by analyzing the dynamic coefficient of friction between combinations of footwear and walking surfaces (5,9,13).

In kinematics, most of the previous studies investigated the changes of gait parameters during the heel contact phase. Increased step length would result in a greater shear force, which in turn increased the slippery likeliness (3). In adapting to slippery walking surfaces, people of all ages shorten step length to reduce the likelihood of slipping (5,7). During walking, heel horizontal velocity rises gradually after take-off of the foot, reaches a maximum during the swing phase, and falls to zero rapidly after heel contact to support the stance leg (6). Failure to achieve zero horizontal velocity at heel contact may result in a slip. However, in slowing down the walking speed, the heel horizontal velocity is not decreased as expected (17). Therefore, heel velocity should be reported in gait analysis. In lower-extremity kinematics, the overall profiles of ankle and knee joint angles were in agreement across the past studies (1,18). At the ankle, there is a dorsiflexion at heel contact, followed by a rapid plantar flexion. At the knee, there is a flexion during the first 30% of the stance and another flexion again during the last phase of the stance, followed by take-off of the foot (14).

Injuries due to slips and falls are not purely random events but rather predictable with known risk factors (4). A local survey (10) reported that 48.7% of the construction workers believed that these injuries could be prevented by working with proper safety equipment, policies, and measures. However, about 34% of the workers found difficulties in learning about safety measures to prevent occupational injury. The purpose of this study was to investigate the lower-extremity preventive measures to slips when walking over potentially slippery walking surfaces in simulated construction worksite environments. The findings from this study would help in discovering the risk factors, understanding the human adaptation to slippery surfaces, and educating construction worksite workers to use safe walking strategies when walking on slippery surfaces to prevent occupational slips and falls.

METHODS

Survey.

Two thousand questionnaires were randomly sent to local construction site workers. The survey aimed to get statistics about the popular footwear used by workers, the nature of the walking surface at the construction site, and the most common types of contaminants on the walking surface, to better simulate the real situation in local construction worksites. The selection criteria of footwear, flooring, and contaminants were the items with top ranking and comparable popularity. From the results of the survey, two types of footwear, two types of flooring surface, and four types of contaminants were chosen. The most popular type of footwear (93.9%) was a kind of safety shoe that passed European Safety standard EN 345 and is currently recommended by the Hong Kong Occupational Safety and Health Council. The second most popular type of footwear was a cloth shoe, which was a kind of lightweight, low-price, traditional sport shoe in Hong Kong and mainland China. Even the second most popular shoe, the cloth sport shoe (2.0%), was far less popular than the safety shoe, and it was also chosen for comparison. The two most popular types of flooring surfaces included cement plates (57.3%) and wooden plates (33.3%), which were chosen because they had comparable popularity. The four selected contaminant conditions included dry, sand (43.4%), water (38.7%), and oil (33.3%). The selected footwear, flooring, and contaminants from the results of the survey made a total of 16 simulated construction worksite environments to be investigated.

Mechanical slip-resistance test.

The DCOF of all 16 footwear/flooring/contaminant conditions were measured for slip-resistance classification. A self-designed pulley system (Fig. 1), which allowed an adjustable horizontal drag force, was used to drag a 11.8-kg–weighted shoe over a testing flooring surface 10 times over a force plate (Kistler 9281CA, Switzerland) (9). Weights were added in the pulley system to increase the horizontal drag gradually until the shoe slid. The sliding velocity, and horizontal and vertical reaction forces during the slide were recorded by the force plate. The DCOF was calculated by dividing horizontal reaction force by vertical reaction force. According to the measured DCOF and the classification scale suggested by Grönqvist et al. (5), the 16 conditions were classified into three groups (very slip resistant, unsure, and slippery) as shown in Table 1. The effect of slipping potential on the lower-extremity kinematic parameters was investigated in a later human walking test.

FIGURE 1— Pulley system in the mechanical slip resistance test.
FIGURE 1— Pulley system in the mechanical slip resistance test.
TABLE 1
TABLE 1:
Sliding speed, DCOF, and slip resistance classification of the 16 simulated environments in mechanical test.

Subjects.

Fifteen Chinese males (age = 21.8 ± 1.3 yr, mass = 64.5 ± 4.6 kg, height = 1.75 ± 0.06 m) with no gait abnormalities and with right-leg dominance were recruited for this study. Written informed consent was obtained from all subjects before the study. The university ethics committee approved the study.

Instrumentation.

A harness system was installed by attaching a harness (Protecta International AB103, U.S.) that conformed to European safety standard EN 361 to a horizontal stainless steel wire by an adjustable connection lanyard (Protecta International AL110C) and a steel safety hook (Protecta International AJ501) that conformed to European safety standards EN354 and EN362, respectively. The horizontal stainless steel wire was 32 feet in length and firmly attached to a wall 2.4 m from the ground at both ends. A pair of safety shoes of size 42 (length = 265 mm), which conformed to European safety standard EN 345, were purchased from a local distributor recommended by the Hong Kong Occupational Safety and Health Council. The sole of the safety shoe fully complied with the main regulations provided by the EEC/89/686 European Directive with harmlessness, comfort, solidity, and protection against skidding risks (UNI 8615/1–DIN 4843). A cloth sport shoe of the same size was purchased from sport equipment shops. The cloth sport shoe was made with a thin layer of cloth shoe last, and a thin and flexible rubber sole. A 5-m walking path was prepared by connecting several cement or wooden flooring plates provided by the university construction work unit. The amounts of the contaminants were about 1 L·m−2 for sand, and 0.5 L·m−2 for water and oil, as they could form a thin layer on the flooring surface without spilling off the surface. The oily condition was prepared with motor oil (elf 10W40 motor oil), which is often used in engines and machines in construction sites (1,6).

Procedure.

Subjects were requested to dress in black and tight clothing that, together with illuminated silvery reflective skin markers, facilitated the autodigitizing process in video data analysis. The reflective skin markers were attached at the major lower-extremity anatomical landmarks on the right side, including the greater trochanter, lateral femoral condyle, lateral malleolus, fifth metatarsal head, and talus (Fig. 2). Ankle and knee joint angles were defined as the included angles (Fig. 2). The harness system was adjusted for each subject so that it would not affect the subject’s perceived normal gait, and it could support and protect the subject in case of a fall (Fig. 3). For both cement and wooden walking surfaces, each subject performed 10 trials of walking on each footwear/flooring/contaminant condition in the sequence of dry, sand, water, and oil. The sequence was designed to avoid the gait alternation effect when walking on a dry surface after a slippery surface as suggested by a previous study (2). During each trial, subjects were instructed to walk at a self-paced normal speed and avoid slipping.

FIGURE 2— Marker positions and angle definitions (1, greater trochanter; 2, lateral femoral condyle; 3, lateral malleolus; 4, fifth metatarsal head; 5, talus).
FIGURE 2— Marker positions and angle definitions (1, greater trochanter; 2, lateral femoral condyle; 3, lateral malleolus; 4, fifth metatarsal head; 5, talus).
FIGURE 3— Subject trying the harness to make sure it can support him in case of a fall.
FIGURE 3— Subject trying the harness to make sure it can support him in case of a fall.

One CCD digital video camera (JVC 9600, Japan) with a 50-Hz filming rate at a 1/250-s shutter speed was used for videotaping human motion in the sagittal plane. The filmed data were processed by a motion analysis system (Ariel Performance Analysis System, U.S.) to obtain two-dimensional coordinates and their derivatives of digitized anatomical markers. Trials with slips were discarded. A slip was defined as the subject requiring support from the harness as reported by the subject, or when the heel horizontal velocity failed to achieve zero within a 3-cm displacement range (8) immediately after the foot strike (2), which was checked by motion analysis.

Data analysis.

Data of the successful trials of walking without slips were averaged for each footwear/flooring/contaminant condition. Gait-pattern parameters including stance, swing and stride time, stride length, heel horizontal and vertical velocity and acceleration at foot strike, and mean propagation speed were obtained from the motion analysis system. Mean propagation speed was measured by the average value of horizontal forward linear velocity of the hip during the stance period. Lower-extremity kinematic data including angular displacement and velocity of ankle and knee joint, and the foot–floor angle were extracted. The profiles of these data were time normalized from foot strike (0% stance) to take-off (100% stance) and were evaluated from foot strike (0% stance) until midstance (50% stance) with 10% stance intervals in between. One-way MANOVA with repeated measures was employed to examine the difference in gait pattern parameters and lower-extremity kinematic data between the classified slip-resistance groups. One-way ANOVA was employed to examine the difference in each gait pattern parameter and in each lower-extremity kinematic datum at selected time points between the groups. Significance level was set at P < 0.05 level. Tukey post hoc pairwise comparisons were conducted between each pair of groups when significant differences reached the P < 0.01 significance level.

RESULTS

Gait pattern.

MANOVA showed that gait pattern was significantly affected by the slipperiness of the walking surface (P < 0.05). The descriptive statistics and the results of ANOVA and Tukey tests are shown in Table 2. Results showed that when the walking surface slipperiness increased from “very slip resistant” to “unsure” and “slippery,” the stance time and stride time significantly increased by about 0.13 s (16%) and 0.14 s (12%), respectively (P < 0.01). Stride length and mean propagation speed significantly decreased from 1.22 to 1.06 m and from 1.01 to 0.80 m·s−1, respectively (P < 0.01). Heel horizontal velocity and vertical acceleration showed a significant decrease in magnitude in the slippery condition (P < 0.01). Heel horizontal acceleration showed a significant decrease in magnitude in the slippery condition at the P < 0.05 level. No significant difference was found among groups in heel vertical velocity at the P < 0.05 level.

TABLE 2
TABLE 2:
Descriptive statistics, results of ANOVA and Tukey tests of gait pattern parameters.

Ankle joint kinematics.

MANOVA showed significant differences among different classes in ankle joint kinematics (P < 0.01). The descriptive statistics and the results of ANOVA and Tukey tests are shown in Table 3. The profile of the ankle angle and angular velocity from foot strike (0% stance) to midstance (50% stance) for the three classes are shown in Figure 4. Similar dorsiflexion trends were found from foot strike to midstance in all three groups. The range of angle changes for the three groups were similar, about 20° from foot strike to midstance. Generally, the included ankle angle in the “unsure” group was significantly larger than the other two groups at all selected time points (P < 0.05). Comparing the trends of the “slip-resistant” and “slippery” groups, the ankle joint in the “slippery” group was more plantar flexed from foot strike to 15% stance and was more dorsiflexed from 15% to midstance. However, no significant differences were found at any time points. The ankle joint angular velocities were all negative from foot strike to midstance, indicating that dorsiflexion occurred all the time in this period. The variation of angular velocity dropped with increasing slipping potential. The range was about 60·s−1 for the slip-resistant group, and was about 30·s−1 for the slippery group. A Tukey test showed significant differences (P < 0.01) between the “very slip-resistant” and “slippery” groups at 10%, 20%, and 40% stance.

TABLE 3
TABLE 3:
Descriptive statistics, and results of ANOVA and Tukey tests of ankle joint kinematics parameters.
FIGURE 4— Included ankle joint profile from foot strike (0% stance) to midstance (50% stance).
FIGURE 4— Included ankle joint profile from foot strike (0% stance) to midstance (50% stance).

Knee joint kinematics.

MANOVA showed significant differences among different classes on knee joint kinematics (P < 0.01). The descriptive statistics and the results of ANOVA and Tukey tests are shown in Table 4. The profile of the knee angle and angular velocity from foot strike (0% stance) to midstance (50% stance) for the three classes are shown in Figure 5. Knee extension occurred during the first 5% stance, followed by rapid knee flexion until midstance. The trends of knee angle and angular velocity of the three groups were similar. No significant differences were found in the knee angle at each time point between three groups. For knee angular velocity, significant differences were found between the “very slip-resistant” and “slippery” groups from 40 to 50% stance (P < 0.05).

TABLE 4
TABLE 4:
Descriptive statistics, results of ANOVA and Tukey tests of knee joint kinematics parameters.
FIGURE 5— Included knee joint profile from foot strike (0% stance) to midstance (50% stance).
FIGURE 5— Included knee joint profile from foot strike (0% stance) to midstance (50% stance).

Foot–floor angle.

MANOVA showed significant main differences (P < 0.01) on overall foot–floor angle parameters between the three slip-resistant groups. The descriptive statistics and the results of ANOVA and Tukey tests are shown in Table 5. The profile of foot–floor angle from foot strike (0% stance) to midstance (50% stance) for the three classes is shown in Figure 6. One-way ANOVA showed significant differences at all selected time points (P < 0.05). Tukey pairwise comparisons showed significant difference between resistant–unsure conditions at foot strike, between 40% and 50% stance (P < 0.05), between resistant–slippery conditions at 30% and 40% stance (P < 0.01), and between unsure–slippery conditions from 20% to 50% stance (P < 0.01).

TABLE 5
TABLE 5:
Descriptive statistics, and results of ANOVA and Tukey tests of foot-floor angle.
FIGURE 6— Foot–floor angle profile from foot strike (0% stance) to midstance (50% stance).
FIGURE 6— Foot–floor angle profile from foot strike (0% stance) to midstance (50% stance).

DISCUSSION

The mechanical slip-resistance test provided a glance at slipping risk. Based on the dynamic data published on human skidding during normal gait, a value of 0.20 was suggested to be a safe limit for slip resistance (15). Two of the sixteen tested construction worksite environments were evaluated as having a slipping hazard, including wearing either a safety shoe or cloth shoe on a wooden surface with the presence of oil contaminant. Wooden surfaces are often present in construction worksites because workers place wooden floorings on top of the finished flooring to protect it against damage and contamination during construction work. Oil contaminants are often present because the workers need lubricant oil for their machines. In the presence of both wooden flooring and oil contaminant, a slipping hazard can arise, even if the workers wear safety shoes as recommended. Therefore, workers should be more careful when walking on wooden surfaces, and should at the same time avoid leakage of machine lubricants.

In this study, the mechanical slip-resistance test was not truly realistic, because no heel–sole contact was simulated as in previous studies, including the programmable slip-resistance test (13) and Grönqvist et al.’s movable artificial foot (5). However, a similar simple mechanical drag test as an alternative low-cost measure was also published (9). The main purpose of this mechanical test was to provide a method to reduce the data groups for later comparison of kinematic parameters in human walking. It was not the focus of this study, and therefore a low-cost protocol that saved time and money was employed. From the mechanical slip-resistance test, 2 of 16 conditions were identified to have a slipping risk. This made the number of trials for slippery and nonslippery groups unbalanced. However, in the simulated environment, most of the conditions were highly slip resistant, and it was a limitation to have a comparable amount of trials for different groups for comparison.

Another limitation was due to the experimental safety measure. In the human walking test, walking with a harness was unrealistic but necessary during the experiment to support the subject in case of a real fall. The harness may provide support for the subject and may alter their normal gait. However, the effect of wearing the harness could not be demonstrated, as no trials were performed without a harness. To minimize this effect, the harness was adjusted every time for each subject so that it would not affect the subject’s perceived normal gait. Moreover, the harness may introduce a psychological effect to subjects, as they know that they will be supported and will not hit the ground in a real slip.

Stance time and stride time significantly increased in unsure and slippery walkway conditions, from 0.79 to 0.92 s and from 1.21 to 1.35 s, respectively (P < 0.01). Moreover, with increasing slipping potential, stride length decreased significantly from 1.22 to 1.06 m (P < 0.01). In shortening the stride length, the foot could be maintained near the body, thus increasing body stability as the line of gravity of the body is closer to the base of support during the foot swing. This finding was in agreement with previous published studies (5). In this study, in shortening the stride length, heel horizontal velocity at foot strike was not significantly decreased as expected. This finding is comparable to Winter et al.’s study (17), which found that the heel horizontal velocity of older adults walking slower was significantly higher than that of younger adults walking faster. A significant decrease in mean propagation speed was found (P < 0.01). With increasing slipping risk, the speed decreased from 1.01 to 0.80 m·s−1. With increasing slipping hazards, the heel horizontal velocity, horizontal acceleration, and vertical acceleration dropped significantly (P < 0.01). This indicated a more gentle foot strike to prevent a slip. These changes indicated that subjects had employed an active strategy to adapt to a slippery walking surface to avoid slips.

The profile of ankle joint parameters suggested that dorsiflexion occurred all the time from foot strike to midstance in all groups. This finding was not in total agreement with the summary of previous studies of gait kinematics without slipping, which stated that the ankle joint was in slight dorsiflexion at contact, followed by a rapid peak plantar flexion at around 10% of stance as the foot rotated down onto the floor (14). However, foot–floor angle data suggested that the heel strikes the ground with an angle and is rotated down onto the floor flat at about 10% of stance, which was in agreement with previous published summarized data (14). The profile of ankle angle in the unsure group was found to be significantly higher. It might be due to the uncertainty of the floor slipperiness. The foot–floor angle at heel strike in the unsure group was significantly smaller. It indicated that subjects tended to land on the floor with a more flat foot and plantarflexed ankle joint during the first 10% stance. Flatfoot landing may help to achieve a reaction force in normal direction instead of in shear direction by flatfoot landing, as the shear force plays important role to initiate slipping. In the very slip-resistant group, the angle was quite steady in the first 20% stance, followed by rapid dorsiflexion until midstance. The change of ankle angle in the slippery group was very steady as reflected by the profile of ankle angular velocity. Such small variations were achieved by maintaining a stiff ankle joint, and may help the subject to better maintain balance and stability on a slippery walking surface.

The knee joint parameters of the three groups showed a similar trend and range. At the first 5% stance, there was a small magnitude of knee extension, followed by rapid knee flexion of about 20° until midstance. This is again not in agreement with the summarized results published (14), which state that the first phase of knee flexion occurred during the first 30%, followed by some knee extension until midstance. Significant differences were only found from 40% to 50% stance for knee angular velocity. From the kinematic results, the ankle joint appeared to play a more active role than the knee joint in preventive measures against slips when walking on a potentially slippery walking surface.

In real construction worksite environments, many other risk factors for slips often occur, including irregular walking surfaces and obstacles on walking surfaces (4). Moreover, workers often need to carry various loads at the worksite. This may introduce a slipping risk and also affect the severity of fall. However, the mentioned factors are difficult to simulate, and therefore their effects cannot easily be investigated. Another important factor is human anticipation (2). The most hazardous situation was believed to be sudden loss of traction due to a sudden drop of available friction in the presence of surface contaminant and without anticipation. Suitable signs or notices should be placed in certain areas in construction worksites to alert workers.

Previous studies mainly investigated the gait changes in the lower extremity (1,2). However, upper-extremity kinematics may also reflect the strategy and adaptation evoked by humans. In normal level walking, the upper-extremity movements are always in opposition to the lower-extremity movements to balance the turning moment along either the sagittal, longitudinal, or frontal axis. From observation, it appeared that the subjects could also alter the upper-extremity pattern to achieve a gentle foot strike and flatfoot landing, and finally reduce the required friction from the ground. Moreover, changes in plantar-pressure distribution during stance may also reflect adaptation strategy. In obtaining kinetic and kinematic data, joint forces and moments could be determined to help understand the human strategy of slip prevention. Future similar studies are suggested to include upper-extremity kinematics, normal and shear reaction forces during stance, and plantar-pressure distribution information to give a better picture of human preventive measures for slips.

CONCLUSION

The presence of an oil contaminant on a wooden walking surface introduced slipping potential in level walking when wearing either safety or cloth shoes. Effective lower- extremity changes to prevent slips evolved by humans in terms of gait pattern included increasing stance time and stride time, shortening stride length, decreasing propagation speed, and having a more gentle foot strike in walking. The ankle joint played an important role in slip prevention strategy. Such a strategy included reducing ankle range of motion, maintaining a stiff ankle joint, and achieving a flatfoot landing or plantarflexed ankle joint during the first 10% stance.

To prevent occupational slips at construction worksites, workers are advised to walk slowly with shortened stride length and longer foot-contact duration. Workers should avoid kicking the floor during the foot strike, as this will increase the heel horizontal velocity and the required friction for walking without slipping. Moreover, to enhance construction worksite safety, the presence of oil contaminants on wooden walkways should be avoided. Suitable signs and notices should be placed in areas with frequent occurrences of wooden walking surfaces and machinery lubricant leakage to attract workers’ attention. Workers should also strengthen their ankle joint mobility by proper exercises. Before working at construction sites, warm-up exercises of ankle joint movement should be performed.

REFERENCES

1. Cham, R., and M. S. Redfern. Lower extremity corrective reactions to slip events. J. Biomech. 34:1439–1445, 2001.
2. Cham, R., and M. S. Redfern. Changes in gait when anticipating slippery floors. Gait Posture 15:159–171, 2002.
3. Grieve, D. W. Slipping due to manual exertion. Ergonomics 26:61–72, 1983.
4. Grönqvist, R., W. R. Chang, T. K. Courtney, T. B. Leamon, M. S. Redfern, and L. Strandberg. Measurement of slipperiness: fundamental concepts and definitions. Ergonomics 44:1102–1117, 2001.
5. Grönqvist, R., J. Roine, E. Jarvinen, and E. Korhonen. An apparatus and a method for determining the slip resistance of shoes and floors by simulation of human foot motions. Ergonomics 32: 979–995, 1989.
6. Lockhart, T. E., J. C. Wodlstad, and J. L. Smith. Effects of age-related gait changes on the biomechanics of slips and falls. Ergonomics 46:1136–1160, 2003.
7. Manning, D. P., I. Ayers, C. Jones, M. Bruce, and K. Cohen. The incidence of underfoot accidents during 1985 in a working population of 10,000 Merseyside people. J. Occup. Accidents 10:121–130, 1988.
8. Maynard, W. S. Tribology: preventing slips and falls in the workplace. Occup. Health Safety 71:134–140, 2002.
9. Newton, R., B. Doan, M. Meese, et al. Interaction of wrestling shoe and competition surface: effects on coefficient of friction with implications for injury. Sports Biomech. 1:157–166, 2002.
10. Occupational Safety and Health Council (Hong Kong). Survey on the community awareness of and attitude towards occupational safety and health in Hong Kong. Hong Kong: Occupational Safety and Health Council Publications, 2000, pp. 7.
11. Occupational Safety and Health Council (Hong Kong). Injury data analysis for the Tuen Mun Community. Hong Kong: Occupational Safety and Health Council Publications, 2000, pp. 8.
12. Occupational Safety and Health Council (Hong Kong). A questionnaire survey on the occupational safety and health of construction workers attending the Green Card course. Hong Kong: Occupational Safety and Health Council Publications, 2003, pp. 16–22.
13. Redfern, M. S., and B. Bidanda. Slip resistance of the shoe-floor interface under biomechanically-relevant conditions. Ergonomics 37:511–524, 1994.
14. Redfern, M. S., R. Cham, K. Gielo-Perczak, et al. Biomechanics of slips. Ergonomics 44:1138–1166, 2001.
15. Strandberg, L. On accident analysis and slip-resistance measurement. Ergonomics 26:11–32, 1983.
16. United States Department of Labor, Bureau of Labor Statistics. Lost-worktime injuries and illnesses: characteristics and resulting time away from work, 1996, USDL 98-157. Washington, DC: U.S. Government Printing Office, 1998.
17. Winter, D. A., A. E. Patla, J. S. Frank, and S. E. Walt. Biomechanical walking pattern changes in the fit and healthy elderly. Phys. Therapy 70:340–347, 1990.
18. Winter, D. A. The Biomechanics and Motor Control of Human Gait: Normal, Elderly and Pathological, 2nd Ed. Waterloo, Canada: University of Waterloo Press, 1991, pp. 87–94.
Keywords:

OCCUPATIONAL SLIPS AND FALLS; INJURY PREVENTION; GAIT ADAPTATION; SLIPPING POTENTIAL MEASUREMENT

©2005The American College of Sports Medicine