University of Southern California Dissertations and Theses

Human and environmental factors contributing to slip events during walking

(USC Thesis Other)

Human and environmental factors contributing to slip events during walking

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content HUMAN AND ENVIRONMENTAL FACTORS CONTRIBUTING TO SLIP EVENTS DURING WALKING Copyright 2003 by Judith Marie Bumfield A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (BIOKINESIOLOGY) August 2003 Judith Marie Bumfield R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. UMI N um ber: 3 1 1 6 6 7 3 Copyright 2003 by Burnfield, Judith Marie All rights reserved. INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3116673 Copyright 2004 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90089-1695 This dissertation, written by J u d ith M arie B u r n fie ld under the direction of h s r dissertation committee, and approved by all its members, has been presented to and accepted by the Director of Graduate and Professional Programs, in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY 'ector Date A u gu st 1 2 . 2003 Dissertation Committp&. ^ 4 AM Chair s.. " ^ . v . R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. DEDICATION To the Bumfield-Wells family and Buddy R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ACKNOWLEDGEMENTS I am truly grateful for the intellectual, financial, and emotional support provided by many individuals and organizations during my doctoral work. I extend my deepest appreciation to faculty, staff, and student colleagues at the University of Southern California, my work colleagues at the Pathokinesiology Laboratory at Rancho Los Amigos National Rehabilitation Center, my technical consultants for slips and falls, and very importantly, my family and friends. You made this possible. Dr. Hislop, thank you for inviting me to join the Biokinesiology program, and believing that I would succeed. Dr. Gordon, thank you for continuing to support my efforts in the program. I am grateful, Dr. Powers, for the time and effort you invested in developing my research skills over your five years as Chair of my Dissertation Committee. Dr. Perry, I have valued your influence on my career and life. Over the last seventeen years, you have continuously challenged and inspired me, and these efforts have been greatly appreciated. Dr. Salem, thank you for your wise words and gentle spirit. Dr. McNitt-Gray, I valued the unique perspective you contributed to the analysis and interpretation of my data. You’ve planted a seed that will grow. Dr. Azen, a special thanks for your many consults related to statistics (not many individuals would make themselves accessible even while vacationing in Europe). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. On a broader level, I would like to express my appreciation to the many faculty members who took the time to critique my work critically. I can only speculate on how much time went into providing feedback with a level of detail that contributed meaningfully to my ability to think. Thank you, as well, to the individuals who worked “behind the scenes” in the Department of Biokinesiology and Physical Therapy. You kept things moving forward and made my life as a graduate student easier. I was fortunate to work in two research labs while completing my dissertation. I warmly appreciate my student colleagues in the Musculoskeletal Biomechanics Research Laboratory in the Department of Physical Therapy and Biokinesiology at USC. Your diverse perspectives and willingness to challenge are greatly appreciated. One particular lab mate, Yi-Ju Tsai, MS, PT, was an invaluable resource during the planning and data collection phases of my dissertation. Thank you. I value, as well, the wealth of experiences possessed by my work colleagues at the Pathokinesiology Laboratory at Rancho Los Amigos National Rehabilitation Center. Thank you for sharing your knowledge, challenging my thought processes, and being flexible as I completed my doctoral work. This dissertation was generously supported, in part, by awards and scholarships from a number of sources including the American Society of Safety Engineers and National Institute for Occupational Safety and Health, the California Physical Therapy Fund, the Jacquelin Perry Scholarship, the Physical Therapy R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. V Foundation, and the Whitaker Foundation. Additionally, John Brault, MS, PT of Macinnis Engineering Associates, Inc., and Janies E. Flynn, P.E. of J2 Engineering, Inc. provided both invaluable technical consults, as well as, generous support for equipment purchases. Thank you. Finally, it is with very deep sense of appreciation that I thank my family and friends for their ongoing support and presence in my life. You have been there when I most needed you. Thank you so much. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. v i TABLE OF CONTENTS DEDICATION.................................................................................................................ii ACKNOWLEDGEMENTS..........................................................................................iii LIST OF TABLES...................................................................................................... viii LIST OF FIGURES...................................................................................................... ix ABSTRACT.................................................................................................................... xi CHAPTER I: OVERVIEW............................................................................................1 Specific Aims........................................................................................................5 CHAPTER II: LITERATURE REVIEW ...................................................................6 Statement of Problem........................................................................................... 6 The Role of Human and Environmental Factors in Slips..................................7 Utilized Coefficient of Friction...............................................................9 Predicting Peak Utilized Coefficient of Friction..................................13 Tribometric Evaluation of Floor Slip Resistance.................................17 Existing Standards for Floor Slip Resistance.......................................23 Predicting Slip Probability.................................................................................25 Summary................................................. 27 CHAPTER III: INFLUENCE OF AGE AND GENDER ON UTILIZED COEFFICIENT OF FRICTION DURING WALKING AT DIFFERENT SPEEDS..........................................................................................28 Introduction.........................................................................................................29 Methods............................................................................................................... 32 Subjects.................................................................................................. 32 Instrumentation......................................................................................33 Procedures.............................................................................................. 34 Data Analysis......................................................................................... 35 Statistical Analysis.................................................................................37 Results................................................................................................................. 38 Peak Utilized Coefficient of Friction................................................... 38 Normalized Stride Length.....................................................................41 Discussion...........................................................................................................43 Conclusions.........................................................................................................45 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. v ii CHAPTER IV: THE ROLE OF CENTER OF MASS KINEMATICS IN PREDICTING PEAK UTILIZED COEFFICIENT OF FRICTION DURING W ALKING............................................................................ 47 Introduction.........................................................................................................48 Methods............................................................................................................... 51 Subjects...................................................................................................51 Instrumentation......................................................................................52 Procedures.............................................................................................. 52 Data Analysis......................................................................................... 54 Statistical Analysis.................................................................................57 Results................................................................................................................. 57 Relation between CM Kinematic Variables and Peak Utilized Coefficient of Friction.............................................................58 Discussion...........................................................................................................62 Conclusions.........................................................................................................65 CHAPTER V: PREDICTION OF SLIPS: AN EVALUATION OF UTILIZED COEFFICIENT OF FRICTION AND AVAILABLE SLIP RESISTANCE 66 Introduction.........................................................................................................67 Methods............................................................................................................... 71 Subjects.................................................................................................. 71 Instrumentation......................................................................................72 Procedures.............................................................................................. 74 Data Analysis......................................................................................... 78 Statistical Analysis.................................................................................80 Results................................................................................................................. 81 Slip Probability based on Slip Resistance Difference.......................... 81 Slip Probability based on Available Slip Resistance........................... 84 Discussion...........................................................................................................85 Conclusions.........................................................................................................88 CHAPTER VI: SUMMARY AND CONCLUSIONS.............................................90 Summary.............................................................................................................90 Conclusions.........................................................................................................94 REFERENCES.............................................................................................................. 95 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. LIST OF TABLES Table 3-1. Physical Characteristics of Subject Participants.......................................32 Table 3-2. Peak Utilized Coefficient of Friction during Walking at Slow, Medium, and Fast Speeds.......................................................................... 38 Table 4-1. Subject Characteristics................................................................................51 Table 4-2. Center of Mass Kinematic Variables at Time of Peak Utilized Coefficient of Friction during Weight Acceptance..................................58 Table 4-3. Simple and Partial Correlation Coefficients for Center of Mass Kinematic Variables and Peak Utilized Coefficient of Friction............. 61 Table 5-1. Physical Characteristics of Study Participants..........................................71 Table 5-2. Logistic Regression Model Details describing the Probability of a Slip Event based on the Slip Resistance Difference......................... 83 Table 5-3. Logistic Regression Prediction of Slip Resistance Difference and Corresponding Probabilities of a Slip Event Occurring...................83 Table 5-4. Logistic Regression Model Details describing the Probability of a Slip Event based on the Available Slip Resistance ......................... 85 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. LIST OF FIGURES Figure 1-1. Onset and recovery from a slip.................................................................. 2 Figure 2-1. Causes of falls in older adults.................................................................... 7 Figure 2-2. Stance phase ground reaction forces and utilized coefficient of friction for a single stride during walking at a self-selected velocity.......................................................................................................10 Figure 2-3. Trigonometric calculations used to estimate impact angle (relative to vertical) and utilized coefficient of friction during walking.......................................................................................................13 Figure 2-4. Prediction of peak utilized coefficient of friction for persons with short versus long legs........................................................................14 Figure 3-1. Representative tracing of ground reaction forces and utilized coefficient of friction during shod walking at a slow speed for a senior female subject....................................................................... 36 Figure 3-2. Between gender and across age group differences in average peak utilized coefficient of friction during shod walking at slow, medium, and fast speeds.................................................................40 Figure 3-3. Average normalized stride length during shod walking at slow, medium, and fast speeds.................................................................41 Figure 3-4. Relationship between normalized stride length and peak utilized coefficient of friction across all walking speeds.......................42 Figure 4-1. Center of mass to center of pressure relationships studied....................56 Figure 4-2. Scatter plot showing the relationship between CM-CPAngie and peak utilized coefficient of friction.................................................. 59 Figure 4-3. Scatter plot showing the relationship between CM-CPvei a p and peak utilized coefficient of friction.................................................. 60 Figure 4-4. Significant predictors of peak utilized coefficient of friction during weight acceptance from sagittal and frontal perspectives............................................................................................... 61 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. X Figure 5-1. The ratio of shear to vertical ground reaction forces used to predict peak utilized coefficient of friction during walking...................68 Figure 5-2. The English XL variable incidence tribometer.......................................73 Figure 5-3. The fall arresting harness, attached to an overhead trolley and track, that was worn by all subjects to ensure safety during walking............................................................ 75 Figure 5-4. Scatter plot showing the relationship between slip resistance difference and the probability of a slip event occurring as calculated by logistic regression..............................................................82 Figure 5-5. Scatter plot showing the relationship between the available coefficient of friction and the probability of a slip event occurring as calculated by logistic regression.........................................84 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. XI ABSTRACT During walking, slips are likely to occur when an individual’s utilized coefficient (COFu) exceeds the friction available on the floor surface. To better understand human and environmental factors that predispose an individual to slip onset, three investigations were undertaken. The first study evaluated the influence of age and gender on COFu during walking. Sixty healthy adults were divided into three groups: Young (20-39 y.o.); Middle-aged (40-59 y.o.); and Senior (60-79 y.o.). Ground reaction forces, recorded as subjects walked at three pre-determined speeds, were used to calculate peak COFu during weight acceptance. Averaged across age groups, females had higher peak COFu than males while walking slowly, while males had higher peak COFu than females while walking fast. Averaged between genders, middle-aged subject’s peak COFu was higher at the medium speed than both young and senior subjects, and than senior subjects at the fast speed. The second investigation determined the extent to which center of mass (CM) kinematics could be used to predict peak COFu during level walking. Ground reaction forces and full-body kinematic data were recorded simultaneously as forty- nine healthy young adults walked. Greater CM to center of pressure angles of inclination and faster velocity of the CM in the anterior direction were associated R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. with higher peak COFu. Combined, these two variables explained 62% of the variance in peak COFu. The third investigation identified the probability of a slip occurring based on the relationship between static measures of floor surface slip resistance and an individual’s peak COFu. Video, kinematic, and ground reaction force data were recorded simultaneously as fifty-two subjects traversed a walkway during conditions of normal and reduced floor surface slip resistance. Knowledge of the available static slip resistance, in combination with an individual’s COFu allowed for a prediction of slips accuracy rate of 89.5%. Collectively, the data generated from this research will be used to identify persons and/or situations posing the greatest risk for slip onset. An understanding of the interaction between these human and environmental factors could serve as a basis for proactive human or environmental interventions to reduce the risk of injurious outcomes. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 1 CHAPTER I OVERVIEW Slips have been identified as a leading cause of falls and injuries in the 6 62 5 23 62 home ’ and work ’ ’ environment. During walking, forces generated by the body are transmitted through the foot to the floor. In order to prevent a slip, sufficient friction at the foot-floor interface is required. When the ratio of shear to vertical forces applied to the floor surface exceeds the available friction at the foot- floor interface, a slip becomes imminent.4 6 When observing a slip, there appear to be two distinct phases: onset and recovery. The onset of a slip can occur at any point while the foot is on the ground, however, the most dangerous slips occur briefly after initial contact as weight is being loaded onto the limb (Figure 1-1 a).9 7 Without sufficient frictional forces, the foot slides away from the body, balance is disrupted, and a fall may occur. Successful recovery from a slip results from an interaction of human and environmental factors to re-align the body over a stable base of support (Figure 1.1b). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 a) ONSET b) RECOVERY Figure 1-1. Onset and recovery from a slip. Inadequate friction at the foot-floor interface is the primary factor contributing to slip onset (a). The success of slip recovery is influenced by multiple factors including changes in available friction, balance reactions, and strength (b). The black line has been aligned with the heel location at initial contact. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3 The purpose of this dissertation is to better understand the human and environmental factors that can predispose an individual to slip onset. To explore this question, it will be necessary to control for a number of factors already known to contribute to slip onset (e.g., footwear characteristics,2 1 ’31,32,37,40,43,58‘60,69'71,80,83, 102,1 1 6 floor type,18 ,1 9 ’54,60,69’91,108,113,1 1 6 and slip anticipation1 7 ) while manipulating variables that are not as well understood (e.g., walking velocity, the slipperiness of the floor surface, gender, age). Efforts aimed at identifying individuals and groups of persons at risk for slip onset due to high friction needs remain limited. The majority of studies evaluating the friction utilized during walking have focused on the requirements to safely negotiate level surfaces.12,16,31,46,61’109,1 1 6 Studies documenting the friction needs during weight acceptance for healthy adults have reported average peak values ranging from n = 0.08 to ju = 0.35. With disability, friction needs are even higher, fi = 0.64.1 2 It is unclear from the existing literature, what influence variables such as velocity, gender, and age have on friction needs. Chapter III will examine the relationship between these human factors and friction needs during walking. Additionally, what remains elusive is the underlying biomechanical mechanism contributing to these apparent differences in friction needs across persons. Elucidation of such mechanisms would allow for the identification of individuals at greater risk for slip onset. This issue will be addressed in Chapter IV. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 Chapter V of this dissertation will address the relationship between human and environmental factors in predicting the onset of a slip. Based on current theory, “sufficient” friction at the foot-floor interface should prevent the onset of a slip. However, no universal method of measuring slip resistance has been established. This is evident by the fact that over 50 different tribometers (devices used to measure slip resistance) have been described in the literature.9,10,20,25,26,30,34,37,41,42,53,6 1 ’7 3 , 74,76,78, 79,84,97,102,108,112,118 different types of tribometers yield different measurements, that frequently do not correlate well with each other.8,26,34,36,41,42,61, 76,78,100,112 As measures of slip resistance vary across the type of device used, it is important to establish the specific relationship between slip resistance values recorded and slip outcome for the more frequently used tribometers. The purpose of Chapter V will be to determine the probability of a slip occurring based on the relationship between static measures of available floor surface slip resistance and utilized coefficient of friction measures. Taken together, the studies within this dissertation provide valuable information that may be used to identify persons at greater risk for slip onset. Collectively, it is anticipated that this information may be used to improve the safety of walkway surfaces in the work and home environments. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 5 SPECIFIC AIMS The principle aims of this dissertation were to: 1. Determine the influence of age and gender on peak utilized coefficient of friction values during walking at different speeds across a level surface (Chapter III). 2. Identify the extent to which center of mass kinematics could be used to predict peak utilized coefficient of friction during the weight acceptance phase of level walking (Chapter IV). 3. Determine the relationship between static measures of slip resistance and peak utilized coefficient of friction values on the probability of a slip occurring during level walking (Chapter V). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 CHAPTER II LITERATURE REVIEW STATEMENT OF PROBLEM Slipping is one of the most common causes of falls,4" 6 ’2 3 ’55,68,9 2 and is of major concern to industry and society due to the associated human suffering and 23 29 55 63 77 financial costs. ’ ’ ’ ’ In the work environment, slipping has been identified as the primary antecedent event to falls on both stairs and level surfaces,5 ’23,6 2 accounting for approximately 62% of underfoot accidents.6 8 Courtney et al,2 3 reported that slipperiness or slipping contributed to 40 to 50% of fall related occupational injuries. Additionally, according to the U.S. Bureau of Labor Statistics (1996), slips, trips and falls accounted for 11% of fatal occupational injuries. In the United States, it has been estimated that the annual costs associated with occupational injuries will exceed $85 billion during the year 2020 when it is projected that more that 17 million falls resulting in injury will occur.2 9 Among older persons, slips have been identified as one of the primary causes of falls.6,6 2 In a one year study of the fall history of community dwelling elders between the ages of 60 and 80 years, slips accounted for 25% of the accidental falls (Figure 2-1).6 Slips were identified as the leading extrinsic (environmental) cause of femoral neck fractures in an investigation of older adults (mean age 78 years) hospitalized for treatment of a femoral neck fracture.9 2 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 Further, slips more frequently result in falls in older compared to younger adults.6 6 This finding is concerning in light of studies which suggest that falls are the leading cause of unintentional injuries resulting in death in persons 65 years of age or older.4 8 Other Sip Loss of Balance Misplaced Step 12% 34% Figure 2-1. Causes of falls in older adults.6 THE ROLE OF HUMAN AND ENVIRONMENTAL FACTORS IN SLIPS The primary factor contributing to the onset of a slip is inadequate friction,3 9 however, additional human and environmental factors may predispose a person to slip.39,1 0 3 For example, the presence of disability has been associated with an increased need for friction from the walking surface.1 2 Knowledge of an impending slip reduces the risk for slip onset. Cham and Redfem1 7 reported that when subjects anticipated that they might slip, they reduced their friction needs by 16-33 % by adopting a safer walking strategy which included R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 reducing stance duration, stride length, heel velocity prior to contact, and loading rate at weight acceptance. Impaired vision, altered cognition, or distractions may reduce a person’s ability to perceive of a contaminant (e.g., an oil or water patch) and adaptively alter their gait to reduce the risk of a slip. Slip onset and recovery are influenced by a number of environmental factors including the presence of a contaminant, and the characteristics of footwear and flooring. The type, amount, and depth of contaminant will alter the friction available from the floor surfaces, as well as the potential to regain traction.37,54,69,90,1 0 2 Shoe tread patterns (e.g., grooved vs. ungrooved), as well as soling material hardness, composition (e.g., leather vs. rubber vs. polyvinyl chloride), and texture (e.g., rough vs. smooth) alter the interaction with slippery substances.21,31,32,37,40,43,58'60,69'7 1 ,8 0 , 83 102 116 ’ ’ Similarly with flooring, the surface coating, composition, and roughness also modify the risk of slip onset.18,19,54,60,69,91,108,113,1 1 6 Recovery from a slip relies, in part, on the ability to perceive and respond to the perturbation. Marigold and Patla7 2 reported that in response to a simulated slip, the average reactive postural muscle onset latency was 146-199 ms, suggesting polysynaptic reflexes associated with the proprioceptive input contributed to the balance response. These data suggest that individuals with impaired proprioception may have a more difficult time with balance recovery during a slip. With aging, reductions in fast twitch muscle fibers6 5 also may reduce the ability to generate forceful postural responses required for slip recovery.114,1 1 5 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 9 In summary, multiple human and environmental factors contribute to both the onset of a slip and to the potential for successful recovery. The focus of this dissertation will be on the biomechanical interactions contributing to slip onset. Utilized Coefficient of Friction During ambulation, slips occur when the utilized coefficient of friction (COFu) exceeds the friction available from the walking surface. In the laboratory setting, an individual’s COFu during walking is calculated from force plate recordings of ground reaction forces (Figure 2-2a).u The COFu is defined as the ratio between the shear (resultant of the fore-aft and medial-lateral forces) and vertical components of the ground reaction forces recorded as a person walks across a dry, non-contaminated surface (Figure 2-2b).1 0 5 Persons with higher COFu are at greater risk of slipping when adequate friction is not available from the floor surface.4 6 During level walking, the COFu peaks during two periods: weight acceptance, and push-off (Figure 2-2b).9 5 The peak in COFu during weight acceptance represents the most critical value. If adequate friction is not available at the floor surface during weight acceptance, then the foot will slip away from the body, creating an unstable base on which to load body weight.9 7 The peak COFu during push-off is less detrimental as it occurs at a time when substantial body weight has already transferred to the leading limb. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. a) FORCES — — Vertical Anterior-Posterior Medial-Lateral 1000 800 « 600 400 | 200 s: a * * -200 -400 b> COEFFICIENT OF FRICTION 0.40 Peak COFu during Push-Off ? & 0.30 'I 0.20 > - 0.10 < 3 £ 0.00 £ -0.10 o u -0.20 Peak COFu during Weight Acceptance 60 -0.30 0 20 40 80 100 % Stance Duration Figure 2-2. Stance phase ground reaction forces (a) and utilized coefficient of friction (b) for a single stride during walking at a self-selected velocity. Utilized coefficient of friction peaks during weight acceptance when a braking force is required to prevent the foot from slipping, and in push-off when a propulsive force contributing to limb advancement. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 11 The majority of studies evaluating COFu during walking have focused on the requirements to safely negotiate level surfaces.12,16,31’3 2 ’46,6 1 ’109,1 1 6 Studies documenting COFu during weight acceptance for healthy adults have reported peak values ranging from n = 0.08 to // = 0.35. What is unclear from the existing literature is what factors contribute to the observed range in peak COFu- For example, Cham et al1 6 reported that the peak COFu was only // = 0.15 (± 0.05) in their study of six young adults (18 to 30 years; two male, four female) walking at an unreported self selected speed. In contrast, Fendley and colleagues3 2 reported a peak of// = 0.35 for a single subject walking at speeds ranging from approximately 102 to 160 m/min. These authors also noted a fair positive correlation (r = 0.41 ;p < 0 .001) between peak COFu and walking speed in the single subject for whom data were reported. Similarly, Skiba1 0 9 in a review article reported that peak COFu increased with walking speed, however, specific subject data were not provided. Buczek et al1 2 reported no difference in peak COFu recorded during self-selected slow and fast walking speeds. However, peak COFu at the slow (ju = 0.49) and fast (ju = 0.52) speeds was averaged across a rather diverse subject group which included younger and older persons (males and females 17 to 71 years of age) with various medical diagnoses (e.g., cerebral vascular accident, above knee amputation below knee amputation, hip fracture) as well as healthy younger adults (males only, 24 to 34 years). Tisserand1 1 6 reported that peak COFu during weight acceptance (range, // = 0.08 to n - 0.15) did not demonstrate a significant relationship to walking speed in a R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 12 study of three subjects (gender and age not reported) walking at speeds ranging from 90 to 165 m/min. Very few studies have focused on the influence of age on utilized coefficient of friction. Lockhart and colleagues6 7 compared slip responses during level walking in younger (18-29 year old) and older individuals (65+ years of age). The authors reported that once a slip was initiated, subjects in the older group, on average, slipped farther (11.8 vs. 5.0 cm) and had higher dynamic COF than their younger counter parts. Peak COFu values were not reported by Lockhart et al. Thus, from the existing literature, it is unclear whether older adults have higher peak COFu values and thus require more slip resistance from a floor surface to prevent slip onset; or whether once a slip is initiated, older adults have a more difficult time recovering, and thus fall. Due to the high risk of injury associated with slips and • O ') falls in older adults, greater attention to the peak COFu requirements of older adults appears warranted. Collectively, these studies highlight one deficit in the literature, namely no study has been conducted to systematically evaluate the influence of variables such as age, gender, and walking speed on peak COFu- As it is well known that walking characteristics differ across the age spectrum7,35,45,56,86,87,89,1 2 0 and between c n o/i o q genders, ’ ’ it is probable that the slip resistance requirements will also be influenced by these variables. The purpose of Chapter III was to investigate the influence of age and gender on COFu at different walking speeds. Understanding R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 13 how these factors may relate to peak COFu, and thus the potential for slip onset, will assist in identifying groups of individuals at greater risk for slipping. Predicting Peak Utilized Coefficient of Friction Exploratory investigations aimed at identifying groups at risk for slip onset could be more focused if the underlying human biomechanical mechanisms which contribute to higher COFu during walking were well understood. One relatively simple model to predict peak COFu used measurements of step length and leg length to estimate the angle of impact of the leg with the ground at initial contact (Figure 2- 27 28 3). ’ A geometric approximation of peak COFu was determined using the tangent of the impact angle to estimate the ratio of horizontal to vertical ground reaction forces. Leg Length Utilized COF = Vi Step Length Horizontal Force (Fh) Vertical Force (Fv) Figure 2-3. Trigonometric calculations used to estimate impact angle (relative to vertical) and utilized coefficient of friction (COF) during walking [Fv = vertical ground reaction force. Fh = horizontal ground reaction force. 0= impact angle (relative to vertical)].1 3 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 14 Based on this model, it would be predicted that individuals who have short legs, yet take long steps would be have higher peak COFu values compared to those with long legs who take short steps (Figure 2-4). A) Person with a shorter leg B) Person with a longer leg Regular Step Long Step Regular Step Long Step W M Figure 2-4. Prediction of peak utilized coefficient of friction for persons with short versus long legs. Static trigonometric estimations of the utilized coefficient of friction generated during walking would predict that for a given step length, a greater utilized coefficient of friction is generated by a person with a shorter leg (a) than a person with a longer leg (b).1 3 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 15 Many state laws and building codes have established that a static COF of /u = 0.50 represents the minimum slip resistance threshold for safe flooring surfaces based, in part, on estimates of peak COFu derived from this model.84,1 0 5 However, subsequent study of this model by Powers et al." revealed that the model overestimates actual peak COFu by approximately 86% at slow walking speeds and up to 131 % at fast walking speeds. Additionally, only 27% of the variance in peak COFu recorded in young adult during walking could be explained based on the relationship between a subject’s leg length and step length." One limitation of the previously described model27,2 8 is that it assumes that at initial contact, the resultant GRF vector followed a line starting at the heel and passing through the hip joint center. This implies that the leg functions as a rigid strut when applying forces to the ground. Studies of human gait, however, do not support this premise. Immediately prior to initial contact, the body goes through a period of rapid descent causing the leading limb to drop vertically by approximately oc 1 cm. This period of descent results in an abrupt vertical loading of the leading limb (60% of body weight in approximately 0.02 seconds),9 8 reducing the ratio of shear to vertical GRFs at initial contact. Also, muscle activity at the hip, present prior to initial contact to decelerate forward momentum of the limb, alters the relative ratio of shear to vertical GRFs.9 8 As COFu is derived from ground reaction forces, and ground reaction forces represent the algebraic summation of all mass-acceleration products of all body R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 16 33 119 • segments ’ it is conceivable that measures describing the motions of the total body center of mass (CM) during walking would serve as predictors of COFu. During healthy, steady-state walking, the CM demonstrates a characteristic pattern of motion in each of the three cardinal planes.14,15,50,52,64,1 1 9 CM displacement in the direction of progression is determined primarily by walking velocity. In the vertical direction, the CM begins to move upward shortly after initial contact, reaching it greatest height at the end of mid stance. It then lowers to its original position by the end of stance. With faster walking velocities, the amount of vertical displacement increases, with reported values ranging from 1.3 cm at 30 m/min to 3.1 cm at 150 m/min.50,6 4 Movement of the CM from side-to-side (medial-lateral) also occurs, creating a sinusoidal pattern in the horizontal plane. Maximal displacement of the CM to the right occurs during mid stance on the right limb, while maximal displacement to the left occurs during mid stance on the left limb. Increases in the base of support width lead to greater medial-lateral excursions of the CM.5 1 CM acceleration also demonstrates a characteristic pattern during walking.1 4 , I c In the anterior direction, the CM accelerates following initial contact and through the first half of loading response. The second half of loading response is characterized by CM deceleration in the anterior direction, reflecting the slowing of the CM anterior movement as the mass rises against gravity. Acceleration of the CM again increases in late stance as the body experiences a controlled forward fall prior to initial contact. In the vertical direction, downward acceleration of the CM initially R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 17 increases as weight loads onto the limb following initial contact. Following limb loading, a gradual upward acceleration of the CM occurs through the middle portion of single limb support. Then in late stance, as the body CM begins a controlled drop, the CM accelerates downward prior to the opposite limb contacting the ground. While characteristic CM kinematic patterns have been documented during normal walking, subtle variations in the relative location as well as the velocity of the CM exist across individuals. These variations could influence the forces applied to the floor and therefore slip potential. The purpose of Chapter IV was to investigate the extent to which CM kinematics could be used to predict the peak COFu during the weight acceptance phase of level walking. Knowledge of the relationship between CM kinematics and peak COFu may make it possible to identify individual anthropometric and gait characteristics that combine to increase the risk of slip onset. Tribometric Evaluation of Floor Slip Resistance The slip resistance provided by a given floor surface is typically measured using a device called a tribometer. Currently, no universal method of measuring slip resistance has been established. Additionally, different types of tribometers yield unique slip resistance values, that frequently do not correlate well with each other.8 , 26,34,36,41,42,61,76,100, 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 Further complicating the issue of measuring floor surface slip resistance is a lack of consensus regarding which friction is most important to measure, static or dynamic. The static coefficient of friction is a measure of the ratio of shear to vertical forces between two contacting objects (e.g., the shoe and floor) at the instant when relative motion begins.7 5 In contrast, dynamic coefficient of friction describes the ratio of forces after the sliding motion has been initiated.75,1 0 5 Adequate static coefficient of friction appears key to preventing the initiation of a foot slip,27,28,38,44, 96,1 1 8 however, once a slip is initiated, sufficient dynamic friction appears necessary to prevent a slip from progressing to a fall.38,44,48,95,' 1 1 8 Perkins and Wilson95,9 1 ’ 118 developed a method for assessing static and dynamic coefficient of friction between the shoe and floor and related these values to slips observed in subjects walking across similar surfaces to those tested. When the available static and dynamic friction was high, no slips occurred. When both available coefficient of friction values were low, slips consistently occurred. Interestingly, when static friction was high, but dynamic friction was low, slips occurred less frequently, but when the individual slipped, it was always “dangerous”. In situations where static coefficient was low, but dynamic friction was high, slips frequently occurred, but subjects did not slip fast enough to lose their balance. Some authors suggest that tribometers are not measuring the actual “friction” between two surfaces when a contaminant is present, as the tribometer test foot may not be in direct contact with the underlying surface. It has been suggested that the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 more general term, “slip resistance”, be used to describe the relative slipperiness between two objects when a contaminant is present.25,8 4 Over 50 different tribometers have been described in the literature.9,10,20,25, 2 6 ,3 0 ,3 4 ,3 7 ,4 1 ,4 2 ,5 3 ,6 1,73,74,76,84,97,102,108,112,118 A u • .(• • * ’ > A brief review of the more common types of tribometers, related standards, and device limitations will be highlighted. The first tribometer in the United States was an articulated strut device developed by Hunter in the 1920’s 4 9 The general operating principle underlying an articulated strut devices is that a known constant vertical force is applied through a strut to a test pad that is in constant contact with the test surface. The angle of inclination of the strut can be gradually increased until the point at which the test pad moves on the surface. The angle of inclination of the strut is then used to estimate the relative ratio of shear to vertical forces, i.e., the available friction. The James Machine, also an articulated strut device, was developed in the 1940s by Sidney James and Underwriter’s Laboratory. Sacher1 0 5 and Marpet7 4 provide a detailed historical overview of the development and subsequent impact that James Sidney and the James Machine had on slip resistance standards. In particular, it was James Sydney who initially recommended that a threshold of /u = 0.5 (as measured by the James Machine) be established as the minimum static coefficient of friction for adequate underfoot safety. Underwriter’s Laboratory currently uses the non-portable James Machine to rate surfaces as “slip resistant” as determined by whether the material or coating achieves apt> 0.50 slip index rating (UL410 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 standard). It is also the device specified for the assessment of the slip resistance of dry floor polishes in ASTM D2047 (Standard Test Method for Static Coefficient of Friction of Polish-Coated Floor Surfaces as Measured by the James Machine) with the method of use specified in ASTM F489 (Standard Test Method for Using a James Machine).1 The Portable Articulated Strut Tribometer (PAST) is a smaller and lighter version (10 lbs. vs. 75 lbs. weight applied through strut) of the James Machine which is approved for testing the slip resistance of dry surfaces (ASTM, F1678- Standard Test Method for Using a Portable Articulated Strut Slip Tester).1 Neither the James Machine nor the PAST has been approved for testing the slip resistance of wet surfaces, due to the “stick-slip” phenomena (commonly referred to as “sticktion”).30,1 1 0 Devices which allow the test foot to reside on the wet surface prior to applying the horizontal force are subject to the build-up of adhesion forces between the test foot and surface.1 0 1 Unrealistically high coefficient of friction readings, due to the adhesion forces, results.6 1 The portable inclineable articulated strut tribometer (PIAST) uses gravity to drop a 10 lb. test foot onto the floor surface at progressively more horizontal angles until a slip occurs.2 5 It avoids the sticktion problem associated with portable articulated strut tribometers by not allowing the test foot to reside on the ground surface prior to horizontal force application. It has been approved for testing both dry and wet surfaces (ASTM F1677).1 Powers et al,1 0 0 however, identified that PIASTs may record extremely low available friction measurements under wet testing Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 conditions The authors attributed these low measurements to water becoming trapped between the test foot and the floor, contributing to a “hydroplaning” effect. The variable incidence tribometer (VIT) also measures static coefficient of friction and is based on the same underlying principles of strut machines. It uses a carbon dioxide cartridge versus a weight to project the test foot to the floor. As the test foot is not allowed to reside on the ground, sticktion is minimized. Also, only the posterior aspect of the test foot initially contacts the ground, and then a flexible joint allows the front portion of the test foot to contact the ground. This test foot contact pattern may explain why the VIT provided higher slip resistance values on wet surfaces than the PIAST,1 0 0 as the hydroplaning effect was likely limited. The VIT has been approved for testing slip resistance on both dry and wet surfaces (ASTM F1679, Standard Test Method for Using a Variable Incidence Tribometer).1 The repeatability and bias of both the PIAST and the VIT have been investigated.1 0 0 During non-slip, dry surface conditions, both tribometers demonstrated good agreement (low level of bias) between the tribometer’s slip-index value and coefficient of friction values recorded from a force plate (ICCs = 0.86 to 0.90). Additionally, good agreement between successive force plate measurements of the horizontal to vertical force ratios generated by each tribometer was documented. Each device, however, gave different slip-resistance values for the same surface condition, indicating that the slip resistance recorded for floor surfaces will vary based on the tribometer used. Therefore, prediction equations, which Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 incorporate tribometric measurements of walkway slip resistance to model the probability of a slip occurring, need to be developed specifically for each tribometer. During the testing of wet surfaces a good level of agreement was demonstrated between the VIT’s slip index value and coefficient of friction values recorded from the force plate (ICC = 0.89).1 0 0 An excellent level of agreement (ICC = 0.99) was documented between successive force plate recordings of COF and the VIT’s slip index values. During testing of the PIAST on a wet surface, abnormally low slip index values (0.001) were recorded. Collectively, the results of this study suggest that the VIT would serve as a better tool from which to develop prediction equations modeling the probability of a slip as the VIT would provide reliable measurements during both dry and wet conditions. Drag sled tribometers exert an increasing horizontal pulling force (parallel to the floor surface) on a known load (either a shoe or other material) until the load starts to move.9 A dynamometer measures the maximal horizontal force achieved before the object starts to move. This horizontal force can be divided by the known load weight to calculate the coefficient of friction.7 6 Drag sled models were developed that could measure both the static and dynamic available coefficient of friction.53,1 0 4 Aberrantly high measurements of available friction on wet surfaces due to the sticktion phenomena61,73,7 6 have limited use of drag sled devices to dry surfaces only (ASTM F609, Standard Test Method for Using a Florizontal Pull Slipmeter;1 ASTM C1028, Standard Test Method for Determining the Static Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 Coefficient of Friction of Ceramic Tile and Other Like Surfaces by the Horizontal Dynamometer Pull-Meter Method). Pendulum testers swing a test foot across the floor surface and calculate dynamic friction based on the reduction in the length of the swing following test foot’s contact with the floor surface.26,4 4 Uneven surfaces can cause wide variations in the contact pressures, and thus, the dynamic friction recorded.4 4 Existing Standards for Floor Slip Resistance Many state laws and building codes have established that a static COF of// > 0.50 represents the minimum slip resistance threshold for safe floor surfaces.84,1 0 5 This threshold arose, in part, from feedback during the early 1950’s to Underwriter’s Laboratories from floor polish manufacturers regarding what was perceived, based on years of experience, to be a safe floor surface.84,1 0 5 Subsequently, a simple model of estimated coefficient of friction requirements, derived from measurements of leg length and step length in 16 subjects, provided additional support for this OH 9R threshold. ’ Estimated coefficient of friction requirements for these subjects ranged from // = 0.30 to // = 0.44, and a safety margin was added to arrive at the threshold value of// > 0.50.27,2 8 Underwriter’s Laboratory currently uses the non portable James Machine to rate surfaces as “slip resistant” as determined by whether the material or coating achieves a// > 0.50 slip index rating (UL410 standard). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 The Americans with Disabilities Act Accessibility Guidelines3 contain advisory recommendations for static coefficient of friction slip resistance values of y > 0.6 for accessible routes (e.g. walkways and elevators) and y > 0.8 for ramps. The foundation for these standards, however, appears to be based on a single study, as a review of the literature reveals only one published investigation exploring the impact 1 9 of disability on slip resistance requirements. Buczek et al. found that persons with a disability had greater slip resistance needs during level walking weight acceptance for both the affected (ju = 0.64) and less affected (ju = 0.48) limbs when compared to the slip resistance used by persons without a disability (ju = 0.31). Additionally, notably high coefficient of friction values (ju = 0.80 to n = 1.10) were reported for selected subjects. The nine subjects who participated in Buczek’s study had a variety of diagnoses including unilateral trans-tibial amputation, transfemoral amputation, unilateral hemiparesis secondary to stroke, hip fracture, broken leg, and osteotomy of the 5th metatarsal. While reduced slip resistance can lead to a slip, excessive friction also can 89 contribute to injuries. Menck and Jorgenson reported on two cases in which excessive frictional forces contributed to sports related ankle fractures. While contact with surfaces with low slip resistance may lead to a slip, it is possible that unexpected contact with a surface that has excessively high friction may contribute to a stumble. Potential adjustments to the minimum threshold for safe flooring must consider the potential negative consequences as well. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 PREDICTING SLIP PROBABILITY While a wide variety of tribometers exist, only two studies have evaluated the probability of a slip occurring based on the relationship between tribometer readings and COFu values.46,6 1 Kulakowski et al6 1 recorded ground reaction forces and calculated peak COFu for five subjects walking at a fast speed across a dry surface. Subjects then walked across three different surfaces on which a film of detergent had been applied, and the incidence of slips was recorded. The authors reported that slips occurred in 79% of the instances when the estimated COFu (projected based on dry walk values) exceeded the estimated available COF (based on previously recorded portable articulating strut tribometer measurements). The accuracy of prediction for this device is difficult to assess for a few reasons. The portable articulating strut tribometer has been approved for testing on dry surfaces only (ASTM FI 678) and thus the validity of the wet surface calculations in this study is of concern due to sticktion. Also, as subjects’ estimated COFu (based on dry walking trials) frequently exceeded the available COF by 200-400%, it is difficult to assess the preciseness of the relationship between available COF and COFu- Finally, as subjects were not masked to the slip trial, and subjects participated in multiple slip trials, it is unclear whether subjects altered their gait, and thus their true COFu, in anticipation of the slip event.1 7 Hanson and colleagues4 6 used a tribometer called the programmable slip resistance tester, and identified that as the difference between COFu (measured with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 force plates) and the available dynamic coefficient of friction (measured with the tribometer) increased, the probability of a slip and/or fall increased in five young males. The COFu was manipulated by adjusting the inclination (0,10, 20 degrees) of the ramp on which subjects walked, and the application of a soapy water solution to the ramp altered the available friction. A threshold COF difference existed, above which, all subjects slipped. Likewise, a threshold COF difference existed, below which, no subjects slipped. Between the upper and lower thresholds, was an intermediate zone in which the probability of slips and/or falls increased as the difference between COFu and available dynamic COF increased. To date, the relationship between COFu and available slip resistance as measured by a device that records static COF has not been established. Given that the variable incidence tribometer is capable of reliably measuring static COF under both dry and wet conditions (i.e., those conditions in which a slip will likely occur), and that a nation-wide standard has been developed for its use (ASTM FI 679), it appears warranted to determine whether this type of tribometer can be used to assess the probability of a slip event occurring during walking. Further, as measures of slip resistance vary across the types of device used, it is important to establish the specific relationship between slip resistance values recorded and slip outcome for the more frequently used tribometers. The purpose of Chapter V was to determine the probability of a slip occurring based on the relationship between available (variable incidence tribometer) and utilized (force plate) coefficient of friction measures. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 SUMMARY Many factors have been identified that can contribute to the onset of a slip, however, several gaps in the existing literature have been recognized. For example, it is unclear whether certain human factors (e.g., as age, gender, walking speed) are related to higher utilized coefficient of friction, and thus greater risk of slip onset. Additionally, the existing biomechanical model27,2 8 which uses leg length and step length to predict peak COFu accounts for only 27% of the variance in peak COFu," suggesting that the existing model does not accurately predict tasks or individuals who might have high COFu. An alternative model, which incorporates measures of the total body CM during walking, may serve as a predictor of peak COFu. Finally, no studies exist relating static measures of coefficient of friction to the probability of a slip event occurring. Thus it is unclear whether static measures, which are frequently used to measure slip resistance in the “real world”, are relevant measures for predicting slip events. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 8 CHAPTER III INFLUENCE OF AGE AND GENDER ON UTILIZED COEFFICIENT OF FRICTION DURING WALKING AT DIFFERENT SPEEDS* A frequently cited theory suggests that ratio of leg length and stride length (i.e., normalized stride length) can be used to predict the utilized coefficient of friction (COF) during walking. As stride length and leg length differs across persons of different ages and genders, it is probable that utilized COF also will vary. The purpose of this chapter was to evaluate the influence of age and gender on utilized COF during non-slip pedestrian gait. Sixty healthy adults were divided into three groups by age (10 males and 10 females in each age group): Young (20-39 y.o.); Middle-aged (40-59 y.o.); and Senior (60-79 y.o.). Ground reaction forces (AMTI forceplate; 600 Hz.) were recorded as subjects walked at slow, medium, and fast speeds. Utilized COF throughout stance was calculated as the ratio of the resultant shear force and vertical force. Reprinted, with permission, from STP 1424 - Metrology of Pedestrian Locomotion and Slip Resistance, copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 9 INTRODUCTION Slipping is a frequent precursor to falls,5 ’6 ’6 8 and is of significant concern among the elderly due to the increased risk of injury.6,29,62,6 3 An investigation of occupational injuries to civilian workers over the age of 55 years, reported that slips accounted for more than half (57%) of the falls occurring on level surfaces.6 2 In a group of community dwelling older adults (60-88 years old), slips contributed to 38% of falls experienced by men and 17% of falls experienced by women during a one year period.6 While one out of every three persons over the age of 65 will fall each year,1 0 6 falls in older women are of even greater concern due to the heightened risk of fractures in the presence of osteoporosis.8 1 As falls are the leading cause of unintentional injuries resulting in death in persons 65 years of age or older,4 8 an understanding of factors that may contribute to slips and falls is critical. Causes of falls include both human and environmental factors. During walking, forces generated by the body are transmitted through the foot to the floor. In order to prevent a slip, sufficient friction at the foot-floor interface is required to counteract the shear forces. When the available friction at the foot-floor interface can not meet the biomechanical demands of walking, a slip becomes imminent.4 6 The forces generated as a person walks across a given surface can be measured by a force plate and used to calculate the utilized coefficient of friction (COF). The “utilized” COF during walking is defined as the ratio between the shear Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 0 (resultant of the fore-aft and medial-lateral forces) and vertical components of the ground reaction force (GRF). A frequently cited theory related to the assessment of walkway slip resistance suggests that the angle form by the lower limb at ground impact is predictive of the utilized COF generated during walking.27,2 8 This theory states that the tangent of the angle formed by the lower limb (relative to vertical) at foot impact is equal to the ratio of shear to normal forces at foot strike (Figure 2-3). This model indicates that, at impact, the angle of the lower limb and the predicted utilized COF would be influenced by two factors: leg length, and step length. Ekkebus and Killey27,2 8 suggested that the most dangerous slip resistance condition would occur when persons with shorter legs were forced to take a longer step, as the utilized COF requirements would be considerably increased (Figure 2-4). As it is well-known that walking characteristics differ across the age spectrum7,35,45,56,86,89,1 2 0 and between genders,57,86,8 9 it is probable that utilized COF also will be influenced by these variables. In healthy adults, gait characteristics such as velocity and stride length remain relatively unchanged until the seventh decade of life.47,9 8 After 60 years of age, reductions in velocity have been documented,47,9 8 and occur, in large part, due to decreases in stride length of approximately 7-20%.45,86,87,93,1 2 0 As stride length decreases with age, static calculations based on these data would suggest that the utilized COF generated by older adults would be less than that generated by younger persons. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 It is also well accepted that on the average, women have a shorter leg length than men.1 1 7 There is also research that suggests that at slower speeds, females use a longer stride length than males (normalized to body height),5 7 while at faster speeds, males use a longer normalized stride length than females.85,8 9 The potential differences in normalized stride length between females and males at different walking speeds would suggest that the ratio of step length to leg length varies between genders. If at the slow speed, females use a longer relative stride length than males, then the model of Ekkebus and Killey27,28 would predict a higher utilized COF for females (Figure 2-4). Similarly, if at fast speeds males use a longer relative stride length than females, then the model would predict that males would have a higher peak utilized COF than females. To date, the influence of age and gender on utilized COF generated while walking at different speeds has not been reported. The purpose of this study was threefold: 1) to quantify age-specific and 2) gender-specific changes in peak utilized COF during walking at different speeds; and 3) to identify the relationship between normalized stride length and peak utilized COF. It was hypothesized that 1) younger adults would generate higher peak utilized COF than older adults; 2) at slower speeds, females would generate a higher peak utilized COF than males, while at fast speeds, males would generate a higher peak utilized COF than females; and 3) normalized stride length would be a predictor of peak utilized COF. Such Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 2 information is quite useful for the development of empirically derived standards for walkway slip resistance. METHODS Subjects Sixty healthy adults between the ages of 23 and 79 participated in this study. Subjects were divided into three groups: Young (20-39 y.o.); Middle-aged (40-59 y.o.); and Senior (60-79 y.o.). Each group consisted of 10 males and 10 females (Table 3-1). Table 3-1 Physical Characteristics of Subject Participants Means (Standard Deviations) Age Group Gender Age (yrs) Leg Length (cm) Height (cm) Mass (kg) Young1 Females (n=10) 28.2 (4.8) 87.2 (3.0) 167.1 (6.5) 60.3 (5.9) Males (n=10) 28.5 (4.6) 90.8 (3.4) 177.0 (5.5) 81.5(11.7) Middle1 Females (n=10) 45.9 (5.2) 88.5 (4.3) 160.9 (12.5) 66.7(10.4) Males (n=10) 47.0 (5.5) 95.6 (6.8) 180.8(6.7) 85.0(12.8) Senior2 Females (n=10) 69.4 (5.3) 85.6 (5.2) 158.9 (5.1) 60.6(11.5) Males (n=10) 71.4 (5.4) 90.3 (5.3) 169.6 (7.1) 79.6(13.3) Total1 Females (n-30) 47.8 (17.9) 87.1 (4.3) 162.3 (9.1) 62.5 (9.7) Males (n=30) 49.0(18.6) 92.2 (5.7) 175.8 (7.8) 82.0 (12.4) Mass, height, and leg length significantly greater for males than females (p < 0.05). 2 Mass and height significantly greater for males than females (p < 0.05). Subjects were recruited from the student and faculty population at the University of Southern California (Los Angeles, CA), as well as by word of mouth in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 3 the local Los Angeles area. Only persons who were capable of independent ambulation without assistive devices were included. Subjects were excluded if they had a known history of neurologic disease or a lower extremity orthopedic condition that would interfere with walking. This was determined through a medical interview. Prior to participation, each subject was fully informed of the nature of the study, and signed a human subjects consent form approved by the Institutional Review Board of the University of Southern California Health Sciences Campus. Instrumentation Ground reaction forces (vertical, fore-aft, and medial-lateral) were recorded using three AMTI force plates (Model OR6-6-1, AMTI Corp., Newton, MA), covered with smooth vinyl composition tile. These force plates were aligned in series and camouflaged within a 10-meter walkway. Force plate data were sampled at 600 Hz, and recorded on a 300 MHz personal computer using a 64-channel analog-to-digital converter. A VICON motion analysis system (Oxford Metrics Ltd., Oxford, England) was used to measure stride length. Kinematic data were sampled at 60 Hz and recorded digitally on an IBM 166 MHz personal computer. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 4 Procedures All testing was performed in the Musculoskeletal Biomechanics Research Laboratory at the University of Southern California. Prior to data collection, the length of each subject’s right lower extremity (anterior superior iliac spine to medial malleolus) was measured with a soft tape measure during standing. To measure stride length, a reflective marker (20 mm sphere) was then placed over the right lateral malleolus. Subjects walked in Oxford-style shoes (Iron-Age, Inc., Endwell, New York) that were provided for use during the walking trials. Subjects were instructed to walk at predetermined slow (57 m/min), medium (87 m/min), and fast (132 m/min) walking speeds along the 10-m walkway. Subjects were instructed to look at a spot on the wall at the far end of the walkway to avoid “targeting” a force plate. The middle six meters of the walkway were delineated by photoelectric light switches, which were used to trigger the data acquisition computer. Subjects performed one trial at each walking speed. Walking speed was calculated following each walking trial, and only trials that were within ±5% of the targeted speed, and in which a clean force plate contact occurred (i.e., the right foot contacted one of the three force plates) were accepted. All other trials were repeated. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 5 Data Analysis Force plate data were analyzed using the VICON Workstation and Reporter software programs (Oxford Metrics, Ltd., Oxford, England). Digitally acquired anterior-posterior, medial-lateral, and vertical forces were exported to ASCII file and imported to an Excel spreadsheet. The anterior-posterior and medial-lateral forces were used to calculate the resultant shear force using the following formula: Resultant Force = ^(Anterior-Posterior Force)2 + (Medial-Lateral Force)2 The utilized COF throughout stance was calculated as the ratio of the resultant/vertical forces. The peak utilized COF value during limb loading, resulting from a shear force that would contribute to the foot sliding anteriorly, was identified. Representative force plate and utilized COF curves for a senior female subject walking at the slow speed are presented in Figure 3-1. Data were screened for spuriously high utilized COF values resulting from the division of small shear and vertical forces.1 1 Typically, non-spurious utilized COF values were observed once the reference limb had been substantially loaded (92 N on the average). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 6 a) Vertical 700 i 600 - 500 - 5 C s s 3 400 - I 300 - % « 200 - | 100 - Medial-Lateral -100 - 0.6 0.2 0.4 Horizontal Time (seconds) b) Spurious Peak in COF Peak utilized COF during limb loading 0.2 - 0.2 0.8 0.4 0.6 Time (seconds) Figure 3-1. Representative tracings of ground reaction forces (a) and utilized coefficient of friction (b) during shod walking at a slow speed for a senior female subject. Note that the initial spuriously high spike in the utilized COF was due to a relatively low vertical ground reaction force. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 7 Kinematic data were analyzed using VICON 370 Workstation software (Oxford Metrics, Ltd., Oxford, England). The reflective marker at the lateral malleolus was identified manually, and three-dimensional marker coordinates were calculated. Stride length was calculated as the horizontal distance, in the direction of progression, of the right lateral malleolus marker from right heel contact to the next right heel contact. Normalized stride length was calculated by dividing each subject’s stride length by his/her measured leg length and expressing it as a percentage of leg length. Statistical Analysis To determine if utilized COF varied between genders and across the three age groups, separate two by three analyses of variance (ANOVA) were performed at each of the walking speeds (slow, medium, and fast). A similar analysis was performed for normalized stride length. For each of the two-way ANOVAs performed, if a significant interaction was found, then the main effects were considered separately through post-hoc testing. To determine if normalized stride length could be used to predict utilized COF, linear regression analysis was performed. All utilized COF values recorded from each subject at each speed were used in this analysis. Statistical analyses were performed using SPSS statistical software (version 10.0; SPSS Inc., Chicago, IL). A significance level ofp< 0.05 was used for all statistical comparisons. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 8 RESULTS Peak Utilized COF The average peak utilized COF values generated by all 60 subjects at slow, medium and fast walking speeds were /u = 0.22, n = 0.24, and jj. = 0.26, respectively (Table 3-2). The highest value recorded for a single subject, // = 0.44, occurred during a fast walking trial. The lowest value recorded for a single subject, p. = 0.13, also occurred during a fast walking trial. Table 3-2 Peak utilized COF during Walking at Slow, Medium, and Fast Speeds SLOW MEDIUM FAST Mean (SD) Range Mean (SD) Range Mean (SD) Range Young Females (n=10) .24 (.05) .20-.35 .24 (.02) .21-.28 .25 (.04) .21-.32 Males (n=10) .19 (.04) .14-.30 .21 (.02) .18-.24 .27 (.03) .23-.31 Middle Females (n=10) .24 (.04) .16-.28 .27 (.02) .23-.31 .26 (.05) .18-.34 Males (n=10) .22 (.05) .17-.33 .26 (.06) .20-.39 .32 (.09) .22-.44 Senior Females (n=10) .23 (-04) .14-.30 .22 (.03) .18-.26 .22 (.06) .13-.30 Males (n=10) .19 (.02) .17-.22 .22 (.04) .17-.36 .24 (.06) .11-31 Totals by Gender Females (n=30) .24 (.04) .14-.35 .24 (.03) .18-.31 .24 (.05) .13-.34 Males (n=30) .20 (.04) .14-.33 .23 (.05) .17-.39 .28 (.07) .17-.44 Overall Total All Subjects (n=60) .22 (.04) .14-.35 .24 (.04) .17-.39 .26 (.06) .13-.44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 9 When averaged between genders, the peak utilized COF varied with age at both the medium (p = 0.001) and fast (p = 0.005) walking speeds. At the medium speed, post hoc analysis revealed that the middle-aged subjects generated significantly higher peak utilized COF than both the young (p = 0.26 vs. p = 0.22; p = 0.001) and senior subjects (p = 0.26 vs. p - 0.22; p - 0.002; Figure 3-2). At the fast speed, post hoc analysis revealed that the middle-aged subjects generated significantly higher peak utilized COF than the senior subjects (p = 0.29 vs. p = 0.23; p = 0.018; Figure 3-2). Peak utilized COF at the slow speed did not vary across age groups. When averaged across age groups, the peak utilized COF varied between genders. During slow walking, females generated significantly higher peak utilized COF values than males (p = 0.24 vs. p = 0.20; p = 0.002; Figure 3-2). In contrast, during fast walking, males generated significantly higher peak utilized COF values than females (p = 0.28 vs. p = 0.24; p = 0.023; Figure 3-2). No difference in peak utilized COF between females and males at the medium speed was observed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 0 -■ -M a le Female 0.35 O u ■ o a > 0.25 P « 4 > * 0.2 0.15 DC DC DC DC SLOW MEDIUM FAST Figure 3-2. Between gender and across age group differences in average peak utilized COF during shod walking at slow, medium, and fast speeds. * = Collapsed across age groups, the average peak utilized COF greater for females than males at the slow walking speed ip = 0.002). * = Collapsed across age groups, the average peak utilized COF greater for males than females at the fast walking speed ip - 0.023). * = Collapsed between genders, the average peak utilized COF greater for middle-aged subjects compared to both young ip = 0.001) and senior subjects ip = 0.002) at the medium speed. § = Collapsed between genders, the average peak utilized COF greater for middle-aged subjects compared to senior subjects ip = 0.018) at the fast speed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 Normalized Stride Length When averaged between genders, normalized stride length did not vary significantly among the young, middle-aged and senior groups at either the slow, medium, or fast speeds (Figure 3-3). When averaged across age groups, normalized stride length did not vary significantly between females and males at the slow, medium, or fast speeds (Figure 3-3). -M ale —A— Fem ale o n a J » < u •J=j = o 210 190 DC g 170 hJ DC < u w 150 130 FAST SLOW MEDIUM Figure 3-3. Average normalized stride length (stride length/leg length x 100) during shod walking at slow, medium, and fast speeds. No significant differences were observed between male and female subjects or across the age groups. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 2 Normalized stride length was found to be a significant predictor of peak utilized COF (r = 0.423; p < 0.001; Figure 3-4). However, only 18% of the variance in peak utilized COF could be explained by normalized stride length (R2 = 0.179). .45 .40 .35 O -30 U § .25 P h .20 .15 .10 120 140 160 180 200 220 240 260 Normalized Stride Length (% Leg Length) Figure 3-4. Relationship between normalized stride length (stride length/leg length x 100) and peak utilized COF across all walking speeds for all 60 subjects (n = 180 data points; r = 0.423; R2- 0.18, < 0.001) □ □ 1 X 1 D JB ■ t " Dn □ “ V t H Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 3 DISCUSSION Age or gender related differences in utilized COF were recorded at each of the three walking speeds. Our initial hypothesis concerning age-related changes in peak utilized COF was partially accepted as the middle-aged group had higher utilized COF than the senior group at both the medium and fast walking speeds. However, there were no differences in utilized COF when the young group was compared to the senior group at any of the speeds, nor was a significant difference identified when the young group was compared to the middle-aged group at the slow speed. Further, the cause of the difference in utilized COF between the middle-aged and senior subjects at the medium and fast speeds could not be explained by normalized stride length as no age-related differences in normalized stride length were observed. With respect to gender, our initial hypothesis was shown to be correct as females had a higher utilized COF during slow walking and men had a higher utilized COF during fast walking. As with the age-related differences however, the cause of gender-related differences also could not be explained by normalized stride length as no gender differences were observed. Normalized stride length was found to be a significant predictor of utilized COF, with longer normalized stride lengths being correlated with greater utilized COF. However, it should be noted that only 18% of the variance in utilized COF could be explained by changes in normalized stride length. This finding suggests Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 4 that factors other than normalized stride length likely contribute to variations in utilized COF during walking. For example, many physical attributes can influence the mechanics of limb loading such as lower extremity strength, the ability to control the center of mass during weight acceptance, lower extremity joint flexibility, and proprioception (particularly at the knee and ankle). Given the complexity of gait and the neuromuscular system, it is not entirely surprising that only a small portion of utilized COF could be explained by the simple geometric relationship suggested by Ekkebus and Killey.27,2 8 Further research is necessary to determine the degree to which these factors influence utilized COF during walking. The average utilized COF recorded for our subjects while walking at slow (// = 0.22) and medium (// = 0.24) speeds were similar to values reported by Skiba1 0 9 (ju = 0.21- 0.23; velocity = 60 to 90 m/min) and Perkins9 6 (ju = 0.22; velocity not reported). Likewise, the average utilized COF recorded for our young male subjects while walking at a fast speed was identical to the ju = 0.27 interpolated (based on a walking speed of 132 m/min) from data presented for a 19 year old male.3 2 In contrast to these similarities, our data differed from values reported by Kulakowski and colleagues6 1 and Buczek et al.1 2 The utilized COF reported by Kulakowski and colleagues6 1 were greater than the values recorded for our subjects during both slow (ju = 0.29 vs. fi = 0.22) and fast {/u = 0.33 vs. // = 0.26) walking, however the apparent trend towards increasing peak utilized COF with higher speeds was similar between studies. Similarly, Buczek and colleagues1 2 reported a higher Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 5 utilized COF value for five young subjects during level walking (// = 0.31 for combined slow and fast walking speeds). Reasons for differences between values recorded in our study and those reported by Kulakowski and colleagues6 1 and Buczek et al.1 2 likely include differences in footwear, floor characteristics, as well as the limited number of able-bodied subjects studied in the other two studies (n = 5 each). Finally, in the current study, a wide range of utilized COF values were recorded within each gender and age group. Collapsed across all subjects and speeds, utilized COF values ranged from p - 0.13 to p - 0.44. Collectively, these data suggest that despite the presence of relatively low mean utilized COF values across the three walking speeds (p = 0.22 to 0.26), wide inter-subject variability exists. As a result of this variability, care must be taken when attributing a specific utilized COF value to a given gender or age group. Further, this variability will likely be important when considering the appropriateness of thresholds used for defining safe flooring. Current recommendations for safe flooring for persons without a disability incorporate a static COF threshold of p > 0.50 (as measured with the James machine). CONCLUSIONS While age and gender related differences in utilized COF exist across walking speeds, these differences could not be attributed solely to the selected Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 6 anthropometric and stride characteristic variables evaluated in this study. The evaluation of the relationship between normalized stride length and utilized COF in the current study revealed that only 18% of the variability in utilized COF could be explained by normalized stride length. Further, while selected differences in utilized COF between senior and middle-aged subjects were identified, the anticipated differences in utilized COF between young and senior subjects did not emerge as predicted. Collectively, these findings suggest that factors, other than age and the selected anthropometric variables considered in this study, likely play a large role in determining utilized COF. These factors may include lower extremity strength, proprioception, and range of motion. Additionally, the wide inter-subject variability in utilized COF demonstrated in this study suggests that minimum threshold levels used to define “safe” walkway surfaces should consider not only average utilized COF, but also the range of values used by individual subjects. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 CHAPTER IV THE ROLE OF CENTER OF MASS KINEMATICS IN PREDICTING PEAK UTILIZED COEFFICIENT OF FRICTION DURING WALKING During walking, slips are likely to occur when an individual’s utilized coefficient (COFu) exceeds the friction available on the floor surface. Although the data presented in Chapter III indicated that COFu values vary across gender and age groups, the ability to identify predictors of individual friction needs, and thus slip initiation risk, remains limited. The purpose of this study was to determine the extent to which center of mass (CM) kinematics could be used to predict peak utilized coefficient of friction (COFu) during level walking. Ground reaction forces (1200 Hz) and full-body kinematic (120 Hz) data were recorded simultaneously as subjects traversed a 10-meter walkway. Stepwise regression analysis was performed to determine whether selected CM variables (i.e., position, velocity) could predict peak COFu during weight acceptance. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 8 INTRODUCTION Slips have been identified as a leading cause of falls and injuries in the 6 62 5 23 62 home ’ and work ’ ’ environments. During walking, forces generated by the body are transmitted through the foot to the floor. In order to prevent a slip, sufficient friction at the foot-floor interface is required. When the shear forces applied to the floor surface exceed the available friction at the foot-floor interface, a slip becomes imminent.4 6 In the research setting, an individual’s friction needs during walking, or utilized coefficient of friction (COFu), is calculated from force plate recordings of ground reaction forces (Figure 2-2a).1 1 The COFu is defined as the ratio between the shear (resultant of the fore-aft and medial-lateral forces) and vertical components of the ground reaction force recorded as a person walks across a dry, non-contaminated surface (Figure 2-2b).1 0 5 Theoretically, persons or groups of individuals who demonstrate higher COFu are at the greatest risk of slipping 4 6 Currently, the ability to identify predictors of peak COFu, and thus risk of slip initiation, remains limited. Two factors that have been proposed to increase COFu during weight acceptance are the angle of impact of the leg with the ground2 7 , 2 8 and walking speed.13,3 1 ,1 0 9 Ekkebus and Killey 2 1 ’2 8 proposed that the impact angle of the leg with the ground, calculated from measures of step and leg length, could be used to predict peak COFu. In their model, the leg was equated to a rigid strut that transmitted Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 9 forces to the ground during walking. The authors theorized that the tangent of the angle formed by the leg at heel contact (using the hip joint center as a reference point) would be equal to the ratio of shear to vertical forces, and thus be predictive of peak COFu at weight acceptance (Figure 2-3). According to this model, greater impact angles would result in higher peak COFu (Figure 2-4). Subsequent investigations, however, have revealed that while the impact angle of the leg does serve as a significant predictor of friction needs, it accounts for only 18 to 27% of the variance in peak COFu.13,9 9 Additionally, the Ekkebus and Killey model has been shown to overestimate peak COFu by 86% at slow walking speeds and up to 131% at fast walking speeds.9 9 Cavagna and Margaria1 5 proposed that greater angles of inclination of the line connecting the whole body center of mass (CM) to the center of pressure (CP) would contribute to higher anterior shear forces during walking. These authors suggested that greater angles of inclination between the CM and CP resulting from faster walking speeds, could be associated with greater anterior shear forces during weight acceptance. Unfortunately, the authors did not calculate peak COFu, so it is unknown what impact, if any, the CM to CP angle would have on the prediction of peak COFu. Walking speed also has been theorized to influence COFu, however, the published literature relating walking speed to COFu is conflicting. While a number i 'X'y 1 no of authors have found that peak COFu increases with faster walking speeds, ’ ’ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 0 other have reported that walking speed is not related to peak COFu.1 2 ’16,1 1 6 Small sample sizes, non-homogeneous populations, and differences in study design, may explain, in part, the differences found across studies. Walking speed in the above noted studies was calculated as the average velocity over a fixed distance, however, it is possible that the instantaneous velocity at the time of peak COFu may be a stronger predictor of peak COFu. For example, Pai and Iqbal9 4 modeled the relationship between the instantaneous velocity of the CM in the anterior direction, the location of the CM relative to the base of support, and the ability to maintain dynamic stability. The authors reported that with increasingly posterior locations of the CM relative to the foot, greater instantaneous CM velocities in the anterior direction would be required to maintain stability. Based on the findings of Cavagna and Margaria,1 5 and Pai and Iqbal9 4 subtle variations in the relative location as well as velocity of the total body CM would be expected to influence the forces applied to the floor and therefore slip potential. Although it is conceivable that measures describing the relative location of the total body center of mass as well as the anterior velocity during walking could serve as better predictors of COFu, there are no data to support this premise. The purpose of this study was to determine the extent to which the CM to CP angle of inclination and the CM velocity could be used to predict peak COFu during the weight acceptance phase of level walking. It was hypothesized that both greater angles of inclination of the CM relative to the CP as well as a faster anterior CM velocity Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 would contribute to higher peak COFu values. Once the relationship between CM kinematics and peak COFu has been established, then it may be possible to identify individual anthropometric and gait characteristics that combine to increase the risk of slip onset. METHODS Subjects Forty-nine persons (28 male, 21 female) between the ages of 22 and 40, participated in this study (Table 4-1). Subjects were recruited from the student population at the University of Southern California (Los Angeles, California), as well as by word of mouth in the local Los Angeles area. Prior to participation, each subject was fully informed of the nature of the study, and signed an informed consent form approved by the Institutional Review Board of the University of Southern California Health Sciences Campus. Only subjects capable of independent ambulation without assistive devices were included. Subjects with known neurologic or orthopedic conditions that would interfere with gait were excluded from the study. Table 4-1 Subject Characteristics Means (Standard Deviations) Male (n - 28) Female (n = 21) Combined (n = 49) Age (years) 27.4 (4.4) 25.1 (2.8) 26.4 (3.9) Height (cm) 180.1 (7.4) 165.7 (8.1) 173.9(10.5) Weight (kg) 86.3 (13.3) 67.7(18.0) 78.3 (17.9) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 2 Instrumentation All walking trials were conducted on a 10-meter walkway with the middle 6 meters designated for data collection. Light sensitive triggers were used to initiate and terminate data collection as subjects traversed the length of the walkway. Three- dimensional motion analysis was performed using a six-camera motion analysis system (VICON, Oxford Metrics Ltd., Oxford, England). Kinematic data were sampled at 120 Hz and recorded digitally on a Pentium III 1GHz personal computer. Reflective markers (20 mm spheres) placed over specific anatomical locations (see below) were used to calculate the total body center of mass location. Ground reaction forces (vertical, anterior-posterior, and medial-lateral) were recorded using a single AMTI force plate (Model OR6-6-1, AMTI Corp., Newton, MA). The force plate was covered with smooth vinyl composition tile in order to camouflage its location within the walkway. Force plate data were sampled at 1200 Hz, and recorded on a Pentium III 1GHz personal computer using a 64-channel analog-to-digital converter. Procedures Each subject was provided with a pair of walking shoes (Rockport World Tour, model M W WT18; The Rockport Company, LLC, Ronks, PA) for use during testing. To allow placement of the markers directly on the skin of the trunk and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 extremities, male subjects wore only sports shorts, while female subjects wore sports shorts and a sports bra. All testing was performed within the Musculoskeletal Biomechanics Research Laboratory at the University of Southern California. To estimate full-body CM location, selected anthropometric measures were obtained including: subject height and mass, bilateral leg length, knee width, ankle width, shoulder offset (vertical distance from base of acromion marker to level of clavicle origin), elbow width, wrist width, and hand thickness (distance between dorsal and palmer surface of hand at level of third metacarpal). Thirty-seven reflective markers were then taped to the following body landmarks: stemo-clavicular notch, xyphoid process, C7 and T10 spinous processes, right mid-scapula, as well as bilaterally over the acromio-clavicular joint, lateral humerus, lateral humeral epicondyle, radial and ulnar styloids, dorsal surface of the third metacarpal, anterior and posterior superior iliac spines, lateral thigh, lateral femoral epicondyle, lateral tibia, lateral malleolus, 2n d metatarsal head, and posterior calcaneous. A head band was used to secure four markers bilaterally over the temple and posterior cranium, forming a plane horizontal with the floor when the subject looked forward. A static calibration trial was recorded to define marker relationships necessary for subsequent kinematic modeling. During the static trial, subjects stood stationary in the center of the calibration field for five seconds. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 4 Next, subjects were allowed several practice trials to accommodate to the markers and head band. All subjects were instructed to walk at a comfortable speed, with kinematic and force plate data being recorded simultaneously. A trial was considered successful if the subject’s right foot landed within the force plate and all 37 markers were visible throughout stance on the force plate. Data Analysis Reflective markers were identified manually using VICON 370 Workstation software (Oxford Metrics, Ltd., Oxford, England) and then automatically digitized. Kinematic data were filtered using Woltring’s general cross-validatory quintic spline routine (predicted MSE = 20). A fifteen segment model (VICON Plug-in-Gait; Oxford Metrics, Ltd., Oxford, England), consisting of six lower extremity links, six upper extremity links, two links for the trunk and one for the head, was used to estimate the location of the total body CM (x, medial-lateral; y, anterior-posterior; and z, vertical coordinates). Masses of each segment were calculated as a proportion of the total body mass using anthropometric relationships reported by Dempster2 4 as well as subject specific anthropometric measures recorded by the investigator. The weighted sum of the CM of each of the fifteen individual segments was then used to compute the 3-D location of the body’s CM. Center of mass displacement data and ground reaction forces were imported into DataPac software (v. 2K2, Run Technologies, Mission Viejo, CA) for further analysis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 5 The velocity of the CM in the anterior-posterior direction (CMye i ap) was calculated in DataPac using the first derivative of the CM y-displacement data. Digitally acquired vertical, anterior-posterior, and medial-lateral forces were filtered compensation.2 2 The center of pressure coordinates (x, y) were calculated in Data Pac using standard formulas provided by the force plate manufacturer. The COFu was calculated by dividing the resultant shear force (computed from the anterior-posterior, F a p, and medial-lateral, F Ml, forces) by the vertical force (Fvert) (Equation 4-1). During weight acceptance, the peak COFu value resulting from a shear force that would contribute to the foot sliding anteriorly, was identified. To avoid spurious COFu values resulting from division by small numbers, the data were screened and only COFu data in which the vertical GRF's exceeded 50N were reaction force exceeded 5N, and the cessation of stance was defined as the time at which the vertical ground reaction force fell below a 5N threshold. in DataPac using a 4 order, 45 Hz, low pass Butterworth filter with a zero-lag COFu = Fvert (4-1) analyzed.1 1 The onset of stance was defined as the time at which the vertical ground Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 6 The CM-CPAngie was defined as the angle formed by the following three points: CP, CM, and the vertical projection of the CM onto the ground surface (Figure 4-1). The position of the CM relative to the CP at the time of peak COFu was calculated in the anterior-posterior (C M a p - CPap), medial-lateral (C M Ml - C P m l), and vertical (CMvert) directions (Figure 4-1). The CM-CPAngie, was referenced to vertical, and was calculated using Equation 4-2: CM-CP,,,,t - Arc Tan ~ C lW + (CM" ,4-2) CMyert Vertical (Z+) CM >Ant (Y+) Med (X+) CP Figure 4-1. Center of mass (CM) to center of pressure (CP) relationships studied. Abbreviations: CM ap - CPap, absolute distance between center of mass and center of pressure in the anterior-posterior direction (y-coordinate); C M m l - C Pm l, absolute distance between center of mass and center of pressure in the medial-lateral direction (x-coordinate); CMyert, absolute distance between center of mass and center of pressure in vertical direction (z-coordinate); CM-CP A n g ie , angle of center of mass to center of pressure (expressed relative to vertical). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 7 Statistical Analysis Descriptive statistics (mean, standard deviation, minimum and maximum) were calculated for all variables assessed. To determine whether CM-CP A n g ie and CMvei ap were predictive of the peak COFu during weight acceptance, stepwise regression analysis using a forward stepwise procedure was performed (F to enter = 0.05; F to remove = 0.10). Independent variables at the instant of peak COFu during weight acceptance included the relative position of the CM to the CP (CM-CPAngie), as well as the velocity of the CM in the anterior-posterior direction (CMvei a p)- Peak COFu was the dependent variable. All statistics were calculated using the SPSS 10.0 statistical software (SPSS Inc., Chicago, IL). RESULTS On the average, self-selected walking velocity of the participants was 1.6 ± 0.2 m/s (range, 1.2 to 2.0 m/s). The mean peak COFu was n = 0.23 ± .03 (range, /u = 0.14 to ju = 0.34), and occurred 90 msec following initial contact (range, 33 to 143 ms). At the instant of peak COFu, the CM was always located posterior (mean = .227 m), medial (mean = .052 m) and superior (mean = .970 m) to the CP. The CM velocity at the time of peak COFu averaged 1.6 m/s in the anterior direction. The average CM-CP A n g ie at the time of peak COFu was 13.6 degrees (Table 4-2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 8 Table 4-2 Center of Mass Kinematic Variables at Time of Peak Utilized Coefficient of Friction during Weight Acceptance (n=49) Kinematic Variables Mean SD Minimum Maximum CMyei ap (m/s) 1.6 0.2 1.3 2.0 CM-CP Angle (° ) 13.6 2.3 9.4 18.9 Abbreviations: CMyei a p, velocity of center of mass in the anterior-posterior direction (+, anterior); CM-CP A n gie, angle of center of mass to center of pressure (expressed relative to vertical). Relation between CM Kinematic Variables and Peak COFu The CM-CP A n gie at the instant of peak COFu was the best predictor of peak COFu during weight acceptance (r = 0.750; p < 0.001), accounting for 56% of the variance. Greater CM-CPAngies were associated with higher peak COFu (Figure 4-2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 9 .40 # * ■ ( o * £ 3 .2 .30 O © o U "g .20 .£ 3 D 3 C u .10 8 16 10 12 18 20 14 Angle CM-CP Relative to Vertical (Degrees) Figure 4-2. Scatter plot showing the relationship between CM-CP A n g ie and peak utilized coefficient of friction (r = 0.750; R2 = 0.563; p < 0.001). The CMyei ap also was a significant predictor of peak COFu (r = 0.591 \P< 0.001) with faster CMye i ap being associated with higher peak COFu (Figure 4-3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 0 .40 ( 3 O > r- <tf o U ■ g -20 N 5 IS .10 1.2 1.4 1.6 1.8 2.0 2.2 CM Anterior Velocity (m/see) Figure 4-3. Scatter plot showing the relationship between CMvei ap and peak utilized coefficient of friction (r = 0.591; R2= 0.349; p < 0.001). The combination of the subject’s CM-CP Angie and the CMvei ap improved the prediction of peak COFu, and together explained 62% of the variance (r = 0.784; p < 0.001; Figure 4-4; Table 4-3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 Center of > . Mass ^ CM Velocity Anterior CM-CP Angle Center of Mass CM-CP Angle _ Center of Pressure Center of Pressure Peak COFu = .0292 + .0090 (CM-CP Angle) + .0493 (CM Velocity Anterior) Figure 4-4. Significant predictors of peak utilized coefficient of friction during weight acceptance from sagittal (a) and frontal (b) perspective. The CM-CP Angle (expressed relative to vertical) in combination with the CM velocity in the anterior direction accounted for 62% of variance in peak COFu (r = 0.784; p < 0.001). Table 4-3 Simple and Partial Correlation Coefficients for Center of Mass Kinematic Variables and Peak Utilized Coefficient of Friction Utilized COF Kinematic Variables r r (partial) CM-CP Angle .750* .634 CMyei AP .591* .345 * Correlations significant atp < 0.001 evel. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 2 DISCUSSION Identification of factors that contribute to high COFu during walking is essential for identification of individuals and/or situation-specific circumstances which present the greatest risk for slip onset. Consistent with the initial hypothesis of this study, both the CM-CPAngie and CMye i ap were significant predictors of peak COFu. The CM-CP A n g ie was the best predictor of peak COFu, with larger CM- CP Angies being associated with greater peak COFu. The finding that peak COFu increased with greater CM-CP A n g le s is consistent with the basic premise proposed by 0 1 OR Ekkebus and Killey. ’ These authors proposed that increases in the initial contact impact angle of the limb with the ground would be associated with greater peak COFu. The Ekkebus and Killey model,2 7 ’2 8 however, was based on assumptions that likely contributed to the inability to accurately predict peak COFu- For example, these authors viewed the leg as a rigid strut which would transmit forces to the ground with a ratio of shear to vertical forces proportional to the angle of impact. In actuality, however, the leg is not a rigid strut as motion98,1 0 7 at the knee and ankle occurs between the time of initial contact and the time of peak COFu.97’1 0 3 In addition, the Ekkebus and Killey model2 7’2 8 used the hip joint center and foot to define the angle of inclination, while the current study used the alignment of the CM to the CP at the instant of peak COFu- In the current model, the CM-CP A n g ie Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 accounted for 56% of the variance in peak COFu during walking compared to the model proposed by Ekkebus and Killey which accounted for less than 30% of the 13 99 variance. ’ Instantaneous velocity of the CM in the anterior direction (CMvei ap) also was a significant predictor of peak COFu, with greater CMV ei a p being associated with higher peak COFu- While the CMvei ap represented the instantaneous velocity of the CM at the instant of peak COFu, post-hoc analysis revealed a strong correlation to , each subject’s average walking speed (r = 0.942; p < 0.001). The finding of faster CMvei ap being associated with higher peak COFu is consistent with a number of previously reported average over-ground velocity-COFu i ^ 'X 'y 1 no relationships, ’ ’ however, others have not found this relationship to be i 'j i f\ 1 1 (\ 'X 'y consistent. ’ ’ Fendley and colleagues noted a fair positive correlation (r = 0.41; p < 0.001) between walking speed and peak COFu in a single male subject walking at speeds ranging from approximately 102 to 160 m/min. Similarly, Skiba1 0 9 reported that peak COFu increased with walking speed, however, specific subject data were not provided. Bumfield and Powers1 3 reported that average peak COFu increased from p = 0.22 at the slow walking speed (57 m/min) to p = 0.26 at a fast walking speed (132 m/min) in their study of 60 subjects between the ages of 23 and 79 years. However, the pattern of increase was not consistent between genders (Figure 3-2). In males, the average peak COFu increased from p = 0.20 at the slow speed to p = 0.28 at the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 4 fast speed (Table 3-2). In females, the average peak COFu (ju - 0.24) did not change across the three walking speeds (Table 3-2). Tisserand1 1 6 reported that peak COFu during weight acceptance (range, // = 0.08 to// = 0.15) did not demonstrate a significant relationship to walking speed in a study of three subjects walking at speeds ranging from 90 to 165 m/min. As the age and gender of this limited number of subjects was not reported, it is difficult to compare these results to other studies. Three mechanisms can be used to increase walking velocity, lengthening stride, increasing cadence, or a combination of both.2,85,88,98,9 9 Increasing velocity by lengthening stride, would likely lead to a greater CM-CPAngie as the CP would be farther anterior to the CM. Based on the results of the current study, a greater CM- CP A n g ie would contribute to a higher COFu than would have resulted from the CMV ei ap increase alone. Such findings suggest that individuals who increase stride length to achieve a faster walking velocity may experience greater increases in peak COFu compared to those who increase cadence to achieve a faster walking speed. The findings of the current investigation also suggest that person-specific anthropometric characteristics may increase an individual’s risk for slip initiation. Individuals with shorter legs who take similar length steps as individuals with longer legs could conceivably have increased CM-CP A n gies, due to the lowered height of the CM. Such results suggest that individuals with shorter legs who take longer steps may experience higher peak COFu during weight acceptance than individuals with longer legs taking a similar step length. In addition, significant trunk leans which Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 5 alter the CM location relative to the CP may also contribute to differences in peak COFu across individuals. Further research is needed to test these assumptions. CONCLUSIONS In summary, both CM-CP Angie and CMyei ap served as predictors of peak COFu during weight acceptance. The identified relationships between CM kinematics and peak COFu provide insight into how gait and individual anthropometric characteristics may increase risk for slip initiation in certain individuals. However, only 62% of the variance in peak COFu could be explained by the model, suggesting further investigation is required to identify other factors that influence COFu values. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 6 CHAPTER V PREDICTION OF SLIPS: AN EVALUATION OF UTILIZED COEFFICIENT OF FRICTION AND AVAILABLE SLIP RESISTANCE It has been reported that slips occur when an individual’s peak utilized coefficient of friction (COFu) exceeds the slip resistance provided by the walking surface. However, relationships between COFu, slip resistance, and the probability of a slip event occurring have only been established using a device that measures dynamic coefficient of friction. This chapter describes an investigation that was conducted to determine the relationship between static measures of slip resistance (measured using a variable incidence tribometer) and peak COFu (as measured from a force plate) on the probability of a slip occurring during level walking in fifty-two young adults. Video, kinematic, and ground reaction force data were recorded simultaneously as subjects traversed a walkway during conditions of normal and reduced floor surface slip resistance. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 7 INTRODUCTION r s o Slipping is one of the most common causes of falls, ’ and is of major concern to industry and society due to the associated human suffering and financial costs.29,6 3 In the work environment, slipping has been identified as the primary antecedent event to falls on both stairs and level surfaces,5 and reportedly accounts for approximately 62% of underfoot accidents.6 8 According to the U.S. Bureau of Labor Statistics (1996), slips, trips and falls accounted for 20% of all nonfatal occupational injuries and 11% of fatal occupational injuries. In the United States, the annual costs associated with occupational injuries due to slips and falls are expected to exceed $85 billion during the year 2020 when it is projected that more 90 that 17 million falls resulting in injury will occur. During walking, a slip is likely to occur when the utilized coefficient of friction of an individual exceeds the available friction at the foot-floor interface. Force plate measurements of ground reaction forces can be used to calculate an individual’s utilized coefficient of friction (COFU ; Figure 5-1), while a device called a tribometer is typically employed to measure the slip resistance of the floor surface. Upwards of 50 different portable tribometers have been described in the literature.9 , 1 0 ,2 0 ,2 5 ,2 6,30,34,37,41,42,53,61, 73,74,76,78,79,84,97,102,108,112,118 j n g e n e r a ] t r i b o m e t e r s can be classified as measuring dynamic friction (reflected by the force required to keep a sliding object in motion) or static friction (reflected by the force required to initiate motion between an object and the surface on which it is resting). Currently, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 8 no universal method of measuring slip resistance has been established. Additionally, different types of tribometers yield different measurements, that do not necessarily correlate.8 ’3 6 ’4 1 ’7 8 ’1 0 0 Medial - Lateral Resultant Shear ■► Anterior - Posterior Resultant Shear Vertical Vertical Figure 5-1. The ratio of shear to vertical ground reaction forces can be used to calculate the utilized coefficient of friction (COFu) during walking. To date, only two studies have evaluated the relationship between COFu and available friction in predicting the probability of a slip event.4 6 ’6 1 Hanson and colleagues4 6 reported that as the difference between COFu (measured with force plates) and the available dynamic coefficient of friction increased, the probability of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 9 a slip and/or fall increased. The utilized coefficient of friction (COFu) was manipulated by adjusting the inclination (0,10,20 degrees) of the ramp on which subjects walked, and the application of a soapy water solution to the ramp altered the available slip resistance. The authors reported that when the available dynamic COF exceeded the COFu by 0.52, the probability of a slip occurring was only 1%. However, when the available dynamic COF exceeded the COFu by 0.16, the probability of slipping increased to 50%. Although this study provided valuable insight into the differences between available and utilized friction, this relationship was established on only five subjects for a tribometer that measured dynamic friction. The relationship between COFu and static measures of floor slip resistance was reported by Kulakowski et al.6 1 Recordings of ground reaction forces were used to calculate peak COFu values for five subjects walking at a fast speed across a dry surface. Subjects then walked across three different surfaces on which a film of water and detergent had been applied. The static available COF of the floor surface was estimated based on previous portable articulating strut tribometer measurements of similar surface and contaminant combinations. The authors reported that slips occurred in 79% of the instances when the estimated COFu (projected based on dry walk values) exceeded the estimated available COF (based on previous portable articulating strut tribometer measurements). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 0 In the study by Kulakowski et al.,6 1 the ability to predict slips based on the relationship between static COF measures of floor slip resistance and an individual’s COFu is limited for a few reasons. The tribometer used to assess the available friction of the floor surface has been approved for testing on dry surfaces only (ASTM FI 678), thus, the validity of the wet surface calculations is of concern. Additionally, the available floor surface slip resistance was based on estimates from previous applications of the contaminant. The authors6 1 noted that variations in the thickness of the water film likely altered the actual available slip resistance. Also, as subjects’ estimated COFu (based on dry walking trials) frequently exceeded the available COF by 200-400%, it is difficult to assess the preciseness of the relationship between available COF and COFu- Finally, as subjects were not masked to the slip trial, and subjects participated in multiple slip trials, it is unclear whether subjects altered their gait, and thus their true COFu, in anticipation of the slip event.1 7 Currently, several nation-wide standards encompass use of tribometers that measure static COF (ASTM D2047; ASTM 489; ASTM F1677; ASTM F1679; ANSI A1264.2). Given the limitations of the single study evaluating relationships between static COF and COFu,6 1 it is unclear to what extent static COF measures can be used to predict the probability of a slip occurring. The purpose of this study was to investigate the relationship between static measures of slip resistance (measured Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 using a variable incidence tribometer) and peak COFu (as measured from a force plate) on the probability of a slip occurring during level walking. METHODS Subjects Fifty-two persons (28 male, 24 female) between the ages of 22 and 40 years participated in this study (Table 1). Subjects were recruited from the student population at the University of Southern California (Los Angeles, California), as well as by word of mouth in the local Los Angeles area. Table 5-1 Physical Characteristics of Study Participants Means (Standard Deviations) Male (n - 28) Female (n = 24) Combined (n = 52) Age (years) 27.4 (4.4) 25.0 (2.7) 26.3 (3.9) Height (m) 1.80 (.07) 1.65 (.08) 1.73 (.11) Mass (kg) 86.3 (13.3) 63.3 (9.3) 75.7 (16.4) Prior to participation, each subject was fully informed of the nature of the study, and signed an informed consent form approved by the Institutional Review Board of the University of Southern California Health Sciences Campus. Only subjects capable of independent ambulation without assistive devices were included. After obtaining informed consent, each subject completed a medical history questionnaire to determine if they could safely participate in the study. Subjects with known neurologic or orthopedic conditions that would interfere with gait were Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 2 excluded from the study. Additionally, subjects with previous back injuries, recent fractures, muscle strains, joint sprains, or potentially pregnant were excluded. Instrumentation All walking trials were conducted on a 10-meter walkway with the middle 6 meters designated for data collection. Light sensitive triggers were used to initiate and terminate data collection as subjects traversed the length of the walkway. Three- dimensional motion analysis was performed using a six-camera motion analysis system (VICON, Oxford Metrics Ltd., Oxford, England). Kinematic data were sampled at 120 Hz and recorded digitally on a Pentium III 1GHz personal computer. Reflective markers (20 mm spheres) placed over specific anatomical locations (see below) were used to assist in determining if a slip had occurred on the trial in which a contaminant was present. A video camera (JVC TK-108V, JVC Inc., Yokohama, JAPAN), positioned perpendicular to the walkway, recorded digital images of all walking trials. Ground reaction forces (vertical, anterior-posterior, and medial-lateral) were recorded using four AMTI force plates (Model OR6-6-1, AMTI Corp., Newton, MA) aligned in series. To camouflage the location of the force plate within the walkway, each was covered with smooth vinyl composition tile. Force plate data were sampled at 1200 Hz, and recorded on a Pentium III 1GHz personal computer using a 64-channel analog-to-digital converter. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 The available coefficient of friction on the floor surface was measured using a variable incidence tribometer (English XL, William English, Inc.; Figure 5-2). This tribometer uses a compressed gas source that is regulated to 25 psi (172 kPa) to power a pneumatic cylinder with a 1.25 in (3.2-cm) diameter test foot covered with Neolite® (Smithers Scientific Services, Inc., Akron, OH). The tribometer provides a slip-index value that is based on the angle of incidence of the test foot relative to the test surface at the angle at which the test foot slips. Figure 5-2. The English XL variable incidence tribometer. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 4 During wet surface conditions, the slip index reading of this variable incidence tribometer has been shown to demonstrate a good level of agreement with coefficient of friction values recorded from the force plate (ICC = 0.89).1 0 0 Additionally, an excellent level of agreement (ICC = 0.99) has been documented between successive force plate recordings of COF and the VIT’s slip index values.1 0 0 To ensure safety during testing, subjects wore a fall-arresting body harness (Miller Model 550-64; Franklin, PA) attached via 8 mm climbing rope (Pigeon Mountain Industries, LaFayette, GA) to an overhead trolley which traversed the length of the walkway (Figure 5-3). To control for the influence of footwear on utilized coefficient of friction, each subject was provided with a pair of walking shoes (Rockport World Tour, model M/W WT18; The Rockport Company, LLC, Ronks, PA) for use during testing. Procedures All testing was performed within the Musculoskeletal Biomechanics Research Laboratory at the University of Southern California. Prior to gait testing, subjects were fitted with the adjustable fall arresting harness and shoes (Figure 5-3), and reflective markers were taped to the posterior calcaneous, bilaterally. Next, subjects were fitted with a pair of adjustable swimming goggles to wear during walking trials to limit their ability to see the contaminant on the floor. Additionally, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 5 the lighting was dimmed by approximately 20% to minimize reflections when the contaminant was present. Figure 5-3. The fall arresting harness, attached to an overhead trolley and track, that was worn by all subjects to ensure safety during walking. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 6 Subjects were permitted multiple practice trials to accommodate to the markers, harness system, and goggles. After the accommodation period, subjects were informed that they would be performing multiple walking trials and that they would be asked to leave the room following every two to three walking trials. Subjects were informed that during their absence, the principle investigator (JMB) would either review data from previous trials or, on one occasion, apply a substance which would make the floor slippery. They were not told during which trial the contaminant would be applied, nor the location of the contaminant application. Prior to each trial, subjects were instructed to walk across the room at a comfortable speed with their eyes focused ahead. At the end of each walking trial, each subject was asked if they felt the floor was slippery and their response was recorded. Subjects who perceived that the floor was slippery were then asked a series of standardized follow-up questions including, “Did you feel as though your shoes slipped on the surface? Did you feel as though you needed support from the harness to recover during the slipping motion? Could you have recovered without support from the harness? Could you tell in advance that the floor was going to be slippery on this trial?” For the non-contaminant walking conditions, a trial was considered successful if the subject’s right foot landed within one of the four force plates. Following 2-3 successful non-slip walking trials, subjects were escorted out of the laboratory to an adjacent room. Following a two minute rest, subjects were brought Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 7 back into the laboratory and the previously described testing procedures were repeated. Approximately 10 successful non-slip trials were captured per subject in order to estimate each subject’s peak utilized coefficient of friction based on speed matched trials. After 10 successful dry trials were obtained, a petroleum-based contaminant (WD-40, WD-40 Company, San Diego, CA) was applied to the final in the series of four force platforms. Typically this occurred during the subject’s sixth to eighth departure from the lab. The application of the contaminant involved spraying WD- 40 onto the surface of the platform from a distance of approximately 15 cm, in an “S” pattern over a 3-5 second period. A soft, clean, dry paper towel, folded in a 5 x 8 cm rectangle, was then used to rub the WD-40 in evenly across the surface using circular motions. When the subject returned to the room, the same testing procedure as previously described was repeated. A slip trial was considered successful if: 1) the subject did not perceive in advance that a contaminant had been applied; and 2) the right foot landed wholly on the region where the contaminant had been applied as determined from a review of the foot imprint in the WD-40. The principle investigator then recorded the available coefficient of friction in the region adjacent to the start of the slip (evident from shoe imprints on the WD- 40). The investigator followed the test protocols described in the Instruction Manual for the variable incidence tribometer and ASTM Standard F 1679-96. The test foot Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 8 was prepared by sanding it in five clockwise and five counterclockwise rotations with 180-grit 3M (3M, Minneapolis, MN) wet-or-dry sandpaper. Four measurements of the available coefficient of friction were taken in orthogonal directions adjacent to the region of slip onset and then averaged. Following the slip trial, the floor surface along the length of the walkway was cleaned by initially wiping the surface with mineral spirits. Then a combination of soap and water was applied, followed by thoroughly rinsing the surface with water, only. After the surface had dried, available coefficient of friction measurements were again recorded to ensure that the surface had been thoroughly cleaned and restored to previous dry surface slip resistance values. Data Analysis The posterior calcaneous markers were identified manually using VICON 370 Workstation software (Oxford Metrics, Ltd., Oxford, England) and then automatically digitized. Kinematic data for the heel, as well as ground reaction force data were imported into DataPac software (v. 2K2, Run Technologies, Mission Viejo, CA) for further analysis. Two sources of data were used to determine if a heel slip occurred during the trial in which the contaminant was present: review of the digitally acquired video as well as the kinematic data. The outcome was qualified as a slip if 1) the video showed that the subject’s heel or foot slipped on the surface during limb loading; or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 9 2) kinematic data revealed that the heel marker continued anterior displacement following the initiation of limb loading. Digitally acquired vertical, anterior-posterior, and medial-lateral forces were filtered in DataPac using a 4th order, 45 Hz, low pass Butterworth filter with a zero- lag compensation.2 2 For each walking trial, the COFu was calculated by dividing the resultant shear force (computed from the filtered anterior-posterior, F a p, and medial- lateral, F m l, forces) by the filtered vertical force (Fven) (Equation 5-1). During weight acceptance, the peak COFu value resulting from a shear force that would contribute to the foot sliding anteriorly was identified. To avoid spurious COFu values resulting from division by small numbers, the data were screened and only COFu data in which the vertical GRF's exceeded 50N were analyzed.1 1 For subjects who did not slip, peak COFu was determined for the trial in which the contaminant was present. For subjects who experienced a heel slip, peak COFu was determined by averaging the peak COFu values recorded from previous walking trials for the right lower extremity. Only trials that were within 5% of the walking velocity of the slip trial were used (average of five trials per subject). COFu = (5-1) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 0 The slip resistance difference was calculated by subtracting each subject’s peak COFu from the available slip resistance of the floor surface (measured with the tribometer). Using this convention, a positive slip resistance difference reflected that the available slip resistance exceeded the subject’s COFu- In contrast, a negative slip resistance difference reflected that the subject’s COFu exceeded the available COF. Statistical Analysis Descriptive statistics (mean, standard deviation, minimum and maximum) were calculated for all variables assessed. To determine the relationship of the observed slip events to the calculated slip resistance difference, logistic regression analysis using a forward stepwise Likelihood Ratio procedure was performed (F to enter = 0.05; F to remove = 0.10). The predictor variable was slip resistance difference (i.e., the difference between available slip resistance and peak COFu). The dependent variable, slip outcome, was coded as a binary variable with 0 equal to “no slip” and 1 equal to “slip”. A second logistic regression was performed to determine the relationship of the observed slip events to knowledge of only the available slip resistance. The predictor variable was available slip resistance (measured with the tribometer), and again, slip outcome was the dependent binary variable. To determine fit of the model due to the inclusion of the predictor variable, the Hosmer and Lemeshow’s measure (Rl ) was calculated. All statistics were calculated using the SPSS 10.0 statistical software (SPSS Inc., Chicago, IL). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 RESULTS Of the 52 subjects tested, 14 (8 males, 6 females) had to be excluded from the final analysis. Three subjects perceived in advance, that the surface might be slippery, while 11 did not fully step on the force plate where the contaminant had been placed. On the average, self-selected walking velocity of the remaining 38 participants was 96.6 ±13.0 m/min (range, 76.9 to 128.2 m/min) and the mean peak COFu wasn = 0.21 ± .04 (range, /x = 0.14 to// = 0.31). The mean available slip resistance, as measured by the tribometer, was fx = 0.23 ± .04 (range, \x = 0.15 to fx = 0.31), and the mean calculated COF difference was fx = 0.02 ± .01 (range, jx = -.09 to {x = 0.14). Fourteen of thirty eight subjects (37%) experienced a heel slip during the trial in which the contaminant was applied. Slip Probability based on Slip Resistance Difference Slip resistance difference (i.e., the difference between the available floor surface slip resistance and an individual’s peak COFu) was a significant predictor of slip outcome (Wald = 8.085, p = 0.004; Figure 5-4; Table 5-2). The logistic regression model relating these two variables provided a good fit to the data, and overall correctly predicted 89.5% of the slip outcomes experienced by our subjects. The model most accurately predicted when a slip did not occur, with correct predictions occurring for 95.8% of the events (23 of 24). The model accurately Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2 predicted 78.6% of the instances when a slip did occur (11 of 14). Overall, the model relating slip resistance difference to slip outcome accounted for 48.5% of the variance in slip outcomes {Ri - .485). Q. • 4 P 4 m e iS* I - J O 00.00 0.10 0.20 Slip Resistance K lfsw iice (Available - Utilized) Figure 5-4. Scatter plot showing the relationship between the Slip Resistance Difference (calculated by subtracting each subjects peak COFu from the available slip resistance on the floor surface) and the probability of a slip event occurring as calculated by logistic regression (Rl = 0.485) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 Table 5-2 Logistic Regression Model Details describing the Probability of a Slip Event based on the Slip Resistance Difference Slip Event Constant, fio (S.E.) -.353 (.492) Predictor Coefficient, /?i (S.E.) -54.740 (19.251) Significance 0.004 Exp f$ (95% Confidence Interval for Exp ($) 1.685 E-24 (6.913 E-41 to 4.105 E-08) Rl .485 Based on the logistic regression model, values of slip resistance difference that would correspond to a specific probability of slip occurring were calculated (Table 5-3). These calculations revealed that there was a 1% probability of a slip occurring when the available slip resistance exceeded the COFu by .077. However, there was a 50% probability of a slip occurring when the available slip resistance was 0.006 less than the COFu, and a 99% probability when the available slip resistance was 0.090 less than the COFu- Table 5-3 Logistic Regression Prediction of Slip Resistance Difference (available floor surface slip resistance - individual’s utilized coefficient of friction) and Corresponding Probabilities of a Slip Event Occurring Probability of Slip Event Occurring Slip Resistance Difference .01 +0.077 .05 +0.047 .50 -0.006 .95 -0.060 .99 -0.090 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 4 Slip Probability based on Available Slip Resistance In the second model generated, available slip resistance (as measured by the tribometer) also significantly predicted slip outcome (Wald = 6.321, p = 0.012; Figure 5-5; Table 5-4). Overall, the model correctly predicted 78.9% of the slip outcomes experienced by our subjects, and was most accurate at predicting when a slip did not occur (87.5%, 21 of 24 cases). The model accurately predicted 64.3% of the instances when a slip did occur (9 of 14). Overall, the model relating available slip resistance to slip outcome accounted for 16.2% of the variance in slip outcomes (Rl2 =.162). 1 . 0 T — ........... — I 0.0 .15 .20 .25 JO .35 Available Slip Resistance Figure 5-5. Scatter plot showing the relationship between the available coefficient of friction (measured with tribometer) and the probability of a slip event occurring as calculated by logistic regression (Rl = 0.162) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 Table 5-4 Logistic Regression Model Details describing the Probability of a Slip Event based on the Available Slip Resistance Slip Event Constant, (S.E.) 5.665 (2.455) Predictor Coefficient, fix (S.E.) -27.472 (10.927) Significance 0.012 Exp (S (95% Confidence Interval for Exp f$) 1.172 E-12 (5.859 E-22 to 2.345 E-03) Rl .162 Based on the second logistic regression model, when the available slip resistance was p = 0.153, there was an 81% chance of slipping. In contrast, when the available slip resistance was /u - 0.308, the chance of slipping was only 6%. At the median available slip resistance level (p = 0.229), the chance of slipping was approximately 35%. DISCUSSION The results from this study indicate that the available slip resistance, as measured by the variable incidence tribometer, can accurately predict slip events. Knowledge of the available slip resistance, in combination with an individual’s COFu allowed for the greatest accuracy in predicting slip outcome (89.5%). With knowledge of only the available slip resistance, the accuracy of prediction was reduced to 78.9%, over the range of floor surface slip resistance values evaluated in this study (p = 0.15 top - 0.31). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 Relatively small changes in slip resistance difference (from an applied perspective) resulted in large changes in the probability of a slip occurring. For example, while the probability of a slip occurring was only 5% when the slip resistance difference was +.047, there was a 50% probability when the slip resistance difference was -0.006. By definition, the magnitude of the slip resistance difference is affected by both changes in available floor surface slip resistance as well as an individual’s COFu. Under conditions of lowered available slip resistance, increases in peak COFu as subtle as 0.05, may contribute to a substantial increase in slip probability. These findings reinforce the need for studies aimed at identifying tasks, as well as individual or group characteristics that require greater friction needs. Based on the logistic regression model developed in the current study, a 5% probability of a slip occurring existed when the available floor surface slip resistance was n - .32. As existing standards for safe flooring have set the threshold for safe flooring at ju> 0.50, the findings of the current study suggest that slips during comfortable speed walking on a level surface by healthy young adults would be unlikely. However, the current study also points to the importance of the relationship between an individual’s peak COFu and the available slip resistance of the floor. Thus statements endorsing an absolute threshold for safe flooring should be interpreted with caution, due to the interaction between the available slip resistance and peak COFu- Faster walking speeds, alternative tasks (e.g., turning, stair negotiation), or different age groups might have higher peak COFu values that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 7 would alter the probability of a slip based on a threshold slip resistance value of a floor surface. It has been suggested that logistic regression can be used to assess the bias of a given tribometer.4 6 Hanson defined bias as the slip resistance difference (i.e, tribometer measurement minus COFu) when the fa ll probability was 50%, with a value of 0.0 indicating no bias. As slip recovery is likely confounded by multiple factors including balance, muscular strength, power, and flexibility it is likely that by using this definition, the determination of bias could be influenced by the population of individuals studied (e.g., healthy and strong vs. deconditioned individuals with pathology that interfered with walking). We elected to define the bias of the tribometer as the slip resistance difference when the slip probability was 50%. Based on this approach, the bias of the tribometer used in the current study was p . - - 0.006, suggesting a small underestimation of the available friction that would lead to a slip. Similar to the estimate of bias, Hanson defined the accuracy of prediction of th th a tribometer as the range of slip resistance difference between the 5 and 95 probability estimates of a fall occurring, with an ideal value of 0.0 indicating the highest accuracy.4 6 For the reasons noted above, we defined the accuracy of the prediction of the tribometer based on the range of slip resistance differences between the 5th and 95th probability estimates of a slip occurring. For the variable incidence tribometer, the accuracy of prediction was .107. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 Our results are comparable to Hanson et al. who reported an overall prediction rate of 96.1 % for slip or fall events.4 6 The notable difference between studies is that the dynamometer used in the current study measured static available COF, while the device employed by Hanson and colleagues4 6 measured dynamic available COF. As several nation-wide standards encompass use of tribometers that measure static COF, it is important to establish the relationship between static COF values recorded and slip outcome. While Kulakowski et al.6 1 recorded measures of static COF and slip outcome, they did not report on the probability of a slip occurring, nor provide data on the accuracy and bias of the static COF device used in their study. Collectively, the findings of Hanson and collagues4 6 and the current study suggest that both static and dynamic testers of available slip resistance provide information useful for identifying human and environmental combinations that may create a greater risk for slip onset. Care must be exercised when generalizing these data to other types of tribometers, however, as different styles of tribometers yield different measurements, that do not necessarily correlate.36,4 1 ’1 0 0 CONCLUSIONS This study demonstrated that knowledge of a person’s COFu and the available friction (as measured by the VIT) can be used to predict the probability of a slip event during level walking in young adults. As measures of static coefficient Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 9 vary across the type of tribometer used, the relationships reported in the current study will likely apply to the variable incidence tribometer only. However, other tribometers, whether measuring static or dynamic coefficient of friction, could be assessed using similar procedures to determine their accuracy, bias, and ability to predict slip probability. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 0 CHAPTER VI SUMMARY AND CONCLUSIONS SUMMARY Slips are a leading cause of falls and injuries in both the work and home environments. While frequently the consequence of a slip is minor, at times, a slip can lead to devastating outcomes including fractures, brain injuries, and even death. Though the underlying biomechanical causes of slips are well understood, an understanding of which individuals or groups might be at greater risk of slipping, why they might be at greater risk of slipping, and how to assess the risk of slipping in the real world environment has been limited. The overall aim of this dissertation was to provide additional insight into these questions. Chapter III of this dissertation focused on determining if differences existed in peak COFu for persons of different ages and genders, while walking at different speeds. A second goal of Chapter III was to determine to what extent an existing model27,2 8 could be used to predict individuals who might have high friction needs. This model, which uses leg length and step length to estimate peak COFu needs, has served, in part, as the underpinning for the current standards for safe flooring (i.e., p i < 0.05). It was hypothesized, based on the model of Ekkebus and Killey,2 7’2 8 that during walking, reductions in stride length in older compared to younger adults would contribute to reduced peak COFu values in the older adults population. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 relevance of the question was multi-fold. If older adults, who slip and fall more frequently than younger adults, simply require additional friction at the foot-floor interface, then interventions which increase the available friction (e.g., changes in flooring or shoe wear) should be targeted. If however, they did not have higher friction needs then their younger counter-parts, then their falls might better be attributed to reductions in the ability to recover following slip onset. From Chapter III, it was determined that older adults did not have higher COFu values, and thus were not at greater risk for slip onset. Based on literature related to the relative steps lengths of women compared to men when walking at different speeds, it also was hypothesized that women would have higher COFu than men at slow walking speeds, but lower COFu than men at fast speeds. While theses gender-related differences in peak COFu were found, the expected differences in stride length did not emerge. The relationship between normalized stride length and COFu in the study indicated that only 18% of the variability in peak COFu could be explained by normalized stride length (a ratio of leg length to step length). In summary, it was determined in Chapter III that age and gender related differences in peak COFu exist across walking speeds, but that these differences could not be attributed solely to the selected anthropometric and stride characteristic variables evaluated in the study. Collectively, these findings suggested that factors, other than age and the selected anthropometric variables considered likely have a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 2 large role in determining peak COFu. Additionally, the wide inter-subject variability in COFu values demonstrated in the study suggested that minimum threshold levels used to define “safe” walkway surfaces should consider not only average utilized COF, but also the range of values used by individual subjects. Chapter IV of this dissertation focused on understanding the underlying biomechanical mechanism contributing to the differences in peak COFu observed across subjects. A 15 segment model was used to estimate the location of the CM of young subjects. Then regression analysis was used and identified that the relative location of the CM relative to the CP (as defined by an angular position relative to vertical) served as a strong predictor of peak COFu, and accounted for 55% of the variance in peak COFu values for our subjects. In addition, the velocity of the CM in the anterior direction also was a significant predictor of peak COFu- This finding was interesting in light of previously conflicting studies on the influence of walking speed on peak COFu values.12,13,16,32,109,1 1 6 While the findings in Chapter IV indicated that the walking velocity was a predictor of peak COFu, a larger portion of the variance was accounted for by the location of the CM relative to the CP. Thus the method used for increasing velocity (e.g., increasing stride length vs. increasing cadence) could conceivably contribute to differences observed in peak COFu across walking speeds. Collectively, the CM to CP angle, and the CM velocity in the anterior direction served as strong predictors of peak COFu during weight acceptance, and accounted for 62% of variance in peak COFu- The findings of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 Chapter IV suggested that person-specific CM to CP relationships may increase an individual’s risk for slip initiation. However, only 62% of the variance in COFu values could be explained by the model, suggesting further investigation is still required to identify other factors that influence peak COFu- Chapter V of this dissertation focused on determining to what extent a slip event could be predicted. Greater than 50 devices exist for the measurement of slip resistance, and the values recorded by the different devices do not correlate well. Nor is it well understood, how the measurements recorded from a single device relate to the probability of a slip occurring. While a n> 0.50 threshold for safe flooring has been established, this standard was developed using a device that is not portable (the James Machine). In this experiment, the portable variable incidence tribometer was used to relate human (an individual’s peak COFu) and environmental factors (available friction on the floor surface) to slip outcome. The variable incidence tribometer is frequently used to assess slip resistance, and has demonstrated acceptable reliability and bias in previous studies.1 0 0 The primary finding of Chapter V was that static measures of available floor surface slip resistance could be used to predict the probability of a slip event occurring. Knowledge of the available static slip resistance, in combination with an individual’s COFu allowed for a prediction of slips accuracy rate of 89.5%. Prior to this study, only dynamic measures of available slip resistance4 6 had been related to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 4 the probability of a slip event occurring. As tribometers providing static measures of COF are frequently used to assess slip resistance, the relationships identified in this study provide data necessary for accurately interpreting the safety of walking surfaces in home, work and community environments. CONCLUSIONS Small changes in peak COFu relative to the available slip resistance of the floor surface (as determined by a tribometer that evaluates static friction) can have a large impact on the probability of a slip occurring. Statements endorsing an absolute threshold for safe flooring should be interpreted with caution, due to the interaction between the available slip resistance and peak COFu. Faster walking speeds, alternative tasks (e.g., turning, stair negotiation), or different age groups might have higher peak COFu values that would alter the probability of a slip based on a threshold slip resistance value of a floor surface. Refinement of models developed to predict individual friction needs should aid in identifying persons with high COFu, and thus greater slip initiation risk. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 5 REFERENCES 1. American Society o f Testing and Materials, Annual Book o f ASTM Standards 2002. West Conshohocken, PA: ASTM International; 2002. 2. Andriacchi TP, Ogle JA, and Galante JO. Walking speed as a basis for normal and abnormal gait measurements. Journal o f Biomechanics. 1977; 10(4):261-268. 3. Architectural and Transportation Barriers Compliance Board. Americans with Disabilities Act (ADA) accessibility guidelines for buildings and facilities. Final guidelines. Federal Register. 1991;56(144):35408-542. 4. Bell JL, Gardner LI, and Landsittel DP. Slip and fall-related injuries in relation to environmental cold and work location in above-ground coal mining operations. American Journal o f Industrial Medicine. 2000;38:40-48. 5. Bentley TA and Haslam RA. Slip, trip and fall accidents occurring during the delivery of mail. Ergonomics. 1998;41(12):1859-72. 6. Berg WP, Alessio HM, Mills EM, and Tong C. Circumstances and consequences of falls in independent community-dwelling older adults. Age & Ageing. 1997;26(4):261 -8. 7. Blanke D, Hageman, PA. Comparison of gait of young men and elderly men. Phys Ther. 1989;69:144-148. 8. Bowman R, Strautins CJ, Westgate P, and Quick GW, Implications for the development of slip resistance standards arising from rank comparisons of friction-test results obtained using different walkway-safety tribometers under various conditions, in Metrology o f Pedestrian Locomotion and Slip Resistance, ASTM ST P 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 112-136. 9. Braun R and Brungraber RJ. A comparison of two slip-resistance testers. American Society o f Testing and Materials. 1978:49-59. 10. Braun R, Roemer D, Hoechst AG, and Gersthofen WG. Influence of waxes on static and dynamic friction. Soaps/Cosmetics/Chemical Specialties. 1974. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 6 11. Buczek FL and Banks SA. High-resolution force plate analysis of utilized slip resistance in human walking. Journal o f Testing and Evaluation. 1996;24(6):353-358. 12. Buczek FL, Cavanagh PR, Kulakowksi BT, and Pradhan P, Slip resistance needs of the mobility disabled during level and grade walking, in Slips, Stumbles, and Falls: Pedestrian Footwear and Surfaces, ST P 1103, G.B. Everett, Editor. 1990, American Society for Testing and Materials: Philadelphia, PA. 39-54. 13. Bumfield JM and Powers CM, Influence of age and gender of utilized coefficient of friction during walking at different speeds, in Metrology o f pedestrian locomotion and slip resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 3-16. 14. Cavagna GA. Force platforms as ergometers. Journal o f Applied Physiology. 1975;39(1):174-179. 15. Cavagna GA and Margaria R. Mechanics of walking. Journal o f Applied Physiology. 1966;21(l):271-278. 16. Cham R, Musolino M, and Redfem MS. Heel contact dynamics during slip events. Proceedings o f the IEA 2000/HFES 2000 Congress. 2000;4:514-517. 17. Cham R and Redfem MS. Changes in gait when anticipating slippery floors. Gait and Posture. 2002; 15:159-171. 18. Chang W-R. The effect of surface roughness on dynamic friction between neolite and quarry tile. Safety Science. 1998;29:89-105. 19. Chang W-R. The effect of surface roughness of the measurement of slip resistance. International Journal o f Industrial Ergonomics. 1999;24:299-313. 20. Chang WR, Gronqvist R, Leclercq S, et al. The role of friction in the measurement of slipperiness. Part II. Survey of friction measurement devices. Ergonomics. 2001;44(13):1233-1261. 21. Chiou S-Y, Bhattacharya A, and Succop PA. Effect of workers' shoe wear on objective and subjective assessment of slipperiness. American Industrial Hygiene Association Journal. 1996;57:825-831. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 7 22. Christina KA and Cavanagh PR. Ground reaction forces and frictional demands during stair descent: Effects of age and illumination. Gait and Posture. 2002;15:153-158. 23. Courtney TK, Sorock GS, Manning DP, Collins JW, and Holbein-Jenny MA. Occupational slip, trip, and fall-related injuries- Can the contribution of slipperiness be isolated? Ergonomics. 2001;44(13):1118-1137. 24. Dempster WT, Space requirements of the seated operator, in WADC Technical Report-55-159.1955, Aerospace Medical Research Laboratory: Wright-Patterson Air Force Base, Ohio. 25. Di Pilla S and Vidal K. State of the art in slip-resistance measurement: A review of current standards and continuing developments. Professional Safety. 2002;June:37-42. 26. Dravitzki VK and Potter SM. The use of the tortus and the pendulum with the 4S rubber for the assessment of slip resistance in the laboratory and the field. Journal o f Testing and Evaluation. 1997;25(1):127-134. 27. Ekkebus CF and Killey W. Validity of 0.5 static coefficient of friction (James machine) as a measure of safe walkway surfaces. Proceedings o f Chemical Specialties Manufacturers Association 67th Mid-Year Meeting. 1971:250- 253. 28. Ekkebus CF and Killey W. Measurement of safe walkway surfaces. Soaps/Cosmetics/Chemical Specialties. 1973(February):40-45. 29. Englander F, Hodson TJ, and Terregrossa RA. Economic dimensions of slip and fall injuries. Journal o f Forensic Sciences. 1996;41(5):733-46. 30. English W, Improved tribometry on walking surfaces, in Slips, Stumbles, and Falls: Pedestrian Footwear and Surfaces, ASTM STP 1103, B.E. Gray, Editor. 1990, American Society for Testing and Materials: Philadelphia. 73- 81. 31. Fendley AE and Marpet ML Required coefficient of friction versus walking speed: Potential influences of footwear and walkway surfaces. Journal o f Testing and Evaluation. 1996;24(6):359-367. 32. Fendley AE and Medoff HP. Required coefficient of friction versus top- piece/outsole hardness and walking speed: Significance of correlations. Journal o f Forensic Sciences. 1996;41(5):763-769. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 8 33. Gillet C, Duboy J, Barbier F, et al. Contribution of accelerated body masses to able-bodied gait. American Journal o f Physical Medicine and Rehabilitation. 2003;82(February): 101 -109. 34. Gooch HD. A comparison of four methods of measuring slip resistance of floor polishes. Chemical Times and Trends. 1981(July):47-51. 35. Grabiner PC, Biswas ST, and Grabiner MD. Age-related changes in spatial and temporal gait variables. Archives o f Physical Medicine and Rehabilitation. 2001 ;82(1):31 -5. 36. Grieser BC, Rhoades TP, and Shah RJ. Slip resistance: Field measurements using two modem slipmeters. Professional Safety. 2002;June:43-48. 37. Gronqvist R. Mechanisms of friction and assessment of slip resistance of new and used footwear soles on contaminated floors. Ergonomics. 1995;3 8(2):224-241. 38. Gronqvist R, Slips and falls, in Biomechanics in Ergonomics, S. Kumar, Editor. 1999, Taylor and Francis: London. Chapter 19, 351-375. 39. Gronqvist R, Abeysekera J, Gard G, et al. Human-centered approaches in slipperiness measurement. Ergonomics. 2001 ;44(13): 1167-1199. 40. Gronqvist R and Hirvonen M. Slipperiness of footwear and mechanisms of walking friction on icy surfaces. International Journal o f Industrial Ergonomics. 1995;16:191-200. 41. Gronqvist R, Hirvonen M, Rajamaki E, and Matz S. The validity and reliability of a portable slip meter for determining floor slipperiness during simulated heel strike. Accident Analysis and Prevention. 2003;35:211-225. 42. Gronqvist R, Hirvonen M, and Tohv A. Evaluation of three portable floor slipperiness testers. International Journal o f Industrial Ergonomics. 1999;25:85-95. 43. Gronqvist R, Hirvonen M, and Tuusa A. Slipperiness of the shoe-floor interface: Comparison of objective and subjective assessments. Applied Ergonomics. 1993;24(4):258-262. 44. Guevin PR. Review of skid and slip resistance standards relatable to coating. Journal o f Coatings Technology. 1978;50(643):33-38. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 9 45. Hageman PA and Blanke DJ. Comparison of gait of young women and elderly women. Physical Therapy. 1986;66(9):1382-7. 46. Hanson JP, Redfem MS, and Mazumdar M. Predicting slips and falls considering required and available friction. Ergonomics. 1999;42( 12): 1619- 33. 47. Himann JE, Cunningham DA, Rechnitzer PA, and Paterson DH. Age-related changes in speed of walking. Medical Science o f Sports and Exercise. 1988;20(2):161-6. 48. Hoyert DL, Kochanek KD, and Murphy SL. Deaths: final data for 1997. National Vital Statistics Report. 1999;47(19):1-104. 49. Hunter RB. Method of measuring frictional coefficients of walkway materials. Bureau o f Standards Journal o f Research. 1929;5:329-347. 50. Ilda H and Yamamuro T. Kinetic Analysis of the Center of Gravity of the Human Body in Normal and Pathological Gaits. Journal o f Biomechanics. 1987;20(10):987-995. 51. Inman VT, Ralston HJ, and Todd F, Human Walking. Baltimore,MD: Williams and Wilkins Company; 1981. 52. Inman VT, Ralston HJ, Todd F, and Lieberman JC, Human walking. Baltimore: Williams & Wilkins; 1981. 53. Irvine CH. A new slipmeter for evaluating walkway slipperiness. Materials Research and Standards. 1967;7(12):535-542. 54. James DI. Rubbers and plastics in shoes and flooring: The importance of kinetic friction. Ergonomics. 1983;26(l):83-99. 55. Jensen OC. Non-fatal occupational fall and slip injuries among commerical fisherman analyzed by use of the NOMESCO injury registration system. American Journal o f Industrial Medicine. 2000;37:637-644. 56. Kerrigan DC, Lee LW, Collins JJ, Riley PO, and Lipsitz LA. Reduced hip extension during walking: Healthy elderly and fallers versus young adults. Archives o f Physical Medicine and Rehabilitation. 2001 ;82(l):26-30. 57. Kerrigan DC, Todd MK, and Della Croce U. Gender differences in joint biomechanics during walking: Normative study in young adults. American Journal o f Physical Medicine and Rehabilitation. 1998;77(l):2-7. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 0 58. Kim I-J. Wear progression of shoe heels during slip resistance measurements. Proceedings ofthelEA 2000/HFES 2000 Congress. 2000;4:498-501. 59. Kim I-J, Smith R, and Nagata H. Microscopic observations of the progressive wear on shoe surfaces that affect the slip resistance characteristics. International Journal o f Industrial Ergonomics. 2001 ;28:17-29. 60. Kim I-J and Smith RM, A critical analysis of the relationship between shoe- heel wear and pedestrian/walkway slip resistance, in Metrology o f pedestrian locomotion and slip resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 33- 48. 61. Kulakowksi BT, Buczek FL, Cavanagh PR, and Pradhan P. Evaluation of performance of three slip resistance testers. Journal o f Testing and Evaluation. 1989;17(4):234-240. 62. Layne LA and Landen DD. A descriptive analysis of nonfatal occupational injuries to older workers, using a national probability sample of hospital emergency departments. Journal o f Occupational and Environmental Medicine. 1997;39(9):855-65. 63. Leamon TB and Murphy PL. Occupational slips and falls: More than a trivial problem. Ergonomics. 1995;38(3):487-98. 64. Lee CR and Farley CT. Determinants of the center of mass trajectory in human walking and running. The Journal o f Experimental Biology. 1998;201:2935-2944. 65. Lexell J. Human aging, muscle mass, and fiber type composition. J Gerontol A Biol SciMedSci. 1995;50A(special issue):l 1-16. 66. Lockhart TE, Woldstad JC, and Smith JL, Assessment of slip severity among different age groups, in Metrology o f pedestrian locomotion and slip resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 17-32. 67. Lockhart TE, Woldstad JC, Smith JL, Hsiang SM, and Li P. Prediction of falls using a robust definition of slip distance and adjusted required coefficient of friction. Proceedings o f the IEA 2000/HFES 2000 Congress (abstract). 2000;4:506-509. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 68. Manning DP, Ayers IM, Jones C, Bruce M, and Cohen K. The incidence of underfoot accidents during 1985 in a working population of 10,000 Merseyside people. Journal o f Occupational Accidents. 1988;10:121-130. 69. Manning DP and Jones C. The effect of roughness, floor polish, water, oil and ice on underfoot friction: current safety footwear solings are less slip resistant than microcellular polyurethane. Applied Ergonomics. 2001 ;32:185- 196. 70. Manning DP, Jones C, and Bruce M. Boots for oily surfaces. Ergonomics. 1985;28(7):1011-1019. 71. Manning DP, Jones C, and Bruce M. Proof of shoe slip-resistance by a walking traction test. Journal o f Occupational Accidents. 1990;12:255-270. 72. Marigold DS and Patla AE. Strategies for dynamic stability during locomotion on a slippery surface: Effects of prior experience and knowledge. Journal o f Neurophysiology. 2002;88:339-353. 73. Marietta W. The effects of wetness on coefficient of friction measurement evaluated with three slip testers with varied residence times on selected materials. IEEE. 1995:167-170. 74. Marpet MI. On threshold values that separate pedestrian walkways that are slip resistant from those that are not*. American Journal o f Forensic Sciences. 1996;41(5):747-755. 75. Marpet MI, Issues in the development of modem walkway-safety tribometry standards: required friction, contextualization of test results, and non- propietary standards, in Metrology o f pedestrian locomotion and slip resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 96-111. 76. Marpet MI and Fleischer DH. Comparsion of walkway safety tribometers: Part two. Journal o f Testing and Evaluation. 1997;25(1): 115-126. 77. McCoy CE, Hadjipavlou AG, Overman T, Necessary JT, and Wolf C. Work- related low back injuries caused by unusual circumstances. Journal o f Orthopaedic and Sports Physical Therapy. 1997;26(5):260-265. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 78. Medhoff H, Fleisher DH, and Di Pilla S, Comparison of slip resistance measurements between two tribometers using smooth and grooved Neolite®- test-liner feet, in Metrology o f Pedestrian Locomotion and Slip Resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 67-72. 79. Medoff H, Brungraber R, Hilferty C, Patel J, and Mehta K, Variable inclineable stepmeter: using test subjects to evaluate walkway surface/footwear combinations, in Metrology o f Pedestrian Locomotion and Slip Resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 51-57. 80. Medoff HP. Viscoelastic characteristics of typical shoe outsole materials and their effects on walkway friction models. Journal o f Forensic Sciences. 1996;41(5):756-762. 81. Melton LJ, III, Thaner M, and Ray NF. Fractures attributable to osteoporosis: Report from the national osteoporosis foundation. J Bone Miner Res. 1997;12:16-23. 82. Menck H and Jorgenson U. Frictional forces and ankle fractures in sports. British Journal o f Sports Medicine. 1983; 17(4): 13 5-136. 83. Menz HB, Lord SR, and McIntosh AS. Slip resistance of casual footwear: Implications for falls in older adults. Gerontology. 2001;47:145-149. 84. Merscher JH. Slip resistance and floor polishes. Maintenance Supplies. 1976;February: 19-20, 72-74. 85. Murray MP. Gait as a total pattern of movement. American Journal o f Physical Medicine. 1967;46:290-332. 86. Murray MP, Drought AB, and Kory RC. Walking patterns of normal men. Journal o f Bone and Joint Surgery. 1964;46A:335-360. 87. Murray MP, Kory RC, and Clarkson BH. Walking patterns in healthy old men. Journal o f Gerontology. 1969;24:169-178. 88. Murray MP, Kory RC, Clarkson BH, and Sepic SB. Comparison of free and fast speed walking patterns of normal men. American Journal o f Physical Medicine. 1966;4:8-25. 89. Murray MP, Kory RC, and Sepic SB. Walking patterns of normal women. Archives o f Physical Medicine and Rehabilitation. 1970;51:637-650. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 90. Myung R and Smith JL. The effect of load carrying and floor contaminants on slip and fall parameters. Ergonomics. 1997;40(2):235-246. 91. Nagata H, An analysis of the sliding properties of worker's footwear and clothing on roof surfaces, in Metrology o f pedestrian locomotion and slip resistance, ASTM STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 58-66. 92. Nyberg L, Gustafson Y, Berggren D, Brannstrom B, and Gosta B. Falls leading to femoral neck fractures in lucid older people. Journal o f the American Geriatrics Society. 1996;44(2): 156-160. 93. Ostrosky KM, VanSwearingen JM, Burdett RG, and Gee Z. A comparison of gait characteristics in young and older subjects. Physical Therapy. 1994;74(7):637-646. 94. Pai Y-C and Iqbal K. Simulated movement termination for balance recovery: Can movement strategies be sought to maintain stability in the presence of slipping or forced sliding? Journal o f Biomechanics. 1999;32:779-786. 95. Perkins PJ, Measurement of slip between shoe and ground during walking, in Walkway Surfaces: Measurement o f Slip Resistance, ASTM STP 649, C. Anderson and J. Senne, Editors. 1978, American Society for Testing and Materials: Philadelphia. 71-87. 96. Perkins PJ, Measurement of slip between the shoe and ground during walking, in Walkway Surfaces: Measurement o f Slip Resistance, ASTM STP 649, C. Anderson and J. Senne, Editors. 1978, American Society for Testing and Materials: Baltimore. 71-87. 97. Perkins PJ and Wilson MP. Slip resistance testing of shoes - New developments. Ergonomics. 1983;26:73-82. 98. Perry J, Gait Analysis, Normal and Pathological Function. Thorofare, NJ: Charles B. Slack; 1992. 99. Powers CM, Bumfield JM, Lim P, Brault J, and Flynn JE. Utilized coefficient of friction during walking: Static estimates exceed measured values. Journal o f Forensic Sciences. 2002;47(6):1303-1308. 100. Powers CM, Kulig K, Flynn JE, and Brault J. Repeatability and bias of two walkway safety tribometers. Journal o f Testing and Evaluation. 1999;27(6):368-374. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 101. Rabinowicz E. Stick and slip. Scientific American. 1956;194(May):109-118. 102. Redfem MS and Bidanda B. Slip resistance of the shoe-floor interface under biomechanically relevant conditions. Ergonomics. 1994;37(3):511-524. 103. Redfem MS, Cham R, Gielo-Perczak K, et al. Biomechanics of slips. Ergonomics. 2001 ;44(13): 1138-66. 104. Robinson WH and Kopf RE. Evaluation of the horizontal pull slipmeter. Materials Research and Standards. 1969;July:22-24. 105. Sacher A. Slip resistance and the James Machine 0.5 static coefficient of friction - sine qua non*. American Society o f Testing and Materials Standardization News. 1993; August:52-59. 106. Sattin RW. Falls among older persons: A public health perspective. Annual Review o f Public Health. 1992;13:489-508. 107. Saunders JBdM, Inman VT, and Eberhart HD. The major determinants in normal and pathological gait. Journal o f Bone and Joint Surgery. 1953;35A(3):543-557. 108. Sigler PA, Geib, M.N., Boone, T.H. Measurement of the slipperiness of walkway surfaces. Journal ofResearch o f the National Bureau o f Standards. 1948;40:339-346. 109. Skiba R. Sicherheitsgrenzwerte zur vermeidung des ausgleitens auf fussboden. Zeitschrift Fur Arbeitswissenschaft. 1988;14:47-51 (in German with English summary). 110. Smith RH, Examination of sticktion in wet-walkway slip resistance testing, in Metrology o f Pedestrian Locomotion and Slip Resistance, STP 1424, M.I. Marpet and M.A. Sapienza, Editors. 2003, ASTM International: West Conshohocken, PA. 73-86. 111. Strandberg L. On accident analysis and slip-resistance measurement. Ergonomics. 1983;26(1):11-32. 112. Strandberg L and Lanshammar H. The dynamics of slipping accidents. Journal o f Occupational Accidents. 1981 ;3:153-162. 113. Szymusiak SM and Ryan JP. Prevention of slip and fall injuries. Professional Safety. 1982;July:30-35. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 5 114. Tang P-F and Woollacott MH. Inefficient postural responses to unexpected slips during walking in older adults. Journal o f Gerontology: Medical Sciences. 1998;53 A(6):M471 -M480. 115. Tang P-F and Woollacott MH. Phase-dependent modulation of proximal and distal postural responses to slips in young and older adults. Journal o f Gerontology: Medical Sciences. 1999;54A(2):M89-M102. 116. Tisserand M. Progress in the prevention of falls caused by slipping. Ergonomics. 1985;28(7):1027-42. 117. Webb A, Anthropometric Source Book. Washington, D.C.: NASA Reference 1024; 1978. 118. Wilson MP and Perkins PJ. Evaluation of a slip-resistance test for shoes. Ergonomics. 1985;28(7): 1081-1085. 119. Winter DA, Biomechanics and Motor Control o f Human Movement. 2 ed. Toronto: John Wiley & Sons, Inc.; 1990. 120. Winter DA, Patia AE, Frank JS, and Walt SE. Biomechanical walking pattern changes in the fit and healthy elderly. Physical Therapy. 1990;70(6):340-347. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Factors contributing to patellofemoral joint stress: a comparison of persons with and without patellofemoral pain

PDF

At what age are gait characteristics mature? Evaluation of gait kinematics, kinetics, and intersegmental dynamics in 7 year-old children

PDF

Bimanual coordination of goal-directed multi-joint rapid aiming movements following stroke

PDF

Biomechanical and neuromuscular aspects of non-contact ACL injuries: The influence of gender, experience and training

PDF

Functional brain correlates for premovement planning and compensatory adjustments in rapid aimed movement

PDF

Biopsychosocial factors in major depressive disorder

PDF

Enabling clinically based knowledge discovery in pharmacy claims data: An application in bioinformatics

PDF

Does young adult Hodgkin's disease cluster by school, residence and age?

PDF

A comparative study of environmental factors associated with multiple sclerosis in disease-discordant twin pairs

PDF

Evaluation of the efficacy of two lipid-lowering treatments using serial quantitative coronary angiography: Two- and four-year treatment results

PDF

Extent, prevalence and progression of coronary calcium in four ethnic groups

PDF

Imputation methods for missing items in the Vitality scale of the MOS SF-36 Quality of Life (QOL) Questionnaire

PDF

Cardiopulmonary coupling in the control of breathing during exercise

PDF

Emotional intelligence and smoking risk factors in early adolescents

PDF

Cost -efficient design of main cohort and calibration studies where one or more exposure variables are measured with error

PDF

Combining compile -time and run -time parallelization

PDF

Critical beauties: Aesthetics, gender, and realism in John Ruskin and George Eliot

PDF

Impact of heterogeneity on evidence for linkage to type 2 diabetes mellitus in the Finnish population

PDF

Biomechanical and neuromuscular aspects of non-contact ACL injuries: the influence of gender experience and trainting

PDF

Associations of physical fitness and other risk factors with cardiovascular and non-violent mortality

Asset Metadata

Creator Burnfield, Judith Marie (author)

Core Title Human and environmental factors contributing to slip events during walking

School Graduate School

Degree Doctor of Philosophy

Degree Program Biokinesiology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag engineering, environmental,health sciences, occupational health and safety,health sciences, recreation,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Powers, Christopher (committee chair), [illegible] (committee member), Azen, Stanley (committee member), Gordon, James (committee member), Perry, Jacquelin (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-629160

Unique identifier UC11334887

Identifier 3116673.pdf (filename),usctheses-c16-629160 (legacy record id)

Legacy Identifier 3116673.pdf

Dmrecord 629160

Document Type Dissertation

Rights Burnfield, Judith Marie

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA