the graded redefined assessment of strength, … · upper limb function is important for...
TRANSCRIPT
THE GRADED REDEFINED ASSESSMENT OF STRENGTH, SENSIBILITY AND
PREHENSION (GRASSP): DEVELOPMENT OF THE SCORING APPROACH,
EVALUATION OF PSYCHOMETRIC PROPERTIES AND THE RELATIONSHIP
OF UPPER LIMB IMPAIRMENT TO FUNCTION
By
Sukhvinder Kalsi-Ryan
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy
Graduate Department of Rehabilitation Science
University of Toronto
©Copyright by Sukhvinder Kalsi-Ryan (2011)
ii
ABSTRACT
THE GRADED REDEFINED ASSESSMENT OF STRENGTH, SENSIBILITY AND
PREHENSION (GRASSP): DEVELOPMENT OF THE SCORING APPROACH,
EVALUATION OF PSYCHOMETRIC PROPERTIES AND THE RELATIONSHIP
OF UPPER LIMB IMPAIRMENT TO FUNCTION
Sukhvinder Kalsi-Ryan
Doctor of Philosophy, 2011
University of Toronto
Graduate Department of Rehabilitation Science
Upper limb function is important for individuals with tetraplegia because upper limb
function supports global function for these individuals. As a result, a great deal of time and
effort has been devoted to the restoration of upper limb function. Appropriate outcome
measures that can be used to characterize the neurological status of the upper limb have been
one of the current barriers in substantiating the efficacy of interventions. Techniques and
protocols to evaluate changes in upper limb neurological status have not been applied to the
SCI population adequately. The objectives of this thesis were to develop a measure; which is
called the Graded Redefined Assessment of Strength Sensibility and Prehension (GRASSP).
Development of the scoring approach, testing for reliability and construct validity, and
determining impairment and function relationships specific to the upper limb neurological
were established. The GRASSP is a clinical measure of upper limb impairment which
incorporates the construct of “sensorimotor upper limb function”; comprised of three
iii
domains which include five subtests. The GRASSP was designed to capture information on
upper limb neurological impairment for individuals with tetraplegia. The GRASSP defines
neurological status with numerical values, which represent the deficits in a predictive pattern,
is reliable and valid as an assessment technique, and the scores can be used to determine
relationships between impairment and functional capability of the upper limb. The GRASSP
is recommended for use in the very early acute phases after injury to approximately one year
post injury. Use of the GRASSP is recommended when a change in neurological status is
being assessed.
iv
ACKNOWLEDGEMENTS The journey of completing my PhD has not been without tremendous support from my colleagues, mentors, friends and family. I am grateful to the many individuals in the Allied Health and Neurosurgery Departments at the Toronto Western Hospital who have provided me with an extraordinary work environment, which has supported me and my endeavors over the past years. We do not always realize what our true capabilities are, however, often others recognize that,
“Everything you need is already inside” -Bill Bowerman
My committee and supervisor have provided me with great support, guidance and encouragement. It is with their help I have realized my true potential. For this I would like to Thank each member of my committee: Dr. Milos Popovic for providing the knowledge and resources to assist in commercializing this work; Dr. Dorcas Beaton for guiding the process and analysis and helping in developing the GRASSP project from the beginning to the end; and Dr. Michael Fehlings for providing great knowledge of the field and providing opportunities which allowed the GRASSP project to grow. My supervisor Professor Molly Verrier has not only inspired me throughout my PhD studies but throughout my entire professional and academic career. Molly has provided for me the tools to build foundations as a clinician and researcher, has helped me grow through this journey and has taught me, how to learn. Molly was well aware of my potential long before I was. I would like to thank the GRASSP International Research and Design Team who placed their faith in me, to lead the GRASSP project and for their support and provision of resources. Particularly, Dr. Armin Curt who supported my development and growth throughout this project. Also, Depeng Jiang who provided assistance with analysis of the data. I am blessed to have two wonderful extended families. Both my husbands and my own, parents and siblings have supported our family so that I was able to succeed in this endeavor. And finally to my family, I am forever grateful; my husband Timothy, who is my rock and the one who never once stopped supporting my goals and endeavors; and my children Alanzo, Maya and Saphira who have given me the most important things in life. Thank you for helping me attain my goals and providing me with the strength and support to finish what I had started. I would also like to acknowledge the Toronto Rehabilitation Institute Student Scholarship Fund, Ontario Neurotrauma Foundation, Christopher and Dana Reeve Foundation, Rick Hansen Foundation, SCI Solutions Network and the Physiotherapy Foundation of Canada for the academic and research support for the GRASSP project.
v
DEDICATION
To my parents, who everyday set an example for me and instilled in me the key to success as being “hard work and only hard work” and that if you believe in something commit yourself to it. “No one ever attains very eminent success by simply doing what is required of him; it is the amount and excellence of what is over and above the required that determines the greatness of ultimate distinction.” -Charles Kendall Adams
vi
TABLE OF CONTENTS
Abstract ii Acknowledgements iv Dedication v Table of Contents vi List of Tables x List of Figures xii List of Abbreviations xiii Glossary of Terms xiv List of Appendices xviii
CHAPTER I: INTRODUCTION AND LITERATURE REVIEW
1.1 Introduction 1
1.2 Literature Review 2
1.2.1 Epidemiology of Spinal Cord Injury 2
1.2.2 Classification of the SCI population 3
1.2.3 Recovery and Function after Tetraplegia 5
1.2.3.1 Neurological Recovery 6
1.2.3.2 Functional recovery 7
1.2.4 The Components and Function of the Upper Extremity and Relevance
in Tetraplegia 8
1.2.5 Available Outcome Measures for Spinal Cord Injury 14
1.2.6 Development of a Measure for Clinical Research 19
1.2.6.1 Rationale for a New Measure 19
vii
1.2.6.2 Purpose and the Conceptual Framework of a Measure 20
1.2.6.3 Psychometric Characteristics of Measures 20
1.3 The Graded Redefined Assessment of Strength Sensibility and Prehension
(GRASSP) 22
1.3.1 Conception and Development of the GRASSP 22
1.3.2 Description of the GRASSP 26
CHAPTER II: RESEARCH OBJECTIVES & HYPOTHESES
2.1 Significance and Purpose 27
2.2 Research Objectives and Hypotheses 27
2.2.1 Development of the Scoring System 28
2.2.2 Psychometric Testing 29
2.2.3 Confirmation of Theoretical Framework 30
CHAPTER III: GENERAL METHODS
3.1 Recruitment of Sites and Examiners 31
3.1.1 Training Workshops 31
3.2 Study Protocol 32
3.2.1 Outcome Measures 32
3.2.2 Sample Size 33
3.2.3 Inclusion and Exclusion Criteria 33
3.2.4 Study Participant Recruitment 33
3.2.5 Study Design 34
3.2.6 Data Management 34
viii
3.3 Sample 35
CHAPTER IV: RESULTS PART 1
4.1 Development of the Scoring Approach for the Graded Redefined Assessment of
Strength Sensibility and Prehension (GRASSP) 39
4.2 Abstract 39
4.3 Introduction 41
4.4 Methods 45
4.5 Results 49
4.6 Discussion 54
CHAPTER V: RESULTS PART 2
5.1 The Graded Redefined Assessment of Strength Sensibility and Prehension
(GRASSP) – Reliability and Validity 59
5.2 Abstract 59
5.3 Introduction 60
5.4 Methods 65
5.5 Results 67
5.6 Discussion 73
CHAPTER VI: RESULTS PART 3
6.1 Sensory and Motor Relationships to Prehension and Upper Limb
Function 78
6.2 Abstract 78
ix
6.3 Introduction 80
6.4 Methods 83
6.5 Results 89
6.6 Discussion 96
CHAPTER VII: GENERAL DISCUSSION
7.1 Summary of Findings 100
7.1.1 Objective One: Development of the Scoring System 101
7.1.2 Objective Two: Psychometric Testing 102
7.1.3 Objective Three: Confirmation of Theoretical Framework 102
7.2 Impact, Significance and Strengths of the Research 103
7.3 Limitations of the Project 106
7.4 Future Directions 108
7.5 Final Summary 109
REFERENCES 110
APPENDICES 126
x
LIST OF TABLES
Table 1: Outcome Measures Used for SCI and Upper Limb Assessment 16
Table 2: Components of the Initial GRASSP and Methods of Administration 25
Table 3: Phases of Development, Objectives and Outcomes for the GRASSP Project 26
Table 4: Description of the Sample and Breakdown of Data received from Participating
Sites (presented by n values) 35
Table 5: AIS Classification of European (n=27) and North American (n=45) Sub Samples
(presented in percent and n values) 36
Table 6: ISCSCI Motor and Sensory Neurological Levels of European (n=27) and North
American (n=45) Sub Samples (presented by percent and n values) 36
Table 7: Demographic, Comparator Measure and GRASSP Results for Total
Sample n=72 38
Table 8: Example of Guttman Scaling Using Scores of Palmar Sensation Items 48
Table 9: Demographics of the Sample Based on ISCSCI sensory and motor levels and AIS
Classification, n=72, (presented as n values and percentages of sample) 50
Table 10: Portion of Sample which presented with and without a Cumulative Predictive
Pattern for each Subtest Total Score in the GRASSP (presented by n value and
Percentage) 51
Table 11: ISCSCI and SCIM benchmarks for GRASSP Discriminative
Score Ranges 53
Table 12: Reliability Values of Subtest Scores Within the GRASSP 69
Table 13: Construct Validity Agreeement/Discordance of Sensory Results
Between GRASSP and ISCSCI (n=72) 71
xi
Table 14: Level of Agreement Between GRASSP-SWM and ISCSCI-LT
for C6, C7 and C8 71
Table 15: Construct Validity Agreement/Discordance of Strength Results
Between GRASSP and ISCSCI (n=72) 72
Table 16: Concurrent Validity of GRASSP Subtests and Functional Measures 73
Table 17: GRASSP Scoring Details (Subtest and Item Scores and Ranges) 83
Table 18: Demographics of the Sample Based on ISCSCI and AIS Classification,
n=72, (presented as n values and percentages of sample) 90
Table 19: Correlation Matrix of all Variables Included in the Multiple Linear Regression
and Structural Equation Modeling Analysis 91
Table 20: Unadjusted Multiple Linear Regression Values for Sensibility,
Stength and Prehension 92
Table 21: Attributes of the GRASSP Version 1.0 104
xii
LIST OF FIGURES
Figure 1: Theoretical Framework for the Construct of Sensorimotor Upper Limb
Function used for the Development of the GRASSP 23
Figure 2: Frequency Histograms Describing ISCSCI Sensory and Motor Levels and AIS
Classification For Total Sample (n=72) 37
Figure 3: GRASSP (Components) Version 1.0 42
Figure 4: A Visual Representation of GRASSP Subtest Scores 57
Figure 5: Summary of GRASSP Components 64
Figure 6: Histograms of Sample Distribution 68
Figure 7: Model D, Hypothetical Model of Impairment and Sensorimotor Upper Limb
Function in Tetraplegia to be tested with Structural Equation Modeling 89
Figure 8: Results of Multiple Linear Regression of Impairment Variables and
Upper Limb Function 93
Figure 9: Model D, Structural Equation Model for Impairment and Sensorimotor
Upper Limb Function after Tetraplegia 95
xiii
LIST OF ABBREVIATIONS
AIS: The American Spinal Injury Associations Impairment Scale
CFI: Comparative Fit Index
CUE: Capabilities of Upper Limb Function Questionnaire
FES: Functional Electrical Stimulation
GRASSP: Graded Redefined Assessment of Strength, Sensibility and Prehension
ICC: Intra class correlation coefficient
ISCSCI: International Standards of Neurological Classification of Spinal Cord Injury
ISCSCI-ML: International Standards of Neurological Classification of Spinal Cord Injury
Motor Level
ISCSCI-SL: International Standards of Neurological Classification of Spinal Cord Injury
Sensory Level
RMSEA: Root Mean Standard Error of Approximation
SCI: Spinal cord Injury
SCIM: Spinal cord Independence Measure
SCIM-SS: Spinal cord Independence Measure, Selfcare Subscale
SS: Somatosensory Stimulation
SRMR: Standardized Root Mean Square
TLI: Tucker Lewis Index
xiv
GLOSSARY OF TERMS
AIS: Refers to the classification of spinal cord injury as defined by the American Spinal Injury Association’s Impairment Scale, and is used in conjunction with the ISCSCI motor and sensory levels (A – sensory and motor complete, B – sensory incomplete, C – motor incomplete, D – sensory and motor incomplete). Anatomic Injury: Location of damage to structures in the spinal colum including ligaments, bones, vessels, cartilage and nerves. Capabilities of Upper Extremity Questionnaire (CUE): Self perceived measure of upper limb function specific to tetraplegia (Marion et al., 1998). Compensation: A method to perform an act by using alternative materials and measures in comparison to the normal method to perform that same act. Component: A factor which contributes to a larger concept. Construct: The ideas and concepts which define the basis of a well designed measure. Dermatome: The area of skin on the body which is mainly supplied by a single spinal nerve. Dichotomous Variables: A variable which is derived from only two response levels, “0” or “1”. Dorsal Sensation: The first subtest in GRASSP which assesses three locations on the dorsal surface of the hand, with a score range between 0 to 12. Element: An aspect which is one of many in relationship to a larger concept.
Feedback Control: The situation in the neurological system when output from previous events influence occurrences of the same phenomenon in the present or future.
Feedforward Control: The mechanism related to the neurological system that monitors performance inputs rather than outputs, and provides information to maintain a specified state, thus preventing or minimising problems, and refining movement. Function: An ability to perform goal oriented tasks and self-care tasks independently in any environment for any purpose. Functional Recovery: To regain loss in function in the direction of returning to the original state. Global Function: Ability to perform tasks independently which enable one to provide care for personal well being including tasks related to the whole body.
xv
Guttman Scaling: The procedure to determine whether a set of items can be rank-ordered in a unidimensional scale (Guttman, 1944; Guttman, 1950). Impairment: A reduction in strength and/or sensation and/or prehension in comparison to the normal neurological state, within the work of this thesis. Independence: The ability to perform functional tasks without the assistance from another individual. Integrated Systems: The incorporation of all the systems which function collectively to perform movement and function. The International Standards for Neurological Classification of Spinal Cord Injury (ISCSCI): A classification measure which defines severity of spinal cord injury according to spinal cord level determined by testing of dermatomes and myotomes. Latent Trait Variable: The unobserved variable, indicated or estimated by other observed variables which is included in path analysis. Level of Impairment: The amount of impairment defined by a measure or value. Level of Injury: The most caudal level, at which sensory and motor spinal cord function are fully intact. Manipulation: Skillful and controlled movements of the five digits and palm of the hand incorporating fine motor control and sensibility. Myotome: A group of muscles innervated by a single spinal nerve. Neural Repair: The restoration or replacement of damaged neural structures. Neurological Integrity: The status of the neurological state specifically in reference to sensation and strength of the upper limb, within the work of this thesis. Neurological Recovery: To regain loss of neurological status (impairment) in the direction of returning to the original state. Palmar Sensation: The second subtest in GRASSP which assesses three locations on the palmar surface of the hand, with a score range between 0 to 12. Path Analysis: One type of structural equation modeling which evaluates the directional dependencies among a set of variables. Polychotomous Variables: A variable which is derived from more than two response levels. Prehension: The act of gripping or grasping something with the hand.
xvi
Prehension Ability: The fourth subtest of the GRASSP which assesses one’s ability to generate three specific prehension patterns, with a score range of 0 to 12. Prehension Performance: The fifth subtest of the GRASSP which assesses one’s ability to perform six tasks and rated according to whether that individual is able to incorporate the expected prehension pattern, with a score range of 0 to 30. Reach: The extension of the arm by way of the shoulder and elbow to approach a target for the hand to perform prehension and manipulation. Recovery: To regain loss. Response Item: The individual component/s that are measured within a test. In the case of GRASSP the response items are the individual muscles, or sensory locations measured. Scaling: The assignment of objects to numbers according to a rule. Semmes Weinstein Monofilaments: Monofilaments of variable stiffness that apply different levels of grams of force used to quantify cutaneous sensation. Sensation: The perception associated with stimulation of a sense organ Sensibility: Functional sensation of the hand, usually associated with palmar sensation. “Sensibility”: Common sense aspects of clinical measure. Sensorimotor: Sensory and motor integration in the context of movement. Spinal Cord Independence Measure (SCIM): Measure of global independence specifically designed for individuals with spinal cord injury (Catz et al., 2004). Spinal Cord Independence Measure Selfcare Subscore (SCIM-SS): Grooming and selfcare subscale within the SCIM which includes the assessment of items related to use of the upper limbs to perform a range of activities of daily living (Catz et al., 2007). Strength: The third subtest of the GRASSP which assesses the muscle force of ten muscles in the arm and hand, with a score range of 0 to 50. Structural Equation Modelling: General, powerful multivariate analysis technique used to determine causal relationships among variables. Tetraplegia: Dysfunction of all four limbs due to the insult or injury to the spinal cord in the cervical region of the spine, including levels from the occiput down to thoracic vertebrae 1. Trait: A distinguishing feature. Traumatic: A physical injury caused by trauma.
xvii
Upper Limb: The complex which includes the shoulder, elbow and hand.
xviii
LIST OF APPENDICES
Appendix 1 126
Appendix 2 160
Appendix 3 162
Appendix 4 165
Appendix 5 180
1
CHAPTER I: INTRODUCTION AND LITERATURE REVIEW
1.1 Introduction
Neurological recovery of the upper extremity after tetraplegia is the main factor
which has been influencing the need for a precise, detailed measure of upper limb
impairment specific to tetraplegia. Classification of spinal cord injury (SCI) is established for
individuals by the International Standards for Neurological Classification of Spinal Cord
Injury (ISCSCI) which includes the American Spinal Injury Association’s Impairment Scale
(AIS) (Marino, 2000); this measure incorporates assessment of the trunk and four limbs and
designates severity of injury with respect to the whole body. Thus, researchers have not been
able to determine subtle neurological change in the upper limb using the ISCSCI because it
was not designed to be specific to the upper limb. As a result of the void for an upper limb
impairment measure specific to tetraplegia, the Graded Redefined Assessment of Strength,
Sensibility and Prehension (GRASSP) was developed.
The GRASSP is an assessment strategy with three domains that provides a detailed
profile (integrated sensorimotor function) of the upper limb for individuals with tetraplegia,
both at a single time point and longitudinally. The International GRASSP Research and
Design Team were brought together in May of 2006 by the Christopher and Dana Reeve
Foundation to design and develop the GRASSP (see Appendix 1).
The research presented in this thesis relates to the development of the scoring approach,
psychometric testing, and the relationships between upper limb impairment and function. The
work presented in this thesis falls within the larger GRASSP project. Chapter 1 provides a
literature review of the pertinent issues related to measurement of the upper limb after
tetraplegia and reviews previous development of the GRASSP. Chapter 2 provides the
rationale, objectives and hypotheses of the study. Chapter 3 presents the general methods
2
section which details the data collection process and sampling. Chapters 4, 5 and 6 present
the results in the form of three manuscripts. Chapter 7 presents a summary and general
discussion. Initial development of the GRASSP including the: theoretical framework, item
generation and item reduction has been reported in a previous manuscript (Kalsi-Ryan et al.,
2009).
1.2 Literature Review
1.2.1 Epidemiology of Spinal Cord Injury
The incidence of traumatic SCI worldwide is approximately 22 persons/million,
essentially creating 130,000 new spinal cord injured individuals yearly. The prevalence is
estimated at 2.5 million and the average age at time of injury is 33.4 years. Fifty percent of
injuries are caused by motor vehicle accidents and yearly health care costs from four nations
(Canada, USA, Australia, and England) total approximately $10 billion dollars per annum
(Campaign for Cures, 2010). A recent systematic review reported that there is a general trend
world wide towards increased incidence in the elderly due to falls and non-traumatic SCI
(van den Berg, et al., 2010). The International Campaign for Cures of spinal cord injury
Paralysis (ICCP) provides the best possible global data; however, the majority of the data is
garnered from developed countries (Campaign for Cures, 2010). Not only does this prevent
the true international incidence and prevalence of SCI from being captured, the data also fails
to impress the significance and enormous strain that SCI places on society and the economy.
In 2001/2002 the incidence of SCI in Canada was reported to be approximately
1050/year or 35 persons/million (Rick Hansen Registry, 2005). Although there is no exact
figure, it is estimated that 41,000 people live with a SCI in Canada.
3
SCI is most prevalent in age groups below 50 years, meaning that individuals will
live close to 40 years with enormous care needs. Not only do these persons require a
significant amount of care, they require it for prolonged periods of time which in turn
generates larger lifetime care costs than other diseases such as stroke. Although Canadian
statistics report that approximately 50% of SCI are cervical injuries, reports from more recent
international reviews including literature from Canada (Wyndaele and Wyndaele, 2006)
report an increase in the ratio of individuals with tetraplegia and incomplete injuries to be
above 50%.
In Canada it is estimated that the lifetime care costs for an individual with paraplegia
is $1.25 million; and for an individual with tetraplegia is around $25 million. Although the
incidence and prevalence of SCI is low in comparison to other diseases, the care costs are
enormous; approximately $1.5 billion per year is spent on managing SCI by the Canadian
Health Care System (Rick Hansen Registry, 2005). With such a significant impact on those
with injury, and the individuals who provide care for people with tetraplegia, across a
lifetime; restoring function is imperative.
1.2.2 Classification of the SCI population
Traumatic SCI encompasses injuries to the spine and spinal cord, which can occur
anywhere along the spinal column (cervical, thoracic, and lumbar) as a result of trauma.
Injuries to the thoracic and lumbar spine result in paraplegia while injuries to the cervical
spine result in tetraplegia. The significant difference between these two general classes of
injury is that an individual with paraplegia would retain normal upper limb function;
however, the level of injury determines whether or not the individual retains postural control.
On the other hand, individuals with tetraplegia not only lose the function of the lower body,
4
but also function of the upper limbs which are vital in maintaining independence after a SCI.
Depending on the level of injury the degree of upper limb impairment varies; an injury to the
higher levels of the cervical spine will cause greater impairment in the upper limbs. The
degree of impairment is reduced as the level of injury moves from the occiput towards T1.
When a spinal cord injury occurs the structures including: bone, ligaments, cartilage,
vessels and nerves are damaged by the trauma. The location of the actual damage to the
structures is called the anatomic injury or spinal injury which should be defined separately
from the neurological injury (Middendorp et al., 2010). For example an individual may
sustain a C5 fracture and dislocation, which would be a C5 anatomic injury. The damage that
occurs to the spinal cord may or may not be at the same level. Although, spinal cord damage
can be measured by MRI imaging, maximal spinal cord compression, lesion length and
spinal cord hemorrhage which do correlate with severity of SCI and prognosis according to
the ISCSCI (Miyanji et al., 2007) are not widely used. The neurological level of injury that
occurs as a result of trauma based on spinal cord integrity defined by the impairment at the
periphery is the most universally used method to define SCI. Thus, the neurological level of
injury is defined by the most caudal level of normal neurological intactness.
SCI is defined by the neurological level of injury which is usually defined by the
most caudal level at which sensory and motor function are fully intact according to the
ISCSCI. Classification of the SCI is defined by the AIS which is part of the ISCSCI and
defines whether the injury is complete or incomplete. A complete SCI occurs when there is
no motor or sensory function preserved in the sacral segments S4-S5 (A). There are three
classes of incomplete SCI. An incomplete injury occurs when there is sensory but not motor
function preserved below the neurological level and includes the sacral segments S4-S5 (B),
or motor function is preserved below the neurological level, and more than half of the key
5
muscles below the neurological level have a muscle grade less than three (C), or motor
function is preserved below the neurological level, and at least half of the key muscles below
the neurological level have a muscle grade of three or more (D) (Marino, 2000). Based on an
international review of SCI epidemiology, the prevalence of completeness is approximately
40% complete and 60% incomplete (Wyndaele and Wyndaele, 2006).
In addition to the neurological level of injury and the classification of completeness,
there is a zone of partial preservation where there is partial integrity of the spinal cord that
does not contribute to the neurological level assigned by ISCSCI.
Traditionally, individuals with complete SCI do not gain any motor or sensory
function below the neurological level of injury, whereas individuals with incomplete injuries
can due in part to the zone of partial preservation. This essentially results in spontaneous
recovery of sensory and motor deficits. However, the degree of recovery and the way in
which the recovery can influence function varies significantly from individual to individual,
and is not always predictable. Therefore, two elements contribute to a heterogeneous
tetraplegic population: the first is the difference in upper limb intactness secondary to the
neurological level of injury and classification and the second is the varying degree of
recovery. The variance in the tetraplegic group creates many complexities in developing an
upper limb measure of impairment.
1.2.3 Recovery and Function after Tetraplegia
The term recovery refers to the changing state (improvement) of neurophysiology
and/or functional capacity. After tetraplegia, even if an injury is complete it is common to
anticipate some type of recovery. In the case of complete injuries, where no neurological
change is expected, the anticipated recovery is functional. In the case of incomplete injuries
6
the anticipated recovery is both neurological and functional. Neurological recovery refers to
changes in impairment, while functional recovery refers to a change in ability; in both cases,
depending on the severity of injury either type of recovery, or both, can be expected.
1.2.3.1 Neurological Recovery
Outside of neural repair there is evidence to support neurological recovery. Motor
recovery of key muscles does occur in the upper limb after complete and incomplete cervical
SCI injuries. Varying degrees of local recovery can be expected in tetraplegic individuals.
Motor recovery can be predicted by the acute motor intactness early after injury (Ditunno et
al., 2000; Ditunno et al., 1987; Fisher et al., 2005). It has also been noted that gains in
strength of elbow and shoulder muscles can continue to occur up to 15 months post discharge
from rehabilitation (Drolet et al., 1999). Individuals with preservation of sensation to pin
prick in a motor segment with grade zero power, showed an 85% chance of motor recovery
to at least grade three (Poynton et al., 1997). Some of the recovery in strength may be
attributed to plasticity within the spinal cord. Corbetta et al. reported that even with greater
than 75% loss in spinal cord area, functional neural connections that traverse the site of cord
injury still exist (Corbetta et al., 2002). Sensory evoked potentials performed in the acute
phase of cervical SCI are indicative of level of SCI and have been found to be predictive for
recovery of hand function (Curt and Dietz, 1996). Although, the regeneration of nerve fibers
is limited in the adult central nervous system, changes in function occurring for several years
after injury can be dependent on the reorganization of circuits which are spared by the lesion
(Green et al., 1999) and cortical reorganization (Jurkiewicz et al., 2007; Raineteau and
Schwab, 2001; Bruehlmeier et al., 1998; Cohen et al., 1991; Levy et al., 1990). With the
cortical reorganization that does occur after SCI, it is theorized that the sensorimotor cortex
7
plays a critical role in the recovery of function. Therefore, the spared sensation at time of
injury which is often followed by motor recovery (Poynton et al., 1997) may partially be
related to the changes that occur in the sensorimotor cortex and reorganization of tracts.
Measurement of global neurological recovery has been demonstrated with the ISCSCI
(Marino et al., 2003), which is a classification measure used to define severity of SCI and
was not designed to be specific to the upper limb. As it does not include sensory testing for
the palmar surface of the hand and it measures the strength of a limited number of muscles of
the upper limb and hand. Therefore, a more comprehensive sensorimotor profile of the upper
limb was necessary to report the subtle changes that occur post tetraplegia and to determine
how these subtle changes could affect one’s level of functional independence.
1.2.3.2 Functional recovery
Recovery of function and independence has also been documented for individuals
with tetraplegia. Individuals post-tetraplegia improve their ability to function after a period of
rehabilitative intervention (Ota et al., 1996), and often continue to improve once discharged
into the community. When improvement of function is documented, both complete and
incomplete groups of SCI show improvement (Welch et al., 1986). This would indicate that
an individual who would make no progress in neurological status can make progress in
functional status, due to skill acquisition or adaptation (Müslümanoğlu et al., 1997). Function
in the cervical SCI population is often assessed by measures of
function/independence/burden of care, which are not all SCI specific. The Functional
Independence Measure (FIM) (Keith et al., 1987); is a generic measure used for many
populations with diverse disabilities and measures burden of care. There are also two
measures of function specific to SCI; the Quadriplegia Index of Function (QIF) (Marino et
8
al., 1999), and the Spinal Cord Independence Measure (SCIM) (Catz et al., 2004). These
measures rate an individual’s level of function with and without assistance and assistive
devices. Although the QIF and FIM results are related to functional recovery over time, both
measures require more sensitivity to be discriminative (Marino et al., 1993) as compensation
may be the factor influencing the changes. Strength in specific muscle groups can predict
greater functional outcomes; elbow flexion, shoulder flexion and wrist extension all correlate
higher with tasks in both the FIM and the QIF (Beninato et al., 2004; Marciello et al., 1995).
Specifically shoulder strength and elbow extension strength show stronger associations to
functional independence than other muscle groups in the upper limb (Fujiwara et al., 1999;
Welch et al., 1986).
Neurological and functional recovery are well evidenced in the SCI literature,
however, it is predominantly based on global function and impairment as defined by the
ISCSCI, FIM and QIF. There remains little evidence that defines the recovery of upper limb
impairment in tetraplegia.
1.2.4 The Components and Function of the Upper Extremity and Relevance in
Tetraplegia
The key actions of upper extremity function are considered to be reach, grasp and
manipulation skills (Shumway-Cook and Woollacott, 2007). The systems theory of motor
control predicts that there are specific neural and musculoskeletal subsystems that contribute
to the control of these three components.
Voluntary performance of an upper limb function is controlled by complex neural
circuits in the brain inter connecting the sensory and motor systems. The sensory, motor and
motivational systems each have anatomically and functionally distinct subsystems (whereby
9
hand manipulation and reach are not controlled by precisely the same subsystem components
in an individual) that perform specific tasks during the performance of voluntary movement
(Shumway-Cook and Woollacott, 2007), such as reaching and grasping. The decision to
initiate a voluntary movement is regulated by the motivational system which acts directly on
the motor system (Kandel et al., 2000). Concurrently the sensory system begins processing
information regarding the targeted task and begins sending information to association areas
of the cortex where movement is planned. The processed sensory information is transmitted
to the motor system. The motor system then executes and regulates the movement based on
continuous sensory information. Direct and indirect pathways relay motor control
information. The direct system regulates the activity of the motor neurons that innervate
muscles of the hand and arm involved in the fine control of movement. The indirect system
plays an important role in the overall regulation of body posture and stability. The indirect
motor system includes synaptic relays, whereas the direct motor system does not.
The general areas of the cerebral cortex which are critical in the control of reach,
grasp and manipulation are the primary motor cortex, the somatosensory cortex, premotor
cortex and areas of the posterior parietal lobe (Filimon, 2010). The cerebellum is also
important for feedforward and feedback control. Afferent fibers carry somatosensory
information from the periphery in the dorsal columns on the same side until the medulla
where the cells synapse in the dorsal column nuclei and cross to the other side and ascend to
the thalamus. In the thalamus sensory input is relayed to the primary sensory areas of the
cerebral cortex as well as information regarding motor behaviour to the motor areas of the
cortex. Axons from the primary motor cortex project directly to motor neurons in the spinal
cord via the corticospinal tract which is a descending tract in the ventral column of the spinal
cord. The descending corticospinal tract crosses to the opposite side of the spinal cord in the
10
medulla just caudal to the dorsal nuclei. Approximately 10% of the fibres remain ipsilateral
until termination in the spinal cord. Corticospinal axons terminate on groups of motor
neurons in the spinal cord that innervate specific limb muscles. The corticospinal tract
primarily controls distal muscles that are important for precise movements, such as those of
the hand. Other motor pathways, which originate in the brainstem nuclei mediate postural
adjustments during movement (for a detailed discussion of this subject please see Kandel et
al., 2000, Chapter 19 for more detail).
Upper limb function is generated by the anatomical structures of the arm. The
anatomy of the upper limb includes five main joints: the sternoclavicular joint, the shoulder
joint, elbow joint, the wrist joint and hand joints. The associated musculature that surrounds
these joints are the muscles: of the shoulder girdle, the upper and lower arm; and the wrist
and hand (Moore, 1985; Netter, 1989). The muscles and overlying skin of the upper limb are
innervated by the anterior and dorsal horns which emerge from the cervical spinal cord
between the fourth cervical vertebrae and the first thoracic vertebrae known as the C5 to T1
nerve roots. Nerve roots which emerge on the right side innervate the right side only and
similarly for the left side. The peripheral representations of the nerve roots for muscles are
known as the myotomes and of the skin are known as the dermatomes. Each myotome and
dermatome has specific representations at the periphery. The representation of the myotomes
and dermatomes, although defined can vary from individual to individual. Specifically,
proximal muscles of the upper limb are innervated by C5 and as the muscles become more
distal, the innervation moves caudally to T1 (see Moore, 1985 for more detail). Sensation of
the upper limb is represented by the nerve roots between C5 and T1. The hand is specifically
represented at the levels of C6, C7 and C8. Due to the specific representations of the spinal
cord at the periphery via myotomes and dermatomes, impairment measured at the periphery
11
is reflective of the neurological integrity at the cord level. Thus the anatomy is often used as
the basis for neurological assessment at the periphery.
Upper limb function is comprised of three components which contribute to
performance of tasks; reach, prehension and manipulation (Shumway-Cook and Woollacott,
2007). Stability is provided by the proximal structures of the shoulder girdle which enable
the arm to move freely and reach to be a form of transport, which is produced by the shoulder
and elbow joint in association with the musculature and innervation that surround these two
joints. These two joints transport the hand to the location of the task. Prehension is produced
by the wrist and hand structures with the associated joint and skin sensory receptors and
musculature, incorporating the actions of grasp and release. Manipulation is produced by the
sensibility of the palmar surface and motor control of the small muscles of the hand and the
associated joints to produce an extraordinary organ which can perform fine movements and
functions (Brand and Hollister, 1991; Shumway-Cook and Woollacott, 2007). Therefore,
assessment of upper limb impairment requires a measure which acknowledges the
components (sensory and motor) as well as upper limb function (integration).
Not only do the upper limbs perform the functions for which normal individuals use
their hands, but they replace the functions for other parts of the body that may no longer have
even partial function (i.e. leg function for walking, replaced by wheelchair propulsion). The
upper limbs for an individual with tetraplegia represent all self-care activities including
locomotion, bowel and bladder function, recreational activities and vocational activities.
Improvement in upper limb function after cervical SCI is one of the most significant factors
in improving quality of life according to individuals with tetraplegia (Anderson, 2004;
Snoek, 2004). Therefore, the more normal and precise the recovered upper limb function is
after tetraplegia the more functional the individual will be. The field of upper limb
12
restoration research is quite new; interventions have only been studied with some emphasis
since the 1960’s due to the fact that survival of high tetraplegics was poor prior to this era
(Lamb and Chan, 1983).
As a result of the emphasis placed on upper extremity function by the SCI population
and society and the increasing potential for optimizing recovery (increased proportion of
individuals with incomplete SCI), intense interest has been generated in the development of
therapies for preserving and restoring upper limb function. Furthermore, the degree of injury
along with the spinal cord structures affected also contributes to the varying degrees of
presentation of impairment and potential for recovery. Therefore, interventions for the
improvement of upper limb function not only need to incorporate compensatory therapeutic
strategies but also restorative interventions that reduce the impairment. The potential to
restore lost upper limb function in SCI is influenced by three factors: 1) a higher rate of
incomplete injuries (Sekhon and Fehlings, 2001; Marino et al. 1999), which will tend to
increase the rates and magnitude of recovery; 2) development of therapeutic interventions
that are applied at the periphery to affect the sensorimotor system (Popovic et al., 2006;
Beekhuizen and Field-Fote, 2005); and 3) development of interventions that are applied to
the central nervous system such as pharmacological agents and biologics with the potential
for neural repair, neuro-protection and regeneration (Schwab et al., 2006; Wells et al., 2003;
Baptiste and Fehlings, 2007).
The potential to target rehabilitative approaches towards specific muscle groups that
have a known potential to recover, and have a greater impact on functional independence is
knowledge available from the current literature that reports and defines recovery (Ditunno et
al., 2000; Drolet et al., 1999; Fujiwara et al., 1999; Welch et al., 1986). Although, the
literature is not all upper limb specific it does provide an understanding of the changing
13
neurological integrity (motor) post SCI, which assists in framing clinical decision making for
rehabilitative approaches for the different levels of tetraplegia.
There is scientific evidence to support the efficacy of stimulation therapies (including
functional electrical stimulation (FES), and somotosensory stimulation to the median nerve
(SS)). Stimulation therapies can be used as assistive technologies or can be applied as
therapeutic interventions to influence the sensory and motor integrity of the system. Assistive
technologies are beneficial for individuals with C5, C6 or C7 complete tetraplegia once
neurologically stable. These technologies are not therapeutic interventions; rather they are
assistive devices that improve hand function when in place and active (Alon et al., 2003;
Taylor et al., 2002; Peckam et al., 2001; Popovic et al., 1999). Both SS and FES can be
applied as interventions which can improve the voluntary control and strength of the hand
musculature. There is significant evidence to support SS as an adjuvant to rehabilitation
therapies, particularly massed practice. SS in association with a comprehensive rehabilitation
program that is functionally oriented results in improved sensation, motor function,
functional ability and cortical changes related to the trained function (Beekhuizen and Field-
Fote, 2008; Hoffman and Field-Fote, 2007; Beekhuizen and Field-Fote, 2005). FES as an
intervention during the sub-acute phase of recovery has proven to be more effective than
conventional rehabilitation alone when administered as a therapy in conjunction with
functional tasks (Popovic et al., 2006; Miller et al., 2008).
Stimulation therapies are beneficial not only as assistive technologies for the
neurologically stable individual, but also as therapeutic interventions for individuals with
complete or incomplete tetraplegia during the rehabilitation phase of recovery (4wks to 18
months post injury) (Kalsi-Ryan and Verrier, 2010).
14
There is less evidence to support the practice of tendon transfer surgery as an
intervention which can be provided once an individual is neurologically stable. Laffont et al
makes a strong argument that with more active joints the greater one’s ability becomes to
grasp, release and manipulate an object (Laffont et al., 2007). Although, some researchers
feel there is adequate research to support tendon transfer surgery, the evidence for tendon
transfer lacks rigor (Vastamaki, 2006; Rothwell et al., 2003; Meiners et al., 2002; Lo et al.,
1998). One of the barriers for successful research related to tendon transfers is the lack of
appropriate and well developed outcome measures.
As a result of the potential spontaneous recovery and the state of upper limb
restoration research in SCI, the need for a specific upper limb impairment measure in
tetraplegia is paramount. According to Wilson et al. the effectiveness of clinical interventions
can only be optimized if the causal relationships between the clinical variables and outcomes
are made, and measures are the link between these two concepts (Wilson et al., 1995). An
upper limb impairment measure for tetraplegia will enable researchers and clinicians to
achieve a degree of precision in the assessment for this specific area. It will also enable an
understanding, an ability to evaluate, and differentiate characteristics of upper limb
impairment in a quantitative manner.
1.2.5 Available Outcome Measures for Spinal Cord Injury
In light of the neurological recovery that does occur post tetraplegia, the necessity to
assess impairment in the form of neurological integrity is crucial. Measuring the upper limb
with a specific and sensitive measure will allow researchers to evaluate the subtle change
occurring in the upper limb secondary to natural recovery and interventions. Although a
number of SCI specific and upper limb (some tetraplegia) specific measures exist,
15
impairment of the upper limb (the construct of interest) is not measured with enough
precision by existing tests. Curtin et al. reported that there was inconsistency in evaluation
and documentation within the rehabilitation therapy community due to inappropriate use of
outcomes measures. The best suited measures required to answer research questions were not
used. Furthermore, the sensitivities of the available measures has not been sufficient to
determine significant but small changes (Curtin et al., 2005). Dunn et al. stated it was crucial
that measures have specificity in order to detect small but clinically significant improvements
in hand function (Dunn et al., 2008).
Many of the measures used to date have not been specifically designed for use in the
SCI population, (Dunn et al., 2008; van Tuijl et al., 2002) and, of the measures that are
specific to SCI, many of the psychometric properties are not well established. Identifying the
measures that are reliable, valid and specific to the tetraplegic upper limb is the key element
to performing successful research in the field. Use of functional tests such as the FIM, SCIM
and QIF alone are not sufficient to determine the efficacy of neurophysiological based
interventions; in that they do not identify the changes in neurological integrity. Although
common measures across interventions and timeframes should be employed, specificity of
the intervention and the phase of recovery also determine the choice of outcome measures.
Table 1 describes the available and commonly used measures in SCI and, their applications,
constructs and psychometric properties if available.
16
Table 1: Outcome Measures Used for SCI and Upper Limb Assessment Construct Validity Measure and Description Construct and
Population Mode InR ItR
TrT KG ConcV Res Qualities
Impairment International Standards of Neurological Classification of SCI (ISCSCI) Classification measure, used to determine severity of injury. (Marino et al., 2003; Priebe and Waring, 1991)
Whole Body SCI
M 35%-93%
▲ ■ ♦
Grip and Pinch Dynamometry Measure of grip force (Mathiowetz et al., 1984)
Hand Strength Peripheral Hand Injury
M *0.98 *0.99 ● ♦
International Classification for Surgery of the Hand Upper limb motor strength test and two point discrimination of digits (McDowell et al., 1986)
Upper Limb Motor Tetraplegia
M ■ ● ♦
Function The Van Lieshout Test Upper limb capacity seated in a wheelchair (Post et al., 2006; Spooren et al., 2006)
Upper Limb Function Tetraplegia
M *0.98 *0.87 with Grasp and Release
*R ▲ ■ ● ◘
Grasp and Release Test Functional test specific to ability after neuroprosthetic implantation (Mulcahey et al., 2004)
Hand Function Tetraplegia
M *0.87 ▲ ■ ● ◘
Capabilities of Upper Extremity Questionnaire Self perception of functional ability with upper limbs (CUE) (Marino et al., 1998)
Upper Limb Function Tetraplegia
SR *0.96 *0.94 *0.74 with FIM
▲ ■ ● ◘
The Jebsen Hand Function Test Generic hand function test, outcome of time (Jebsen et al., 1969; Sears and Chung, 2010; Beebe and Lang, 2009)
Hand Function Hand Injury
M ● ◘
The Sollerman Hand Function Test Hand Function Test (Sollerman and Ejeskar, 1995)
Hand Function Musculoskeletal
M *0.98 R ▲ ● ◘
Action Research Arm Test (ARAT) Upper limb reach and grasp test. (Lin et al., 2009)
Upper Limb Function Stroke
M *0.92 *0.97 *0.81 with WMFT
● ◘
Wolf Motor Function Test (WMFT) Hand Function Test (Lin et al., 2009)
Hand Function Stroke
M *0.92 *0.97 *0.81 with ARAT
● ◘
Fugl-Meyer Hand Subtest Hand Function Test (Fugl-Meyer et al., 1975; Lin et al., 2009)
Hand Function Stroke
M *0.92 *0.97
*0.81 with ARAT
● ◘
17
Global Function/Independence Functional Independence Measure A generic outcome which measures burden of care. (Keith et al., 1987; Kidd et al., 1995)
Independence All disabilities
O 0.70
The Spinal Cord Independence Measure II (SCIM) Measure of global independence (Catz et al., 2004; Itzkovich et al., 2007)
Independence SCI
O *0.94 *0.79 with FIM
▲ ■ ◘
Quadriplegia Index of Function A measure of global function (Marino et al., 1999)
Independence Tetraplegia
O ■ ◘
Lamb and Chan Questionnaire ADL inventory specific to tendon transfer (Lamb & Chan,1983)
Independence Tetraplegia
SR ■ ● ◘
Legend for Table 1: InR – Inter rater reliability, ItR – Intra rater reliability, TrT – Test retest reliability, KG – Known Groups, ConcV - Concurrent Validity, Res – Responsiveness. Mode is the manner by which the measure is administered: M-clinician administers measure and rates individual, O-clinician observes activity of an individual and rates the individual, SR-self reported measure. Psychometric Properties: * - inter rater reliability above 0.8 intra class correlation coefficient, intra rater or test retest reliability above 0.8 intra class correlation coefficient, concurrent validity above 0.7 correlation with a comparator measure, construct validity established, R-responsiveness established at acceptable level Qualities = Qualities of the Measure: ▲- Psychometric properties evaluated with the SCI population, ■ – Construct of measure specific to SCI, ● – Construct of measure specific to the upper limb, ♦ - Construct of measure specific to impairment, ◘ - Construct of measure specific to function
18
While it appears that there are a number of available measures, impairment of the
upper limb is not the defined construct for any of the available assessments. In summary the
measures used often with the SCI population are either too global in nature (incorporate the
whole body), designed for different populations, assess a construct other than impairment of
the upper limb, do not include sensory testing specific to the hand, or do not possess the
psychometric properties required (see Table 1). Measuring function and independence are
very important, however, a sound understanding of the core and integrative elements of
neurological integrity which underlie any level of function are also important to understand
because neurological integrity is the factor being targeted with treatment and how it changes
is of interest. The body of literature established by researchers and clinicians in the field of
“Hand Rehabilitation” definitively places a significant emphasis on impairment as well as
functional assessment. The reason being that the core impairments, and how they relate to
the performance of functional tasks is understood (Lundborg and Rosen, 2007; Mackin et al.,
2003) and known to be a key concept in therapeutic interventions. Sensation plays a
significant role in recovery of function. As such training protocols specifically addressing the
sensorimotor relearning process substantially increase the possibilities for improved
functional outcome after nerve repair. Based on the anatomy of the upper limb and the
integrative nature of refined movement; assessment of core elements (sensation and strength)
and integrative elements (performance of functional tasks) of impairment are essential in
making connections between the state of impairment or residual neurological integrity and
function.
19
1.2.6 Development of a Measure for Clinical Research
In order to create a measure that will hold strong psychometric properties once
completed the initial development must be sound. The various steps involved in
measurement development should be included in the process to ensure the measure fulfils its
intended purpose.
1.2.6.1 Rationale for a New Measure
Prior to initiating the process of developing a measure, it is necessary to determine
the rationale for a new measure. “The clearest rationale for a new measure is an absence of a
scale directed toward the phenomenon of interest” (Portney and Watkins, 2000). The
phenomenon of interest in this work was, “the assessment of upper limb impairment as a
result of tetraplegia”. The ISCSCI is one measure of impairment for SCI; however, it is a
classification measure for the whole range of SCI, making it non-specific to the upper limb.
The International Classification for Surgery of the Hand (McDowell et al., 1986) is a detailed
motor and sensory assessment of the hand and forearm which has been designed to screen
and classify individuals specifically for tendon transfer surgery; its use for any other purpose
within tetraplegia has not been reported. Grip and pinch dynamometry are not sufficient to
describe all of the impairment that occurs in the whole upper limb. There are a number of
scales specific to the upper limb (not all specific to SCI) which are commonly used to assess
individuals with tetraplegia (Post et al., 2006; Mulcahey et al., 2004; Jebsen et al., 1969;
Sollerman and Ejeskar, 1995; Lin et al., 2009). However, the construct for these functional
tests is hand function or upper limb function, precluding their use for the assessment of
neurological integrity of the upper limb. Therefore, the rationale for developing the GRASSP
20
was based on the absence of an impairment measure specific to the upper limb after
tetraplegia.
1.2.6.2 Purpose and the Conceptual Framework of a Measure
The initial work required to design a measure includes establishing a purpose and a
conceptual framework which is used as the basis for measure development (McHorney,
1999). The purpose of the measure defines the phenomenon that is being assessed. In the
case of the GRASSP, the aim was to assess impairment of the upper limb after tetraplegia.
The conceptual framework (Figure 1) that was developed, defined the construct,
“sensorimotor upper limb function” and the domains that were contained within (McDowell
and Jenkinson, 1996). The theoretical framework provided the roadmap for compiling the
GRASSP, which guided the processes of item generation and reduction (McHorney, 1999;
Streiner and Norman, 1995). Section 1.3 defines the theoretical framework designed for the
development of the GRASSP. Once the actual measure was constructed, the GRASSP was
then tested for sensibility, reliability, validity and ultimately is being tested for
responsiveness.
1.2.6.3 Psychometric Characteristics of Measures
“Sensibility” is the aggregate of properties that make up the common sense aspect of
an instrument including: face and content validity (Rowe and Oxman, 1993). Feinstein
defined that the dimensions of sensibility should include: comprehensibility, replicability,
suitability of scale, ease of usage, face validity, content validity and scale purpose (Feinstein,
1990). “Sensibility” is established by having experienced clinicians administer the measure
and then complete a sensibility questionnaire (see Appendix 2). The results of the
21
questionnaire provide the required information for modifications to the test. A sensibility
assessment of the GRASSP was conducted during the training workshop in Toronto (January
2007) and the questionnaire results were incorporated into modifications made to the initial
GRASSP which rendered GRASSP Version 1.0. A measure that has sensibility is practical,
assesses the intended population, is easy to administer and demonstrates a comprehensible
results.
Reliability of a test is when the same measurement is obtained when repeated by the
same rater (intra rater) or different rater (inter rater). Test retest reliability replaces intra rater
reliability when the repeated tests are conducted more than three days apart. (Norman and
Streiner, 1995; Portney and Watkins, 2000). Reliability is considered to be a basic and
essential quality of a scientific measure. Scientific investigation can only be performed with a
reliable measure, as a change in measurement can more confidently be attributed to clinical
change; and the number of individuals required in a clinical trial can be reduced. Reliability
of a measure is established when a measure is administered repeatedly to the same stable
individual by the same and/or different examiners. The agreement among results renders the
reliability of a measure. Most commonly agreement statistics or intra class correlation
coefficients (ICC) are used for analysis. Reliability is a necessary step in establishing the
usefulness of measurement; however, it alone is not sufficient.
Validity is when a measure assesses what it is “intended to measure”. Face validity
exists if the measure appears to assess what it is intended to measure, and content validity
exists if the components represent the construct that the measure is intended for. Face and
content validity are established by a review of the measure, usually by an expert or a panel of
experts; this is often incorporated into the “sensibility” questionnaire. Construct validity is
usually established by comparing the new measure to an existing measure which assesses the
22
construct of interest. However, if a comparable measure does not exist, other available
measures in the field of interest can be used (Streiner and Norman, 1995). Unlike the analysis
of reliability a wide variety of approaches can be used to establish construct validity.
1.3 The Graded Redefined Assessment of Strength Sensibility and Prehension
(GRASSP)
The Graded Redefined Assessment of Strength Sensibility and Prehension (GRASSP)
is a multi-domain clinical impairment measure of the upper limb for individuals with
tetraplegia (Appendix 1).
1.3.1 Conception and Development of the GRASSP
Figure 1 defines the theoretical framework conceived to assemble the GRASSP.
Integration of systems includes: cognition, sensory and motor substrates which are controlled
by feedforward and feedback processing. The integration is the “medium” with which
functions of the upper limb are performed (Kandel et al., 2000). Sensorimotor control and
sensorimotor learning principles are part of the framework as they are the “process” by which
the body learns and maintains refined movement (Shumway-Cook and Woollacott, 2007).
When these two concepts of the framework are considered they produce the construct of
“senorimotor upper limb function”. The construct of “sensorimotor upper limb function”
being an integrated construct is comprised of three domains, in this case sensation, strength
and prehension. Sensation and strength are the two core elements of impairment and
prehension is the integrated element of impairment. In order to specify the framework for
tetraplegia, characteristics related to the spinal cord segments and recovery after injury were
incorporated. In order to connect impairment to function; characteristics of the task oriented
23
approach were incorporated into the framework as well. Consideration of integration and
processing, principles, task orientation and SCI pathology were incorporated into the
theoretical framework to ensure the elements contributing to impairment would be
components of the measure developed.
Subtests of the GRASSP were selected based on how well they assessed the domains
within the construct. Items within the subtests were selected to ensure that the three
components (reach, prehension and manipulation) of upper limb movement were captured
(Shumway-Cook and Woollacott, 2007; Kapandji, 1970; Brand and Hollister, 1991). The
domains (sensation, strength and prehension) represent core and integrated elements of
impairment. The purpose of incorporating an integrated domain was to provide the
opportunity to assess how sensation and strength impairments contributed to an integrated
(prehension) function which may be increasingly important in understanding the recovery
process.
Figure 1: Theoretical Framework for the Construct of Sensorimotor Upper Limb Function used for the Development of the GRASSP
Figure 1 illustrates the concepts, principles and structures associated with upper limb function. All elements contribute to the construct “sensorimotor upper limb function” and the domains within, sensation, strength and prehension.
24
Items for the GRASSP were generated from existing tests and measures which were
reviewed to determine suitability. The theoretical framework for the GRASSP (Figure 1)
provided a conceptual roadmap to guide initial item generation. In addition to the theoretical
framework suitability for test selection was based on three rationales: a inter and intra rater
reliability of greater than 0.8, evidence of development of the test and availability of
normative data and the tests ability to measure the underlying construct of the domain. Parts
from the Link Hand Function Test (Link, 2004), The Tetraplegia Hand Measure (Kalsi-Ryan,
2006), generic assessment methods for peripheral hand injury (Mackin et al., 2003) and the
Sollerman Hand Function Test (Sollerman and Ejeskar, 1995) were selected, modified and
incorporated to create the GRASSP.
Components adapted from the Link Hand Function Test included five prehension
tasks. In a similar fashion, the sensory module, part of the motor testing, and the scoring
scale from the prehension tasks were adapted from the Tetraplegia Hand Measure. All
components included in the GRASSP are presented in Table 2. The sensibility domain
included Semmes Weinstein monofilaments (SWM) for light touch and static two-point
discrimination (S2PD) for functional sensation. The strength/tone domain was assessed using
manual muscle testing (MMT) for strength (Daniels and Worthingham, 1995; Kendall et al.
1993) and the Modified Ashworth Scale for tone (Bohannon and Smith, 1993). Both
descriptive and performance-based prehension tests were incorporated to address the
prehension domain. The prehension ability test evaluates whether the thumb and digits can
assume three specific grasps or can perform any active movement at all. The prehension
performance test is a modified version of the Sollerman Hand Function Test (Sollerman and
Ejeskar, 1995). The Sollerman was modified by Link and Kalsi-Ryan et al. during the
development of the Link Hand Function Test and the Tetraplegia Hand Measure. The
25
prehension domain in the GRASSP retains the Sollerman concept of evaluating specific
activity of daily living tasks performed with specific grasps for evaluation.
Table 2: Components of the Initial GRASSP and Methods of Administration Components of the GRASSP Method for administration Test
Details
Rationale
Position time required
How to test
Sensibility domain: test sites selected by dermatome Light Touch/ SWM (Mackin et al., 2003)
6 palmar/dorsal test sites
Inter/intra-reliability = 0.965
Supine/ sitting, 10 min
Apply monofilaments to all test locations Summate the score for each hand separately
Static 2 Point Disc (Mackin et al., 2003)
3 palmar test sites
Inter/intra-reliability = 0.989
Supine/ sitting, 5 min
Apply stimulus to all test locations Summate score for each hand
Strength and tone domain: muscle selection based on myotomes Strength (Cuthbert and Goodheart, 2007)
MMT-4 arm & 7 hand muscles
Inter reliability = 0.880
Supine/ sitting, 10 min
Assess each muscle and grade Summate all scores for each hand
Tone (Bohannon and Smith, 1993)
Modified Ashworth for hand & arm
Inter-reliability = 0.750
Supine/ sitting, 5 min
Assess elbow and hand for flexor/ extensor tone
Prehension domain: segmental influence movement pattern Prehension Ability
3 grasps rated on scale of 0–4
Supine/ sitting, 5 min
Have subject perform grasps and rate
Prehension Performance (Sollerman and Ejeskar, 1995)
5 grasps/6 tasks rated on scale of 0–5
Adapted from Sollerman, inter-reliability = 0.980
Sitting, 15 min
Set patient up in sitting at table and have patient perform all 6 tasks for each hand separately
The first phase of item reduction was based on pilot testing and clinical expertise
which rendered the initial GRASSP (Appendix 1 and Table 2), and was used in the cross
sectional study. Further item reduction was conducted by using psychometric analysis to
determine relevance of items and subtests. This stage of item reduction eliminated static two
point discrimination testing, testing for tone, and abductor pollicis brevis from the strength
items. This step rendered GRASSP Version 1.0 (see Appendix 1).
26
The GRASSP project includes the entire process associated in the development of the
measure. Table 3 defines the development for each phase.
Table 3: Phases of Development, Objectives and Outcomes for the GRASSP Project Development Phase Objectives Outcomes Phase I – Theoretical Development and Pilot Testing
- Development of the Theoretical Framework - Assembly of the GRASSP Items - Pilot Testing - Item Reduction Based on Linear Regression
Initial GRASSP
Phase II – Cross sectional Study
- Development of Scoring System - Inter/Test Retest Reliability and Construct and Concurrent Validity - Confirmatory Analysis of the Theoretical Framework
- Scoring System Developed and Included in GRASSP Version 1.0 - Inter/Test Retest Reliability, Construct and Concurrent Validity Established - Theoretical Framework Proven to be a Robust Model for GRASSP Development
Legend for Table 3: Bolded sections represent phase of GRASSP Project included in the PhD thesis and conducted by Sukhvinder Kalsi-Ryan.
1.3.2 Description of the GRASSP
The purpose of assembling the GRASSP was to develop a clinical research tool that
could characterize upper limb impairment from the cervical (CO-T1) SCI population, obtain
isolated and integrated sensory and motor impairment data, and define level of impairment
according to the upper limb and spinal cord lesion. The GRASSP was required to: a) be
highly responsive (sensitive) to change over time; b) assess the extent of spontaneous
(natural) recovery; and c) be applicable for use in clinical trials to evaluate the effect of new
therapies. The GRASSP is intended to measure change in neurological status from the acute
to chronic phases of recovery after tetraplegia. The GRASSP is a three domain impairment
measure which contains five subtests and measures each hand separately.
27
CHAPTER II: RESEARCH OBJECTIVES & HYPOTHESES
2.1 Significance and Purpose
Upper limb function is of great importance to individuals with SCI (Anderson, 2004;
Snoek, 2004. Specific and precise methods must be implemented to establish efficacy of
methods available to restore upper limb function. However, a challenge that researchers face
is the lack of a precise impairment measure specific to the upper limb after SCI.
Development of the GRASSP was initiated to create an impairment measure to fill this void.
The availability of well developed measures specifically designed to assess the “phenomenon
of interest” is imperative to measure the influence of new interventions. The research
conducted to develop the GRASSP has occurred in two Phases as per Table 3. Phase I
included: development of the theoretical framework, compilation of the items proposed for
the initial GRASSP and pilot testing. Phase II included: item reduction to render GRASSP
Version 1.0, development of the scoring system, psychometric testing, and confirmatory
analysis to determine the design of the theoretical framework.
Items bolded in Table 3 refer to work done by the PhD candidate Sukhvinder Kalsi-
Ryan independent of the larger GRASSP Research and Design Team and items not bolded
represent work done by the GRASSP Research and Design Team collectively. Phase II
represents the work done and included in the PhD thesis.
2.2 Research Objectives and Hypotheses
The content of this thesis establishes the scoring approach and sets initial benchmarks
for levels of GRASSP scores, establishes the psychometric properties of reliability and
validity, and presents relationships of impairment to upper limb function. Additional
objectives and outcomes related to development of the GRASSP are identified in
28
Table 3, however, not addressed in this thesis.
2.2.1 Development of the Scoring
The output of a measure is intended to define the phenomenon desired to be measured, in this
case impairment of the upper limb for individuals with tetraplegia. Essentially the score of a
test defines what the state of the phenomenon is. The score or values derived by a measure
should be understandable and represent the state of what they are intended to measure
accurately, so that comparisons with repeated measurement and comparator measures can be
made. If a score is to be cumulated from the items tested, an understanding of what the
cumulated score represents should also be defined. A total score that presents with Guttman
characteristics (Guttman, 1944; Guttman, 1950) predicts item response by knowing only the
total score. Therefore, establishing and understanding the scores of the GRASSP and what
they represented was imperative for the results to be meaningful. Chapter IV presents the
analysis, and results related to the development and explanation of the scoring approach. The
objectives of the analysis in Chapter IV were to:
1) Investigate the cumulative pattern predictability (Guttman pattern) of the items
within each subtest;
2) To identify the discriminative GRASSP score ranges and the associated ISCSCI
levels, AIS classification and SCIM selfcare subscale scores (SCIM-SS); and
3) To determine a method to report GRASSP scores.
The hypotheses were:
H1: Guttman characteristics would not be a dominant pattern of the items
in the GRASSP subtests;
29
H2: Discriminative GRASSP scores would not define the sample similarly to ISCSCI
levels, AIS classification and SCIM-SS scores; and
H3: A total GRASSP score would not represent meaningful findings assessed by the
GRASSP subtests.
2.2.2 Psychometric Testing
Psychometric properties of a measure are fundamental qualities which are essential to
establish for a measure to be used in clinical trials and in clinical settings. Therefore,
establishing the reliability and validity of the GRASSP was a necessary process to undertake
in order to qualify the measure for use in clinical and research settings. The objectives of the
analysis presented in Chapter V were:
1) To establish the inter rater and test retest reliability;
2) To establish construct validity using the ISCSCI as a comparator; and
3) To establish concurrent validity with the SCIM and CUE.
Therefore, in Chapter V it was hypothesized that:
H1: The reliability (inter rater and test retest) for all subtests within the GRASSP
would demonstrate an intra class correlation coefficient of at least of 0.80;
H2: The GRASSP sensation and strength subtests would be significantly more
sensitive and precise in defining neurological integrity at the periphery than the ISCSCI
motor and sensory levels; and
H3: The concurrent validity of the GRASSP would be defined by moderate (above 0.6
Pearson correlation coefficient) and significant associations to the CUE, SCIM and SCIM-SS
(this would indicate, as neurological integrity decreases function improves).
30
2.2.3 Confirmation of the Theoretical Framework
Assessment of impairment is important to determine the neurological status, but also
important for understanding the relationships between impairment and function. The analysis
presented in Chapter VI explores the relationships between the impairment domains
(sensibility, strength and prehension) as defined by the GRASSP to a latent trait variable of
“sensorimotor upper limb function” and observed functional values (SCIM-SS and CUE) that
represent upper limb function. The objective of the analysis presented in Chapter VI was:
1) to determine the influence of sensibility, strength and prehension impairment on
sensorimotor upper limb function.
Therefore, it was hypothesized in Chapter VI that:
H1: a relationship between sensation, strength and function could be demonstrated
through general linear modeling and structural equation modeling, determining the degree
and type of relationship would provide information to further develop therapeutic
interventions.
31
CHAPTER III: GENERAL METHODS SECTION
3.1 Recruitment of Sites and Examiners
All centres were engaged by the members of the GRASSP International Research and
Design Team. Seven centres from Europe and North America were involved in Phase II of
the GRASSP project. Each site engaged two examiners, either occupational or physical
therapists that had experience with the treatment of individuals with SCI. Ethical approvals
were obtained for all sites through local research ethics boards. Fourteen clinicians in total
were engaged in the study (12 occupational therapists and 2 physical therapists).
3.1.1 Training Workshops
All examiners in Europe were trained at a workshop in Zurich, at University Hospital
Balgrist and all examiners in North America were trained at a workshop in Toronto, at the
Toronto Rehabilitation Institute. Workshops consisted of general information regarding the
rationale and study protocol (how the protocol was to be implemented, processes involved in
data management and communications with the central site for assistance and
reimbursement). Demonstration and information regarding the administration of the
GRASSP, ISCSCI, AIS, SCIM and CUE were provided. The examiners were also provided
with the opportunity to perform all measures in the study on individuals with SCI during the
workshop. The workshops were conducted by Sukhvinder Kalsi-Ryan (PhD candidate) in
Toronto and Claudia Rudhe (Clinical Study Nurse) in Zurich.
32
3.2 Study Protocol
The GRASSP was the measure of interest and the comparator measures were the
ISCSCI, SCIM and the CUE. The general methodology of the study was designed to address
the testing of reliability and validity of the GRASSP (see Appendix 3 for details regarding
study design).
3.2.1 Outcome Measures
The ISCSCI was selected for use in the study to define the sample according to an
international classification method and to define the severity of injury for individuals
involved in the study. Being the only true measure of impairment for SCI the ISCSCI is the
most meaningful and widely used measure to define SCI and was used as a comparator
measure to establish validity (further details re: ISCSCI are available in Appendix 4). The
SCIM was incorporated into the study to be used as a comparator measure; although, it is a
global measure of independence it is SCI specific and is a well established measure in the
field. The SCIM is a disability scale that has been specifically developed to evaluate the
functional outcomes of patients with traumatic and non-traumatic SCI. The SCIM assesses
function in three core areas: 1) Self-care, which includes feeding, bathing, dressing and
grooming, and is scored between a range of 0 to 20; 2) Respiration and sphincter
management are scored between a range of 0 to 40; and lastly 3) Mobility, also scored
between a range of 0 to 40 (further details re: SCIM are available in Appendix 4) (Catz et al.,
2004). The SCIM was used to provide data to confirm the relationships between impairment
and function. The CUE is a 32 item questionnaire developed to assess difficulty in
performing certain activities with the upper extremities. The CUE is a self perceived measure
of upper limb function which incorporates components of reach, grasp and manipulation. The
33
scoring is based on degree of difficultly in performing tasks which is an important domain in
defining functional limitation (Marino et al., 1998). The CUE was selected as a comparator
measure to establish validity and determine relationships between impairment and self
perceived function (further details re: CUE are available in Appendix 4). Properties of the
measures are available in Table 1.
3.2.2 Sample Size
A sample size of 40 individuals was calculated based on an estimated ICC of 0.80 or
greater with an alpha of 0.05. The minimum estimated sample size was 39 for 3 repeated
measures according to Donner & Eliasiw’s estimation curves (Donner and Eliasiw, 1987). As
a result of the multi centre collaboration a greater number of datasets were collected
increasing the sample size to 72 allowing for analysis of item reduction, scoring approach
and impairment/function models.
3.2.3 Inclusion and Exclusion Criteria
Individuals with chronic traumatic tetraplegia who were neurologically (community
dwellers) and medically stable, between the ages of 16 and 65 and able to provide informed
consent were included in the study. Individuals with moderate brain injury who were
neurologically unstable or individuals with any pathology other than tetraplegia causing
upper limb impairment were excluded.
3.2.4 Study Participant Recruitment
Study participants were recruited by the examiners or research coordinators at each
respective site. Study participants were recruited from outpatient programs, programs where
34
individuals with chronic tetraplegia were admitted to hospital and the community via
advertisement with SCI specific associations. Potential study participants who met the
inclusion criteria were enrolled consecutively. All participants were screened, informed, and
consented by the research coordinator or examiners.
3.2.5 Study Design
The study protocol in Europe was slightly different from the protocol conducted in
North America. All sites in Europe conducted a protocol which rendered data for inter rater
reliability and construct and concurrent validity. The study participant was assessed twice
with the GRASSP and once with the comparator measures. During the first visit, examiner
one administered the GRASSP. During the second visit examiner two administered the
GRASSP and examiner one administered the comparator measures.
In North America two examiners were engaged at each site. The study participants
were scheduled for two visits. During the first visit examiner one administered the GRASSP
followed by examiner two repeating the administration of the GRASSP. During the second
visit examiner one repeated administration of the GRASSP and examiner two administered
the comparator measures. Examiners were assigned as one and two at each site (see
Appendix 3).
3.2.6 Data Management
All data was de-identified prior to paper entry. All data was initially recorded onto
paper copies of the measures for each study participant. Data collected in Europe were
packaged and couriered to the central site in Toronto, where all data were entered
electronically by the same individual. Data collected in North America were abstracted by
35
the respective examiners and entered electronically. All electronic data were transferred to
the central site. All data were pooled to render 72 datasets in total and stored on a secure
server. All analyses were conducted using SPSS 17.0, SAS 9.1 and M-Plus 5.2. Table 4
presents the sample, location of data collection and inclusion for analysis. All data were
collected between January 2007 and September 2007.
Table 4: Description of the Sample and Breakdown of Data received from Participating Sites (presented by n values)
North America
Inter rater
Reliability(n)
Test Retest
Reliability(n)
Construct Concurrent
Validity (n)
Confirmatory Analysis for Theoretical Framework
(n) 1. Rehabilitation Institute of Chicago (Chicago, USA)
10 10 10 10
2. Toronto Rehabilitation Institute (Toronto, Canada)
15 15 15 15
3. Vancouver Coastal Health (Vancouver, Canada)
10 10 10 10
4. Thomas Jefferson University (Philadelphia, USA)
10 10 10 10
Europe 5. University Hospital Balgrist (Zurich, Switzerland)
9 9 9
6. Krakenhaus Hohe Worte (Bayreuth, Germany)
8 8 8
7. Traumacenter Murnau (Murnau, Germany)
10 10 10
Total n Used for Analysis 72 45 72 72
3.3 Sample
A total of 72 datasets were collected. Forty five individuals were enrolled in North
America and 27 were enrolled in Europe. All data was combined for inter rater reliability and
validity analyses. Test retest reliability was established with North American data only.
Tables 5 and 6 describe the European and North American samples separately according to
AIS Classification and ISCSCI sensory and motor neurological levels. Demographic data
36
regarding gender, age and time post injury were incomplete across samples therefore; only
ranges for the demographics of age and time since injury were available. However, all data
elements required for all aspects of analysis were consistent among both samples; there were
no missing data elements. The total sample is described according to the ISCSCI sensory and
motor levels and the AIS by frequency histograms in Figure 2. The five figures represent the
motor and sensory levels for both left and right and the classification of injury. Demographic
data and results on comparator measures are summarized in Table 7.
Table 5: AIS Classification of European (n=27) and North American (n=45) Sub Samples (presented in percent and n values) A
% (n) B
% (n) C
% (n) D
% (n) Europe 40(11) 15(4) 19(5) 26(7) North America 37(17) 31(14) 16(7) 16(7) Legend for Table 5: AIS – ASIA Impairment Scale, A – Complete, B – Sensory Incomplete, C – Motor Incomplete, and D – Sensory and Motor Incomplete Table 6: ISCSCI Motor and Sensory Neurological Levels of European (n=27) and North American (n=45) Sub Samples (presented in percent and n values) C2toC4
%(n) C5
%(n)C6
%(n) C7
%(n) C8
%(n) T1
%(n)T2toS5%(n)
Europe 22(6) 15(4) 30(8) 15 (4) 7(2) 11(3) ISCSCI Motor Neurological Level North
America 16(7) 16(7) 31(14) 24(11) 4(2) 9(4)
Europe 48(13) 15(4) 15(4) 7(2) 4(1) 4(1) 7(2) ISCSCI Sensory Neurological Level North
America 49(22) 16(7) 31(14) 2(1) 2(1)
Legend for Table 6: ISCSCI – International Standards of Neurological Classification for Spinal Cord Injury, C2 to S5 – the neurological level according to ISCSCI
37
Figure 2: Frequency Histograms Describing ISCSCI sensory and motor levels and AIS Classification for Total Sample (n=72)
Legend for Figure 2: ISCSCI – International Standards of Neurological Classification for Spinal Cord Injury, C2 to S5 – the neurological level according to ISCSCI, AIS – ASIA Impairment Scale, A – Complete, B – Sensory Incomplete, C – Motor Incomplete, and D – Sensory and Motor Incomplete
38
Table 7: Demographic, Comparator Measure and GRASSP Results for Total Sample n=72
Demographic Mean SD Range Median
Age * in years 39.7 10.7 16 - 65 42 Time Post Injury *
in years 7.6 6.1 6 months –
20 yrs 6.2
Comparator Measure
SCIM (0-100) 0 – 100 45.1 21.1 11 - 99 38 SCIM-SS 0 – 20 9.8 5.7 0 - 20 9 CUE 0 – 124 78.8 29 4 - 124 78
GRASSP R L R L R L R L Strength 0 – 50 24.3 25.1 13.0 13.5 5-50 1-50 18 19 Dorsal Sensation 0 – 12 6.5 6.7 3.2 3.1 0-12 0-12 6 6 Palmar Sensation 0 – 12 7.1 7.2 3.6 3.3 0-12 0-12 7 7 Prehension Ability 0 – 12 4.9 5.1 4.5 4.3 0-12 0-12 2 2 Prehension Performance
0 - 30 15.6 14.7 9.6 8.9 0-30 0-30 12 12
Legend for Table 7: SD – Standard Deviation, R – right, L – left, *values based on n=42 SCIM – Spinal Cord Independence Measure, SCIM-SS Spinal Cord Independence Measure Self care Subscale, CUE – Capabilities of Upper Extremity Questionnaire
39
CHAPTER IV: RESULTS PART 1
4.1 DEVELOPMENT OF THE SCORING APPROACH FOR THE GRADED
REDEFINED ASSESSMENT OF STRENGTH SENSIBILITY AND PREHENSION
(GRASSP)
4.2 Abstract
GRASSP is a clinical impairment measure designed to assess the upper limb
following tetraplegia. Three domains within the GRASSP include five subtests when
administered represent neurological integrity. Defining the numerical values representative of
the spinal cord deficits based on anatomy and defining the predictability of scores through
rank-ordering of the items within subtests was required to give the scoring meaning.
Objectives: 1) To establish the frequency of predictable item response for each subtest
within the GRASSP; 2) to establish the discriminative GRASSP score ranges and compare to
the ISCSCI and SCIM-SS; and 3) to derive a method to present GRASSP scores. Methods:
A cross sectional study was conducted; all participants were tetraplegic and neurologically
stable (n=72). The study protocol included administration of the GRASSP, ISCSCI and
SCIM during the same session. Analysis: Guttman scaling was applied to all items in each
subtest to determine cumulative predictive patterning. The sample was organized by subtest
total (ascending order) for each subtest, in order to establish ranges of GRASSP scores.
Results: All subtests within the GRASSP demonstrated a significant degree of predictable
item response based on the Guttman matrices; predictable item response ranged between
78% and 90% for the five subtest total scores (dorsal sensation, palmar sensation, strength,
prehension ability and prehension performance). GRASSP score ranges were defined
according to spinal cord levels (rostral to caudal) which were not congruent with the
40
discrimination of levels derived by the ISCSCI sensory and motor testing, however, similar
to the discrimination of scores derived by the SCIM-SS total. The five subtest scores were
cumulative in nature, therefore, a cluster bar graph of subtest total scores was used to present
GRASSP results. Conclusion: Items within each of the five subtests can be added and
defined by subtest total scores. The subtest total scores are interpreted to mean more
impairment or less impairment when moving from a lower numeric value to a higher numeric
value respectively. Subtest total scores can be interpreted by way of a cluster bar graph.
41
4.3 Introduction
It is recognized that there is a deficiency of adequate outcome measures for the
assessment of impairment of the upper limb for individuals with tetraplegia (Dunn et al.,
2008; Miller et al., 2008; van Tuijl et al., 2002; Curtin et al., 2005). Due to this deficiency
many researchers are faced with the challenge of using inadequate outcomes to follow
neurological recovery post tetraplegia. Due to the fact that impairment is not usually the
construct measured in existing assessments, definitive relationships between impairment and
function remain a phenomenon that is not well understood.
Impairment of the upper limb is extremely important to understand and measure in
order to progress the development of new restorative techniques. In order to measure upper
limb impairment there are a number of components and subcomponents which need to be
considered as part of the measure. There are specific neural and musculoskeletal subsystems
which contribute to the control of upper limb function (Shumway-Cook and Woollacott,
2007). Impairment of the upper limb must be measured by multiple domains to be
comprehensive. Assessing the impairment of the upper limb with precision is the first step in
determining the contribution of impairment to function. Thus, the representation of
impairment should define the construct in this case “sensorimotor upper limb function” with
numeric values which have context.
The ISCSCI (Marino, 2000) is a diagnostic measure used to define the severity of
SCI, by defining the most caudal level of normal neurological spinal cord function. The
ISCSCI does not measure the upper limb comprehensively, particularly sensation of the
palmar surface of the hand. As ISCSCI was not intended to be used as a measure of upper
limb impairment, a new approach was required. Therefore, in an effort to respond to this gap
in available impairment outcome measures, the GRASSP was developed (Kalsi-Ryan et al.,
42
2009). The GRASSP is a clinical measure of upper limb impairment which can be
administered in a clinical setting and incorporates three domains that are vital to upper limb
function: sensibility (sensation), strength (motor) and prehension. The GRASSP includes
Semmes Weinstein Monofilament (SWM) testing on both the dorsal and palmar surface of
the hand, motor testing of 10 muscles of the hand and arm, performance of three prehensile
positions (grasps) and performance of six task oriented prehension skills. Figure 3 presents a
summary of the domains, subtests and items of the GRASSP.
Figure 3: GRASSP (Components) Version 1.0
Figure 3 presents the domains, 5 subtests, items and scoring.
The GRASSP was developed to contain more than one domain to accommodate the
multiple elements related to upper limb impairment in tetraplegia. Three domains are
43
represented by five subtests with each subtest containing up to ten items. Right and left hands
are tested separately, resulting in five subtest results for each hand. Because the response
range for items among subtests is not consistent, defining the meaning of GRASSP subtest
scores through the process of scaling was required in order to determine the optimal and most
meaningful method of presenting the responses. The separate domains (sensation, strength
and prehension) are not intended to have internal consistency, instead they have definitive
contributions and interactions with one another as well as to overall upper limb function.
Figure 1 in Chapter 1 presents the theoretical framework which defines the concepts,
theories, construct and domains considered when designing the GRASSP. Therefore, adding
scores across subtests for a global score would not represent the multi dimensional construct
of “sensorimotor upper limb function” adequately. Attention to the scoring system and best
possible options for use are significant in development of the measure as a whole, however,
the meaning of individual subtest total scores is of greater significance to defining
neurological deficit than a single global score. The manner in which the scoring approach is
defined should reflect the deficit as close as possible (Samejima, 1969). When there are,
numerous domains within a single measure and numeric values are used to represent
neurological integrity, development and explanation of the scoring approach is a
requirement.
Measurement refers to the assignment of objects to numbers according to a rule
(Stevens, 1946). Guttman scaling (Guttman, 1944; Guttman, 1950) is one method used in
developing a measure. The Guttman scaling methodology was applied to each subtest for the
interpretation of the numeric values derived by the response items, in order to establish a
continuum for the concept that was being measured. Essentially, determining a scale to have
Guttman characteristics means that the total score is predictive of the order of item response
44
(Trochim, 2010). Guttman scaling is traditionally applied to dichotomous variables to
determine whether a set of items can be rank-ordered on a one-dimensional scale. However,
the concept can be applied to polychotomous variables (van Schuur, 2003) as well. The
process creates a scale where, when the second item response is positive, the preceding
response would also be positive. With polychotomous variables the process creates a scale
where, when the response is at a specific level, each preceding variable has a response of the
same level or close to it. There is some variation due to the complexity of polychotomous
variables. However, establishing a polychotomous scale to show a Guttman pattern means
that the scale predicts item response based on the total score. The total cannot be predictable
of item response 100% of the time. However, awareness of the frequency of a predictable
pattern provides insight to the examiner as what a subtest total score should mean and what it
could mean. Rasch analysis is one statistical method that can be used to determine the
predictable pattern of item response in a unidimensional scale (Norquist, et al., 2004),
however, a rather large sample size is required to conduct this analysis. Mokken Scale
Analysis is another statistical method that can be used with both dichotomous and
polychotomous variables to determine predictability of item response, again a large sample
size (at least n=200) (van Abswoude et al., 2004) is required for this analysis. Therefore,
Guttman scaling was performed manually.
Known group’s method is a typical method to support construct validity and is
provided when a test can discriminate between groups of individuals known to have differing
levels/severity of a trait (Portney and Watkin, 2000), in this case severity of SCI. In the initial
work done to establish known groups for the GRASSP, ranges of GRASSP subtest scores
were benchmarked against the ISCSCI sensory and motor levels and SCIM-SS. The purpose
of this step was to define the place of GRASSP scores against a known measure of
45
impairment (ISCSCI) in SCI and a known functional measure (SCIM-SS) specific to SCI.
The ISCSCI assigns impairment according to the most caudal “normal” neurological level;
any sensory and motor function below the designated level is not accounted for except by the
zone of partial preservation which is also a sensory or motor level. Therefore, the ISCSCI
does not provide a true representation of impairment measured over the body. Using ISCSCI
to identify levels of GRASSP scores would not provide distinct groupings, only ranges,
whereas the GRASSP impairment measure accounts for the partial neurological status
between neurological levels, by presenting the deficit as a numeric value which represents
impairment or lack there of in the upper limb.
Based on the available measures in the field and the available sample, the scoring
system was developed with Guttman scaling and benchmarking ranges of GRASSP scores to
well known standards used in the field. The three objectives addressed in this chapter were:
1) to establish the frequency of predictable item response for each subtest within the
GRASSP; 2) to define the meaning and ranges of GRASSP subtest total scores according to
the ISCSCI and SCIM-SS; and 3) to define the methodology that would be used to interpret
GRASSP scores.
4.4 Methods
Ethical approval was attained at all institutions participating in the study. Seven
centres participated in the trial (three European and four North American). Data was
collected on a cross section of individuals with chronic traumatic tetraplegia, who were
medically and neurologically stable, and able to provide informed consent (n=72).
Individuals with moderate brain injury and/or any other pathology affecting the upper limb
were excluded. Sample size calculation was based on an estimated intra class correlation
46
coefficient of 0.80 or greater with an alpha of 0.05. The minimum estimated sample size was
39 for 3 repeated measures according to Donner & Eliasiw’s estimation curves (Donner and
Eliasiw, 1987). Due to the multi-national collaboration 72 datasets were collected. Two
workshops (one in Europe and one in North America) were conducted to train the examiners
regarding the study protocol and appropriate use of all study measures. Instructions and
demonstrations in the administration of the primary measure, GRASSP and secondary
measures, SCIM, ISCSCI and the CUE were provided to all examiners. All data were de-
identified and sent to the central site from all participating centres where it was aggregated.
European and North American data was pooled to render 72 data sets.
Analytic Plan: A priori the following was hypothesized: 1) a cumulative predictive
pattern of response items would not exist thereby causing the developers to establish an
alternative method for summating subtest scores to present results; 2) subtest total scores
would define groups of the sample according to upper limb impairment, not necessarily
according to level of lesion as designated by ISCSCI; and 3) a method to present and
interpret the results of the GRASSP scores would be derived.
Guttman scaling methodology (see Appendix 5 for details of method used) was
applied to the five subtests and the items within (dorsal sensation, palmar sensation, strength,
prehension ability and prehension performance) separately. With the number of items in each
subtest being small and the sample being relatively small for this type of analysis, the rank-
ordering was done manually. Guttman scaling was performed by: 1) re-ordering the items in
each subtest to reflect spinal cord levels (anatomy, rostral to caudal) 2) matrices were
constructed for each subtest and the items within which displayed all responses for the total
sample and each matrix was sorted so that individuals who responded with more impairment
were moved to the top and those with less impairment were moved to the bottom (Trochim,
47
2010). See Table 8 for an example of the process and Appendix 5 for all matrices. The
process was repeated until all individuals in the sample fit a pattern of cumulative
predictability; individuals that did not show cumulative predictability were considered the
“exceptions” and were separated (matrices for all five subtests are available in Appendix 5).
Cumulative meaning items that were rostral anatomically were intact prior to more caudal
items being intact. When the process was complete the matrices showed a cumulative pattern
when read from left to right across the columns (items) for each individual. The frequency of
cumulative pattern predictability determined whether the subtest scale could be used as a
cumulative score and interpreted to have item prediction.
48
Table 8: Example of Guttman Scaling Using Scores of Palmar Sensation Items Table 8 illustrates the Palmar Sensation (PS) item and total scores of the first 60 of 72 individuals in the sample. The table on the left side presents the sample as the data was entered and the table on the right side presents the sample ordered by cumulative predictability of the item scores, with descending order of less impairment. This process was repeated for all five subtests.
Subtest total scores were placed in ascending order to establish the ranges of
GRASSP scores and associated with ISCSCI sensory and motor levels, zone of partial
49
preservation scores, AIS classification and SCIM-SS. Table 11 defines the item scores and
ranges, the subtest total scores and ranges for each subtest and the meaning of ranges within
the scores. The numeric value derived from the subtest total score was intended to represent
the neurological integrity of the upper limb which would represent the spinal cord levels
secondary to the dematomal and myotomal patterns followed. For example, if the total
possible score for dorsal sensation is 12, the meaning of the numeric value would be as
follows: 0-4 would represent any degree of innervation between C5 and C6 with 4
representing normal sensation of C6, 5-8 would represent normal C6 innervation and any
degree of innervation between C6 and C7 with 8 representing normal C7 sensation and so on
and so forth. Therefore, the range of numeric values represents a neurological level of
sensation with the lower end of the range representing more deficit and the higher end of the
range representing less deficit.
4.5 Results
Sample - The data used in this analysis included a multi centre/multi national cross
section with the total sample consisting of 72 individuals with chronic tetraplegia ranging
from 6 months to 20 years post injury. Distribution of the sample according to the ISCSCI is
defined in Table 9. Approximately 52.5% of the individuals fell into the C6 - C7 motor level
group and approximately 66% fell into the C4 - C6 sensory level group. The AIS indicates
that approximately 61.1% of the sample was incomplete.
50
Table 9: Demographics of the Sample Based on ISCSCI sensory and motor levels and AIS Classification, n=72 (presented as n values and percentages of sample)
ISCSCI Levels Motor Level
n (%)
Sensory Level
n (%)
AIS Classification
n (%) Right Left Right Left A B C D
C2-C4 10 (14) 14 (19) 29 (40) 29 (40) C5 10 (14) 9 (13) 11 (15) 9 (13) C6 23 (32) 21 (29) 17 (24) 19 (26) C7 17 (24) 16 (22) 8 (11) 6 (8) C8 4 (5) 5 (7) 1 (2) 2 (3)
T1 and below 8 (11) 7 (10) 6 (8) 7 (10)
28 (39)
18 (25)
12 (17)
14 (19)
Legend for Table 9: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, AIS – ASIA Impairment Scale
The Guttman Scaling process defined the frequency of cumulative predictability for
each subtest and Table 10 defines the predictable and non-predictable portions of the sample.
The characteristics of the individuals as designated by the ISCSCI sensory and motor levels,
AIS and zone of partial preservation scores are also defined for the predictable and non
predictable portions of the sample. The portion of the sample that showed cumulative pattern
predictability in the subtest items were not defined by any specific characteristics according
to the ISCSCI and AIS designations, although the portion of the sample that showed non-
predictive patterning did reveal some characteristics as defined by the ISCSCI and AIS.
Individuals with non-predictive patterning of palmar sensation were incomplete (B, C or D).
Individuals with non-predictive patterning of strength tended to have normal strength in the
upper cervical region. The remaining subtest non-predictable patterning portions of the
sample did not present with any specific and distinct characteristics according ISCSCI and
AIS designations.
51
Table 10: Portion of Sample which presented with and without a Cumulative Predictive Pattern for each Subtest Total Score in the GRASSP (presented by n value and percentage of sample) Subtest Predictive
N (%)
ISCSCI Characteristics of
Predictive
Non-Predictive
N (%)
ISCSCI Characteristics Non-Predictive
AIS S/M Level
ZPP AIS S/M Level
ZPP
Dorsal Sensation
63 (87) A-D C4-T1 C5-C8 9 (13) A,B,C C4-T1 C8
Palmar Sensation
65 (90) A-D C4-T1 C5-C8 7 (10) B,C,D C5,C6 C8
Strength 56 (78) A-D C4-T1 C5-T1 16 (22) A-D C4-C7 C6-T1
Prehension Ability
59 (82) A-D C4-T1 C5-T1 13 (18) A-D C6,C7, T1
C6, C7, T1
Prehension Performance
60 (83) A-D C4-T1 C5-T1 12 (17) A,B,C C6,C7, C8
C6, C7, C8
Legend for Table 10: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, AIS – ASIA Impairment Scale, Predictive – Group of sample with cumulative predictive pattern; Non-predictive - Group of sample without cumulative predictive pattern (exceptions); ZPP – Zone of Partial Preservation Score; S/M Level – Sensory or Motor Level as designated by ISCSCI Shaded cell is referred to in the discussion
During development of the GRASSP, items that were included represented
dermatomal and myotomal patterns (Moore, 1985). Therefore, items a priori had or
represented a degree of neurological integrity. The values in Table 11 define the subtest
scores, ranges and the intended meaning of scores based on their representation of the spinal
cord. Each subtest had ranges within it that represented neurological integrity of the upper
limb as defined by the state of deficit in the peripheral limb. Because the location of anatomy
assessed is associated to spinal cord levels the peripheral representation also defines spinal
cord level. However, the range between levels was also accounted for by the GRASSP. In the
case of the strength subtest total score (50) a total score between 0-9 represents innervation
52
rostral to C5 and 10 represents C5 innervation, 11-15 represents innervation between C5 and
C6 and 15 represents C6 innervation and so on and so forth. The intended meaning of scores
which are defined for all the subtests and a glossary for what the numeric values should
represent is provided. These ranges of scoring are based on the items and associated
dermatomes and myotomes. It was not determined how often the scores represented their
intended meaning, with the limited sample. However, it was certain at this stage of score
development that an increase in subtest total score value represented: 1) an increase in
neurological integrity or decrease in deficit, and 2) the increase was moving in the direction
from rostral to caudal approximately 80% of the time with this specific sample.
ISCSCI motor and sensory level designations, AIS and zone of partial preservation scores
were also used to benchmark the ranges of GRASSP scores. The ISCSCI and AIS did not
discriminate the levels of GRASSP scores similarly. The difference in groupings was
predominantly due to the method in designation of levels and scores. The SCIM-SS was also
used to benchmark the levels of GRASSP scores as it represented upper limb function. The
SCIM-SS did provide more definitive groupings in relation to the GRASSP subtest scores.
53
Table 11: ISCSCI Sensory and Motor Levles and SCIM-SS Benchmarks for GRASSP Discriminative Score Ranges Subtest
# of items, item score range subtest score range
Intended Meaning of Score (levels of GRASSP scores)
SCIM-SS Score
ISCSCI and AIS Classification S or M Level ZPP
Dorsal Sensation 3, 0-4 0-12
0-4 – C6 5-8 – C7 9-12 – C8
0-10 5-15 10-20
A-D A,B,D A-D
C4-C7 C4-C7 C4-T1
C4-C8 C7-C8
C8 Palmar Sensation
3, 0-4 0-12
0-4 – C6 5-8 – C7 9-12 – C8
0-10 5-15 10-20
A,B,D A-D A-D
C4-C6 C4-C7 C4-C8
C5-C8 C7-C8
C8 Strength 10, 0-5 0-50
0-10 – C5 11-15 – C6 16-25 – C7 26-40 – C8 41-50 – T1
0-5 5-10 5-15 10-20 10-20
A A,B
A,B,C B,C,D B,C,D
C4-C5 C4-C6 C6-C7 C4-C8 C6-T1
C5-C7 C5-C7 C6-C8 C7-T1
T1 Prehension Ability
3, 0-4 0-12
0-6 – C5-C6 7-12 – C7-T1
0-10 11-20
A-D A-D
C4-C7 C7-T1
C5-C8 T1
Prehension Performance 6, 0-5 0-30
0-5 – C5-C7 6-10 – C5-C7 1-15 – C5-C7 16-20 – C5-T1 21-25 – C5-T1 26-30 – C5-T1
0-10 0-10 0-10 11-20 11-20 11-20
A A,B
A,B,C A,B,C B,C,D A-D
C4-C5 C5-C7 C5-C7 C6-C8 C6-C8 C5-T1
C5-C7 C6-C7 C6-C8 C7-T1 C7-T1
T1 Legend for Table 11: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, AIS – ASIA Impairment Scale, SCIM-SS – Spinal Cord Independence Measure Selfcare Subscale, ZPP – Zone of Partial Preservation Score; S or M Level – Sensory or Motor Level as designated by ISCSCI, AIS – Classification (A,B,C,D), Shaded cells are referred to in the discussion.
54
4.6 Discussion
The GRASSP is an impairment measure for the upper limb after tetraplegia which
includes three domains (sensation, strength and prehension) and five subtests. Each subtest
includes anywhere from three to ten items with various item ranges and all item response
ranges are polychotomous. The GRASSP was designed so that numeric values would
represent impairment; a numeric value of zero or close to zero represents more deficit and a
numeric value close to or at the maximum value represents less or no deficit.
Because impairment is being represented by numeric values it was essential to define
the meaning of the numeric values that are rendered by the GRASSP. Items for each subtest
were generated to ensure there was representation of all spinal cord levels with at least two
items. Although it was not expected that such a strong representation of an anatomical
pattern would be seen in the GRASSP results, it was present. The initial part of score
development for the GRASSP entailed the process of determining the scale characteristics.
Frequency of cumulative pattern predictability for each subtest was used to determine if the
subtests presented with Guttman characteristics. All five subtests within the GRASSP
presented with Guttman characteristics when the items were ordered according to spinal cord
anatomy. The Guttman pattern mimicked the anatomical spinal cord levels moving from
rostral to caudal for the sensation and strength domains. Therefore, incorporating the
anatomical concepts in the development of the GRASSP has provided a basis to understand
the cumulative numeric values; however, it is not the only factor in interpretation of the
scores because the impairment and numeric values are derived over the upper limb. The
numeric values have some merit in representing the level of spinal cord integrity;
nonetheless, they also represent the neurological integrity over the upper limb which may not
be a one to one correlation to the level of injury. Thus numeric values should be considered
55
as general representations of the level of innervation at the spinal cord as per Table 11 and
the value should be interpreted as a portion of normal sensory or motor function. For
example someone with six out of twelve for palmar sensation would have about half of the
normal sensation in the hand and it would be C6 to partial C7 innervation.
Ideally, if a larger sample were available, Rasch analysis or the Mokken Scale
Analysis would be more appropriate methods to determine the predictability of item
response. Therefore, the sample size is one limitation of this particular analysis. However,
this work does provide the preliminary evidence to proceed with a Rasch or Mokken Scale
Analysis when the available sample is large enough.
When interpreting subtest total scores for dorsal sensation, palmar sensation, strength,
prehension ability and prehension performance a score closer to zero represented more
deficit, while a score closer to the maximum represented less deficit. These numerical values
when interpreted according to Table 11, enable the assessor to understand the general degree
of impairment an individual would possess.
In addition, the ISCSCI characteristics of the individuals who did not demonstrate
cumulative predictable patterning were still presented so that there is an understanding that
subtest totals were only predictable approximately 80% of the time. Subtest totals
represented the individual with enough predictability that the numeric value could be
interpreted to represent level cumulativeness (rostral to caudal). The remaining 20% of the
time the pattern was unpredictable. Since the ISCSCI designated sensory and motor levels
according to the most normal caudal level of innervation, ISCSCI levels did not definitively
define the predictable and non predictable groups. However, classification at times defined
the unpredictable individuals. For individuals who were incomplete (B, C, D), palmar
56
sensation was manifested in a less predictable pattern; similarly for individuals that were A,
B or C, strength and prehension were also manifested in a less predictable pattern.
Since cumulating the subtest items was qualified by the scaling process, the total
subtest scores were also benchmarked for the first time. Table 11 presents the range of each
item and subtest total. Ranges of the subtest totals were intended to represent specific sensory
and motor levels (both partial and full) numerically. The intended meaning of the numeric
values, otherwise known as the levels of GRASSP scores are presented in Table 11. Ranges
of GRASSP scores do show some congruency with SCIM-SS groupings which represent
upper limb function. Function is more discriminative at the upper and lower cervical ranges
of tetraplegia whereas the scores and levels that represent the mid-cervical region tend to be
less discriminative. This is seen in both the sensation and strength data. The group of
individuals with sensation that generally represent C6 innervation or rostral to that (0-4)
generally score between 0-10 on the SCIM-SS. The group of individuals with sensation that
generally represents C7 to C8 innervation (9-12) generally score between 11 and 20 on the
SCIM-SS The group of individuals with sensation that generally represents C6 to C7
innervation score on the SCIM-SS somewhere in between those two distinct groups (5-15).
In the strength scores we see something similar where the upper and lower cervical levels of
innervation (C5, C8, T1) are more discriminative with SCIM-SS scores than the mid-cervical
levels of innervation (C6, C7).
Due to the fact that subtest items can be added for subtest total scores and cannot be
added up across subtests the most meaningful way to observe GRASSP subtest totals is to
plot the scores on a cluster bar graph for each hand. Scores are normalized before they are
plotted on the graph. Normalizing the subtest scores before combining was a method
employed by Rosen and Lundborg in the development of a multi-domain measure for
57
peripheral hand injury (Rosen and Lundborg, 2003). Figure 4 illustrates three individuals
from the sample (right hands only) and the associated subtest scores. This method allows one
to view multiple individuals or repeated measures of the same individual collectively in one
diagram. Scores defined in this method enable the assessor to make comparisons over time.
Figure 4: A Visual Representation of GRASSP Subtest Scores
Figure 4 illustrates three representative cases from the sample. The legend defines the ISCSCI sensory/motor/AIS scores for the right side only. In summary, items of each subtest can be summated; approximately 80% of the time
the scores follow a cumulative predictive pattern which means the subtest total is cumulative
across spinal cord levels rostral to caudal. Individual ranges of subtest total scores represent
58
innervation according to spinal cord level and a percent of the total score established during
development of the measure.
The discriminative GRASSP score ranges in some cases better reflect neurological
status; specific subtests within the GRASSP reflect impairment better than other subtests. For
example as the ISCSCI sensory and motor level is more caudal the scores of the GRASSP
strength subtest tend to increase, see Table 11. Table 10 shows that individuals who present
with a non predictable pattern for palmar sensation tend to be incomplete (B, C or D). SCIM-
SS divisions tend to represent levels of GRASSP scores more discriminatively for the dorsal
and palmar sensation and strength subtest scores, see Table 11. It will be important with a
larger sample to see if these trends are confirmed as definitive known groups in the future.
59
CHAPTER V: RESULTS PART 2
5.1 THE GRADED REDEFINED ASSESSMENT OF STRENGTH SENSIBILITY AND
PREHENSION (GRASSP) – RELIABILITY AND VALIDITY
5.2 Abstract
With the advent of new interventions targeted at both acute and chronic SCI, it is critical
that techniques and protocols are developed that reliably evaluate changes in upper limb
impairment/function. The GRASSP protocol, which includes five subtests, is a quantitative
clinical hand impairment measure designed for use in acute and chronic SCI. Objectives of the
study were to: 1) establish the inter rater and test retest reliability and 2) establish the construct
and concurrent validity with the ISCSCI, SCIM and the CUE. The study protocol included
repeated administration of the GRASSP on a cross section of individuals with tetraplegia who
were neurologically stable (n=72). Two examiners assessed the individuals over a seven day
period; a complete ISCSCI, CUE and SCIM were also administered. Reliability was tested with
intra class correlation coefficients; construct validity was established by comparing the value of
the additional components in the GRASSP to the ISCSCI, and concurrent validity was tested
with Pearson correlation coefficients. Inter rater and test retest reliability for all subtests within
the GRASSP were above the hypothesized value of 0.80 (0.83-0.96 and 0.86-0.99, respectively).
Construct validity was confirmed by discordance of the sensibility and strength subtests within
the GRASSP; the subtests showed concurrence with the SCIM, SCIM self-care subscale and
CUE. Impairment revealed the strongest concurrence with self perception of function (0.57-0.83,
p < 0.0001). The GRASSP demonstrates reliability, and construct and concurrent validity for use
as a standardized upper limb impairment measure for individuals with tetraplegia.
60
5.3 Introduction
Individuals with tetraplegia, researchers and society in general acknowledge that residual
hand and upper extremity function are critical determinants of independence for people with
tetraplegia (Anderson, 2004; DeVivo, 1997). As a result, any gain in arm and hand function is
important to enhance quality of life (Anderson, 2004; Snoek et al., 2005) for this population. For
this reason, there are many concentrated efforts towards the development of methods to restore
upper limb function after spinal cord injury (SCI) (Popovic et al., 2006; Taylor et al., 2002;
Prochazka et al., 1997). Scientists conducting research in SCI report that there is a lack of
appropriate outcome measures for the assessment of impairment and function for the upper limb
after cervical SCI (Dunn et al., 2008; Miller et al., 2008; van Tuijl et al., 2002). Researchers are
faced with the challenge of using inadequate outcomes to prove efficacy of novel interventions
and as a result, measures that are non-specific to tetraplegia are common limitations in efficacy
studies using new therapies (Prochazka et al., 1997; Fujiwara et al., 1999; Kohlmeyer et al.,
1996).
The upper limb acts as a kinetic chain with the proximal joints, placing the more distal
ones in space for prehension and manipulation of the environment (Kapandji, 1970). The
components of upper limb function are defined as reach, prehension and manipulation
(Shumway-Cook, 2007; Brand and Hollister, 1999). However, no single method of assessment
has been developed to capture these impairments in a manner sensitive to small but important
changes that might occur as treatment in SCI advances. Acknowledgement of the integration of
levels of impairment into functional tasks is important for individuals with tetraplegia as many
adaptive techniques make use of available motor and sensory power to execute a task.
61
Concurrently, there is no agreed upon outcome that captures the sensory, motor and functional
findings in a manner that will be sensitive to neurological change in these patients.
The need for a sensitive battery is more important because of two main factors. First,
there is now greater potential for neurological recovery as there are an increasing proportion of
incomplete injuries; almost 50-60% of SCI are incomplete (Marino et al., 1999; Sekhon and
Fehlings, 2001). Second, novel interventions are also targeting a means to enhance neurological
recovery and require outcomes that can substantiate these gains in order to prove efficacy
(Steeves et al., 2007).
Although a number of researchers throughout the world have been working on the
development of upper limb measures specific to the SCI population, the measures developed are
predominantly functional tests of capacity. The Link Hand Function Test (Link, 2004) and the
Van Lieshout Test (Post et al., 2006) are quantitative functional tests for the upper limb. The
Grasp and Release Test (Smith et al., 1996) and the Toronto Rehab Hand Function Test (TR-
HFT) are also tests of capacity, but are specifically designed to evaluate outcomes after
functional electrical stimulation (Popovic et al., 2006). The TR-HFT is the only test which
evaluates stereognosis through the tasks. All of these tests are useful during the post acute and
rehabilitation phases, however, they are not feasible for use in the acute phase when subtle
changes in impairment are critical to establish the relationships of multiple interventions. Due to
the fact that impairment represents the subsystems which comprise the components of upper
limb function (reach, prehension and manipulation), assessing the impairment is an important
factor in defining what subcomponent of upper limb function is affected by an intervention
and/or natural recovery.
62
Measuring hand impairment and function post injury is complex and dependent on
multiple factors such as the interaction between the sensory and motor aspects of movement. The
ISCSCI (Marino, 2000) is the current standard used to assess recovery after SCI by designating
the most normal caudal level of sensory and motor functioning. The ISCSCI classifies by
normal spinal cord level and in doing so, may be relatively insensitive to measure upper limb
impairment in a manner that can quantify change. An outcome measure that is sensitive and
responsive to change is warranted; one that can be used to determine the extensive individual
variability in sensory and motor change post injury. This measure should be able to track the
natural history of recovery and the response of individuals receiving treatment.
The GRASSP was developed by a team of six researchers to fill this void and in 2006 the
International GRASSP Research and Design Team was brought together to establish an upper
limb impairment measure. The GRASSP is a clinical measure of upper limb impairment which
incorporates three domains vital to upper limb function: sensibility (sensation), strength (motor)
and prehension (grasp and pinch). During the theoretical development of the GRASSP an
anatomical, neurophyisological and functional approach were incorporated into all domains
included. Sensory elements which represented physiology and function were incorporated
including palmar test locations and a sensory modality with well established measurement
properties (Mackin et al., 2003). Motor elements with a stronger representation of neurological
levels (Moore, 1985) and an increased number of muscles tested per myotome were included to
enhance accuracy and precision of the measure. In order to ensure that the presence (absence) of
movement of the hand during the early stages post injury was not missed, a prehension ability
test of three grasps was incorporated. A prehension performance test was incorporated to make
certain that movement was assessed within a functional paradigm and to determine how
63
movements were performed and contextualized. The GRASSP includes Semmes Weinstein
Monofilament (SWM) testing on the dorsal and palmar surface of the hand, motor testing on 10
muscles of the upper limb (three arm, seven hand), performance of three prehensile positions and
performance of six task oriented prehension skills. All tests are graded using a scoring system
that focuses on how movement is performed. Each subtest renders a total score which can be
used for comparison with corresponding subtest scores and the use of a cluster bar graph has
been employed for reporting results of repeated measurements. Please see Figure 5 for a
summary of the components of the GRASSP and an example of the current scoring system.
The GRASSP has been developed in stages. The overall objective was to develop a
comprehensive outcome measure capable of the following: able to assess upper limb
impairment/function that is responsive to change, able to assess the extent of natural recovery
over the acute to chronic phases, capable of being used in both clinical research, and the
evaluation of novel interventions. Development has also included evaluation of the psychometric
properties of reliability, validity and responsiveness. Theoretical and clinimetric development of
the GRASSP including the rationale and analysis used to determine inclusion of subtests has
been previously reported (Kalsi-Ryan et al., 2009). The three objectives addressed in this
manuscript are: 1) to establish the inter rater and test retest reliability of the subtests within the
GRASSP; 2) to establish the construct validity (agreement and discordance of GRASSP) against
the ISCSCI and 3) to establish the concurrent validity of the GRASSP with the SCIM (Catz et
al., 2004) and CUE (Marino et al., 1998).
64
Figure 5: Summary of GRASSP Components
Figure 5 presents the domains, items, scores and global scoring.
Although, the ISCSCI is not specific to the upper limb, it was selected because of its use
as a primary outcome measure in many SCI studies. The SCIM was selected as a measure of
function and independence for comparison to the impairment elements in the GRASSP. The
65
CUE was selected as a self-perceived measure of function due to the fact that consumer input in
terms of an individual’s own evaluation of their functional status is becoming increasingly
important in the field (Anderson, 2004). Although the measures selected for validity comparison
do not share the same constructs as the GRASSP, these measures were the only available
measures with well established psychometric properties.
5.4 Methods
Ethical approval was attained at all institutions participating in the study. A cross-
sectional multi centre trial was conducted to establish the reliability and validity of the GRASSP.
Inclusion/Exclusion: Individuals with chronic traumatic tetraplegia who were neurologically and
medically stable, between the ages of 16 and 65 and able to provide informed consent were
included in the study. Individuals with moderate brain injury who were neurologically unstable
or individuals with any other pathology causing upper limb impairment were excluded from the
study. A sample size of 40 people was calculated based on an estimated ICC of 0.80 or greater
with an alpha of 0.05. The minimum estimated sample size was 39 for 3 repeated measures
according to Donner & Eliasiw’s estimation curves (Donner and Eliasiw, 1987). As a result of
the multi centre collaboration a greater number of datasets were available, allowing the n value
to increase to 72.
Seven centres participated in the trial (3 European and 4 North American). Two
workshops (one in Europe and one in North America) were conducted to train the examiners
regarding the study protocol and appropriate use of all study measures. Instructions and
demonstrations on the administration of the primary measure, GRASSP and secondary measures,
SCIM, ISCSCI and CUE were provided to all examiners. Training was provided to reduce the
66
observer variability (Wright and Feinstein, 1992). The protocol conducted in Europe consisted of
two examiners assessing each study participant once; during the first visit examiner one
administered the GRASSP and during the second visit examiner two administered the GRASSP
and all secondary measures. European data was used in the analysis to establish inter rater
reliability and validity.
The protocol conducted in North America consisted of two examiners assessing each
study participant a total of three times. During the first visit examiner one and two administered
the GRASSP approximately one hour apart and during the second visit examiner one
administered the GRASSP and all of the secondary measures. North American data was used in
the analysis of inter rater, test retest reliability and validity. The European and North American
data was pooled to render 72 data sets for the analysis of inter rater reliability, construct and
concurrent validity. Forty five data sets were used for the analysis of test retest reliability. All
data were de-identified, entered into excel tables and sent to the central site. Data was aggregated
and analyzed using SAS 9.1.
Analysis – A priori we anticipated the following: 1) Inter rater and test retest reliability
for subtest scores would be greater than or equal to an ICC value of r=0.80. According to
Streiner and Norman, reliability is considered to be good if the ICC is above 0.75 (Portney and
Watkin, 2000; Steiner and Norman, 1995). 2) Construct validity (Patrick and Erickson, 1993;
Steiner and Norman, 1995) would be demonstrated by GRASSP sensibility and strength subtests
defining sensory and motor impairment with only a slight to moderate agreement when
compared to the ISCSCI.
Reliability was analyzed by conducting intra class correlation coefficients for non-
parametric variables with a two-way random effect on the GRASSP subtest total scores (Portney
67
and Watkin, 2000). The GRASSP was designed to have a broader range of items and response
levels related to the upper limb in comparison to the ISCSCI, therefore, it would be expected that
the GRASSP sensibility and strength subtests and items would define the sample with greater
accuracy. Construct validity was analyzed by comparing the descriptive frequency data of the
ISCSCI sensory and motor test items with the corresponding GRASSP subtest items. Instead of
comparing total scores or levels derived, item comparisons were conducted to ensure summed
data did not influence results. Kappa coefficients were also conducted in the sensory
comparisons to determine the degree of agreement and discordance between both measures.
Agreement is established by the following standards as established by Landis and Koch: [ .00 -
poor, .01-.20 – slight, .21-.40 – fair, .41-.60 – moderate, .61-.80 – substantial and .81-1.00 –
almost perfect (Landis and Koch, 1977)]. Concurrent validity was analyzed by conducting
Pearson correlation coefficients between GRASSP subtest scores with SCIM total scores, SCIM
SS, and CUE scores. The a priori hypothesis was that concurrent validity with functional tests
would be demonstrated by a moderate association of GRASSP subtest scores with the SCIM,
CUE, and SCIM-SS.
5.5 Results
Sample - The data used in this analysis included a multicentre/multinational cross section of data;
the total sample consisted of 72 individuals with chronic tetraplegia ranging from 6 months to 20
years post injury. Distribution of the sample according to the ISCSCI is defined in Figures 6.1
and 6.2, with Figure 6.1 showing subgroups according to ISCSCI sensory levels and Figure 6.2
showing the distribution of the sample according to the ISCSCI motor levels (right and left data
are presented side by side). Approximately 52.5% of the individuals fall into the C6 - C7 motor
68
levels while approximately 66% fall into the C4 - C6 sensory levels. The ISCSCI Impairment
Scale (AIS) which defines completeness of the sample identified 39% (n=28) of the sample as
AIS A complete and 61% (B-25% n=18; C-19% n=14; D-17% n=12) of the sample as AIS B, C
or D incomplete.
Figure 6: Histograms of Sample Distribution. 6.1 6.2
Figure 6.1 defines the sensory level distribution of the sample (n=72) as designated by ISCSCI and histogram 6.2 defines the motor level distribution of the sample as designated by ISCSCI. Both left and right are presented. Note: motor levels between C1 and C4 are based on sensory function.
69
Reliability – Inter rater and test retest reliability were established for each subtest total,
right and left sides separately. Table 12 lists the non-parametric ICC values for the subtests
within the GRASSP, including the confidence interval for each value. ICC for inter rater
reliability ranged between 0.84 to 0.96 and test retest reliability ranged between 0.86 to 0.99 with
a significance level of p < 0.0001 for all values.
Table 12: Reliability Values of Subtest Scores within the GRASSP Subtest Inter rater Reliability Test Retest Reliability
ICC CI ICC CI SWM Right 0.84 0.75-0.89 0.95 0.91-0.97 SWM Left 0.91 0.86-0.94 0.86 0.76-0.92 Strength Right 0.95 0.93-0.97 0.98 0.98-0.99 Strength Left 0.95 0.92-0.97 0.98 0.96-0.98 Prehension Ability Right 0.95 0.92-0.97 0.98 0.96-0.99 Prehension Ability Left 0.95 0.92-0.97 0.98 0.97-0.99 Prehension Performance Right 0.95 0.92-0.97 0.93 0.88-0.96 Prehension Performance Left 0.96 0.93-0.97 0.96 0.93-0.98 Legend for Table 12: All values with significance level of p<0.0001
Construct validity is defined by the broader range of findings of impairment determined
by the sensibility and strength subtests within the GRASSP. Precision of the GRASSP was
established by comparing the sensibility and strength subtest items to the sensory and motor
upper limb items in the ISCSCI. It is known that the ISCSCI derives neurological levels based on
the complete intactness of normal innervations and in doing so, sensory and motor intactness
below the most normal level is not incorporated in defining the sensory and motor level.
Essentially, the impairment is not fully defined by the designation of a level in the ISCSCI. With
GRASSP, levels are not derived from the numeric values and as a result, in order to compare the
tests, items within each subtest were compared. It should be noted that individuals with a sensory
level of C5 or any level rostral to C5 would score 0 on the sensory testing in the GRASSP. The
70
individual sensory (SWM) test locations and the individual muscles tested in the GRASSP were
compared to the sensory test locations for light touch and muscles for the upper limb in the
ISCSCI respectively. Referring back to the introduction for sensibility, the following two
additional components were added to the GRASSP: a greater number of response levels based on
the modality used and a greater number of test locations specific to the hand. ISCSCI sensory
and motor levels were used only to group the sample for presentation of the data; the items are
presented both in subgroups and as the whole sample. All comparisons were made for right and
left separately. Each study participants sensory information was compared (both dorsal and
palmar components from GRASSP were combined) and three conditions of comparison existed:
1) There was agreement between GRASSP sensibility and ISCSCI sensory items (absent,
impaired or normal) 2) There was discordance between GRASSP sensibility and ISCSCI sensory
items due to the added palmar test locations; and 3) There was discordance between GRASSP
sensibility and ISCSCI sensory items due to the increased response levels used in the GRASSP.
Table 13 defines the proportions of the subgroups and the whole sample that fall into these three
different conditions. On average 54% of the sample showed discordance in sensory innervation
when assessed with GRASSP due to the additional components of sensory testing included.
Table 14 shows the level of agreement between ISCSCI-Light Touch (ISCSCI-LT) and
GRASSP-SWM for the C6, C7, and C8 dorsal test locations. The kappa coefficients of C6, C7,
and C8 reveal that the level of agreement is not substantial; the statistical analysis indicates the
two tests demonstrate different results. Essentially ISCSCI-LT and SWM provide different
results regarding the sensory status of individuals with tetraplegia.
71
Table 13: Construct Validity Agreement/Discordance of Sensory Results between GRASSP and ISCSCI (n=72)
AGREEMENT DISCORDANCE ISCSCI
Sensory Levela
N
N (%) 1
N (%) 2
N (%) Right Total Sample 72 32 (44) 16 (22) 24(33)
Left Total Sample 72 34 (47) 13 (18) 25 (35) Right C1 – C4 29 14 (19) 7 (10) 8 (11) Left C1 - C4 29 12 (17) 6 (8) 11 (15)
Right C5 11 5 (7) 4 (6) 2 (3) Left C5 9 5 (7) 3 (4) 1 (1)
Right C6 17 6 (8) 5 (7) 6 (8) Left C6 19 8 (11) 4 (6) 7 (10)
Right C7 8 4 (6) 0 (0) 4 (6) Left C7 6 3 (4) 0 (0) 3 (4)
Right C8 and below 7 3 (4) 0 (0) 4 (6) Left C8 and below 9 6 (8) 0 (0) 3 (4)
Legend of Table13: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, Agreement – GRASSP and ISCSCI are consistent with assessment of sensation, Discordance 1 – due to added palmar test locations in GRASSP; Discordance 2 – due to the increased response levels (SWM) used in the GRASSP; aISCSCI levels are used only to subgroup the whole sample
Table 14: Level of Agreement between GRASSP-SWM and ISCSCI-LT for C6, C7 and C8 ISCSCI-LT GRASSP-
SWM C6 C7 C8
Right C6 Left C6
0.412 0.442
Right C7 Left C7
0.474 0.459
Right C8 Left C8
0.511 0.503
Legend of Table 14 presents the Kappa coefficients for the comparisons made between C6, C7 and C8 test locations on the dorsal surface of the hand. The comparisons are made between the LT and SWM measurements made over the same test locations. All values have a significance level of p < 0.001.
Each study participants strength information was compared and three conditions of
comparison existed: 1) There was agreement between GRASSP strength items and ISCSCI upper
72
limb motor items; 2) There was discordance between GRASSP strength items and ISCSCI upper
limb motor items due to the added muscles in GRASSP; and 3) There was discordance between
GRASSP strength items and ISCSCI upper limb motor items due to the derivation of a level
from the motor testing in the ISCSCI. Table 15 defines the proportions of the subgroups and the
whole sample that fall into these three different conditions. On average 53% of the sample
showed a different degree of motor innervation when assessed with GRASSP due to the
additional muscles of the motor testing and the use of a summative numeric score to define
strength.
Table 15: Construct Validity Agreement/Discordance of Strength Results between GRASSP and ISCSCI (n=72)
AGREEMENT DISCORDANCE ISCSCI
Motor Levela
N
N (%) 1
N (%) 2
N (%) Right Total Sample 72 36 (50) 19 (26) 17(24)
Left Total Sample 72 34 (47) 20 (28) 17 (24) Right C1 - C4 10 1 (1) 6 (8) 3 (4) Left C1 – C4 14 6 (8) 6 (8) 1 (1)
Right C5 10 3 (4) 2 (3) 5 (7) Left C5 9 3 (4) 1 (1) 5 (7)
Right C6 23 13 (18) 3 (4) 7 (10) Left C6 21 11 (15) 5 (7) 4 (6)
Right C7 17 9 (12) 6 (8) 2 (3) Left C7 16 6 (8) 7 (10) 3 (4)
Right C8 4 2 (3) 2 (3) 0 (0) Left C8 5 1 (1) 1 (1) 3 (4)
Right T1 and below 8 8 (11) 0 (0) 0 (0) Left T1 and below 7 7(10) 0 (0) 0 (0)
Legend for Table 15: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, Agreement – GRASSP and ISCSCI are consistent with assessment of strength, Discordance 1 – due to added muscles in GRASSP; Discordance 2 – due to deriving a level from numeric score; aISCSCI levels are used only to subgroup the whole sample
Concurrent Validity: Table 16 displays all of the concurrent validity values. Right and
left data were combined for this analysis and Pearson correlation coefficients were conducted to
73
establish the association between GRASSP subtests and the CUE, SCIM and SCIM-SS. It was
noted that the total SCIM score shows the least association with subtests of the GRASSP,
however, the association is positive and is significant with a p-value of less than 0.0001. The
SCIM-SS shows a stronger association with the subtests of the GRASSP. This is similar to the
result reported by Rudhe and van Hedel (2009) showing the specificity of the self-care subscale
in SCIM to the upper limb motor testing done in the GRASSP. The CUE shows the strongest
associations with the GRASSP subtests which represent a strong association between self-
perceived function and impairment. As the secondary measures become more specific to the
upper limb, it is noted that the association becomes stronger for all domains. Subtests within the
GRASSP demonstrate moderate to substantial concurrent validity with the SCIM and the CUE.
This is indicative of a positive relationship between impairment, function and independence. It is
pertinent to note that the strongest associations linking impairment (GRASSP subtests) are those
between self perceived function (CUE), confirming that individuals perception of their own
ability is comparable to quantitative testing in the chronic SCI population.
Table 16: Concurrent Validity of GRASSP Subtests and Functional Measures Subtest Score SCIM SCIM-SS CUE
SWM Total (R + L) 0.57 0.74 0.77 Strength Total (R + L) 0.59 0.74 0.76 Prehension Performance Total (R + L) 0.68 0.79 0.83 Legend for Table 16: All values with significance level of p < 0.0001, Pearson Correlation Coefficient; Moderate Concurrence=0.61- 0.79, Substantial Concurrence=0.80-1.00; SWM - Semmes Weinstein Monofilaments, R - Right, L - Left
5.6 Discussion
In order for the GRASSP to be accepted as an adequate measure to define upper limb
sensorimotor deficit associated to cervical SCI, psychometric testing with the tetraplegic
population was a requirement. As a result of this study, the psychometric properties of reliability
74
and validity have been established. The GRASSP was designed to be a sensitive clinical
impairment measure specific to the upper limb with a sound theoretical framework and relevant
domains.
Reliability is considered good for group level analysis when the ICC is greater than 0.75
(Streiner and Norman, 1995) and good for individual decision making at a level of 0.90
(Nunnally, 1996). Semmes Weinstein Monofilament testing (Mackin et al., 2003), motor testing
(Brandsma et al., 1995) and prehension skills (Sollerman and Ejeskar, 1995) have been shown to
be reliable with populations and components non specific to SCI. This study provides strong
evidence to support the reliability of the GRASSP when administered by trained clinicians
(ICC’s ranged between .84 - .98) and repeated assessments done by the same or different
assessor/s render reliable results. A change in the measurement of the GRASSP can confidently
be attributed to a clinical change in impairment of the upper limb; And the sample size required
for clinical trials using the GRASSP as an outcome measure would be reduced.
Validity - Two types of validity were established in this study. First, construct validity
was demonstrated by using the theoretical basis to develop a measure that was able to define a
broader range of findings (more sensitive), and in this particular case broader than the current
“gold standard”, ISCSCI. Greater accuracy was one of the underlying requirements set for the
GRASSP during the developmental phase as it was intentionally created to be a more accurate
representation of impairment in the case of sensation. This was to be accomplished by using a
more reliable sensory modality (SWM) with a greater range of response levels and by including
palmar test locations. To ensure a more accurate strength test, more than one muscle per
myotome was incorporated. Although the sample is small when grouped by ISCSCI level, the
specific muscles: anterior deltoid (C5), extensor digitorum, flexor policis longus (C8) and first
75
dorsal interosseus (T1) provide important information regarding innervation. The greatest
amount of discordance was level specific (C5, C8 and T1) and predominantly due to the above
mentioned muscles. The added value of the additional elements to the sensibility and strength
testing showed that upper limb impairment was more accurately defined by the GRASSP than
previous approaches. ISCSCI sensory assessment fails to accurately represent the sensory status,
as sensation from only the dorsal side of the hand is assessed. In addition, the use of light touch
and pinprick as the test modalities in the ISCSCI is likely to affect the robustness of the construct
validity. Furthermore, the ISCSCI motor assessment inflates motor impairment (individual’s
appear more impaired than they really are), due to the method by which myotomal levels are
assigned.
Concurrent validity of a measure is determined by comparing a new test to related,
existing measures in the field. Since a specified upper limb measure of impairment for
tetraplegia did not exist, the best available functional measures were used. Concurrent validity is
determined when a new test shows the anticipated associations to the comparator measures used.
The subtests within the GRASSP show positive and significant associations with the functional
measures used in the field. The subtests which reflect impairment show moderate associations
with the SCIM and CUE, and the subtests which measure impairment within a functional
paradigm (prehension) show stronger associations with the SCIM and CUE. Finally, self
perception of function had the highest association with the GRASSP lending support to the
theory that patients can detect meaningfulness during reporting based on their perceptions.
The purpose of the GRASSP is to describe a profile of upper limb impairment after
tetraplegia. The GRASSP defines the impairment of the upper limb and how the sensory and
motor components affect hand function and is intended for use with the acute, post-acute and
76
chronic tetraplegic individual. Repeated use over time should be able to provide a recovery
profile, whether due to spontaneous recovery, pharmacological interventions, or restorative
upper limb therapies. The GRASSP has been designed to capture changes in neurological
recovery rather than function alone although not all changes in neurological status are large
enough to be realized functionally. Furthermore, improved function, in some cases, may not be
associated with neurological recovery and functional measures in interventional studies are not
sensitive enough to detect small gains. Therefore, measuring impairment over the post injury
course is imperative to determine how much neurological recovery is actually occurring, and
how it relates to functionality.
We tested the GRASSP with a sample of individuals with chronic SCI in order to
diminish the effects of maturation when establishing reliability, thereby ensuring that the
individuals in the sample were considered stable as per the criteria of Wright and Feinstein,
(1992). The strength of the relationship between impairment and function may not be as strong in
more acute populations. Our ongoing longitudinal study is examining the temporal changes in
the three domains to determine their relationship to functional outcome of the upper limb.
We have shown the reliability of the GRASSP, demonstrating its use as a repeatable
measure. The strategies employed to reduce observer variability including the engagement of
experienced clinicians as examiners (occupational therapists/ physiotherapists) with SCI
experience, and the comprehensiveness of the training undoubtedly contributed to our findings.
Having relevant clinicians involved in assessment of neurological status improves consistency.
The results have provided the necessary evidence to confirm that the GRASSP has the
psychometric properties of reliability and validity and is ready for widespread use in cross
sectional studies. The global score will continue to be used as a composite of five subtest scores
77
as there is importance in having a score or set of scores that distinguish change among the
different domains tested by the GRASSP. Sensation and strength scores can be utilized to define
deficit of the upper limb as it relates to spinal cord function and the prehension scores can be
utilized to determine how the components (sensation and strength) impact function. The
individual domains are important to upper limb function individually, therefore, presenting five
subtest total scores on a cluster bar graph will remain the process of presentation of the GRASSP
scores.
78
CHAPTER VI: RESULTS PART 3
6.1 RELATIONSHIPS OF SENSORY AND MOTOR IMPAIRMENT TO
PREHENSION AND UPPER LIMB FUNCTION
6.2 Abstract
It is understood that the degree of impairment is negatively associated to declined
function such that as impairment decreases functional ability improves. The proportion of
change in function that is attributed to sensation and strength has not been well defined for
the upper limb after tetraplegia. While the association of strength to function is well
established in SCI, the association of sensation to function is not. As a result, the role of
sensation and the need to assess it has received less attention in the field. Objectives: 1) to
examine the relationship of impairment (sensibility, strength, and prehension) domains
within the GRASSP to upper limb function. Methods: A cross sectional study was
conducted; all participants had varying degrees of tetraplegia and were neurologically stable,
(n=72). The study protocol included administration of the GRASSP, ISCSCI, CUE and
SCIM during the same session. Hypothetical models which represented the relationships of
impairment and function were tested with both analyses. Analysis: Multiple linear regression
analysis and structural equation modeling were conducted to estimate the effect of the
impairment domains on upper limb function. Results: Multiple linear regression confirmed a
strong association between sensation and strength to prehenssion, but not a strong association
of strength to upper limb function. Structural equation modeling confirmed the association
of sensation, strength and prehension were strong with upper limb function. Structural
equation modeling showed a perfect fit with the data and the goodness of fit indices were
well above the accepted thresholds in [square brackets] (Chi-square = 14.3, p = 0.11,
79
[insignificant p-value]; Comparative Fix Index = 0.99, [above 0.96]; Tucker-Lewis Index
=0.97, [above 0.96]; Root Mean Square Error of Approximation = 0.09, [less than 0.10];
Standardized Root Mean Square = 0.02, [less than 0.09]). The model explained 72% of the
variance in “sensorimotor upper limb function”. Conclusions: The results indicated both
sensation and strength influenced upper limb function directly and indirectly and supported
the hypothesis that prehension was the mediator of sensation and strength for upper limb
function. Strength and sensation played significant roles in ability to perform prehension
tasks which subsequently had a strong association to upper limb function. The results are
significant because the effect of sensation and strength on prehension and upper limb
function provide substantive evidence that supports: 1) therapeutics to be targeted not only
towards strengthening of musculature of the upper limb and also towards promoting sensory
recovery of the palmar surface of the hand, 2) therapy to be applied within a meaningful
functional context, and 3) focus on interventions that target the components of upper limb
function that require the integration of sensory and motor function.
80
6.3 Introduction
Upper limb function is integral to global function and independence for individuals
who suffer from traumatic tetraplegia, therefore, restoration of upper limb function is of great
significance to this population (Anderson, 2004). The more normal and exquisite the upper
limb function becomes after tetraplegia, the more functional and independent an individual
can become. Assessment and restoration of upper limb function for individuals with
tetraplegia entails a multi faceted approach. To date, the impact that an improvement in
impairment has had on function has been defined as positive (Dahlberg et al., 2003; Tooth et
al., 2003; Curt and Dietz, 1996; Schonherr et al., 1999; Mizukami et al., 1995), however, the
amount of change in impairment that influences a specific or meaningful change in function
remains undefined. Furthermore, the type and nature of changed impairment that influences
function also remains undefined. Ultimately, the relationship between impairment status and
functional status is not well established or understood (Ota et al., 1996; Sinnott et al., 2004)
and a stronger understanding of impairment and its role in upper limb function could provide
valuable information to support clinical decision making. This is turn would enhance
recovery through the application of targeted therapeutic interventions (Kirshblum and
O’Connor, 2000).
Exploring the role of specific impairment domains (sensation, strength and
prehension) which contribute to “upper limb function” (reach, prehension, and manipulation)
was the concept of interest. It was conceptualized in the theoretical framework (Figure 1,
Chapter 1) that all three domains did play a role in upper limb function but the contribution
of each component was unknown and where the intermediate relationship (integration)
existed among impairment domains was not fully understood. We did hypothesize within the
framework that sensation and strength are integrated and the mediator was prehension. There
81
is limited literature that discusses these relationships specific to tetraplegia; motor strength
and its relationship to global function is often documented (Fisher et al., 2005) while, the
sensory relationships specific to global function and upper limb function remain under
reported. The specialized field of tendon transfer in tetraplegia reports the significance of
intact sensation (measured by two point discrimination) pre-operatively and its positive
impact on outcomes post-operatively (Moberg, 1976; Sollerman and Ejeskar, 1995).
Furthermore, some preliminary evidence from the Field-Fote laboratory, at the Miami
Project, Cure for Paralysis has reported improvements in sensation, measured by Semmes
Weinstein Monofilaments after a three week intensive massed practice and somatosensory
stimulation protocol. They also found that overall hand function was most improved for
individuals receiving massed practice and somatosensory stimulation, versus just massed
practice or somatosensory stimulation, or conservative management alone (Beekhuizen and
Field-Fote, 2005; Hoffman and Field-Fote, 2007; Beekhuizen and Field-Fote, 2008).
The relationship between the recovery of sensation of the hand is considered to be
activity-dependent (Lundborg, 2000); alterations in hand activity and tactile experience play
a role in the recovery of hand function. The feedback system between the hand and the brain,
otherwise known as sensation operates with continuous proprioceptive and tactile input,
which is coordinated with the feed forward systems in the brain. These two systems are
prerequisites for the regulation of grip force, grasp, and grip speed (Johansson, 2002;
Lundborg and Rosen, 2007). These concepts are well defined within the peripheral hand
injury literature; however, have yet to be established for individuals with tetraplegia. Thus,
based on the significance of sensation to the recovery of hand function in other populations,
sensation was included as a separate domain within the construct of the GRASSP. There is
some evidence to support intact sensation precedes motor recovery (Poynton et al., 1997)
82
specific to SCI, which can be related to post injury cortical and tract plasticity. Ultimately,
establishing the significance of sensation and its role in recovery in tetraplegia is not as
advanced as in other upper limb and hand research fields such as: stroke and peripheral hand
injury.
One reason why impairment had not been clearly connected to function for the
tetraplegic population was because specific tests which quantify impairment with sensitivity
had not been used in the SCI population, particularly sensory tests. Furthermore, the methods
used to analyze the relationships between impairment and function have not been
sophisticated enough to define the factors which clinicians would know to be significant and
sometimes mediators. The GRASSP (Kalsi-Ryan et al., 2009) was developed to incorporate
three domains which, in detail, assess sensation, strength and prehension of the upper limb
for individuals with tetraplegia. The impairment data from the GRASSP has provided an
approach to examine sensorimotor components and their relationship to upper limb function.
Ultimately, the objective of this analysis was to determine the association between
impairment domains (sensation, motor and prehension) and the concept of “sensorimotor
upper limb function” for individuals with chronic tetraplegia. Table 17 provides the scores of
each subtest in the GRASSP for reference.
83
Table 17: GRASSP Scoring Details (Subtest and Item Scores and Ranges) Subtest Item Score
Number of Items
Subtest Total Score
Range
Intended Meaning of Score
Dorsal Sensation 0-4 3 0 - 12 0-4 – C6 5-8 – C7 9-12 – C8
Palmar Sensation 0-4 3 0 - 12 0-4 – C6 5-8 – C7 9-12 – C8
Strength 0-5 10 0 - 50 0-10 – C5 11-15 – C6 16-25 – C7 26-40 – C8 41-50 – T1
Prehension Ability 0-4 3 0 - 12 0-6 – C5-C6 7-12 – C7-T1
Prehension Performance
0-5 6 0 - 30 0-5 – C5-C7 6-10 – C5-C7 11- 15 – C5-C7 16-20 – C5-T1 21-25 – C5-T1 26-30 – C5-T1
6.4 Methods
Ethical approval was attained at all institutions participating in the study and the
analysis was conducted on a dataset collected to develop and establish the psychometric
properties of the GRASSP. A sample size of 40 people was calculated based on an estimated
ICC of 0.80 or greater with an alpha of 0.05. The minimum estimated sample size was 39 for
3 repeated measures according to Donner & Eliasiw’s estimation curves (Donner and
Eliasiw, 1987). Design: A cross sectional, multi centre trial was conducted to collect
impairment and functional data on a sample of n = 72. Seven centres participated in the trial
(three European and four North American). Inclusion/Exclusion: Individuals with chronic
traumatic tetraplegia who were neurologically and medically stable, between the ages of 16
and 65 and able to provide informed consent were included in the study. Individuals with
84
moderate brain injury who were neurologically unstable, and those with any other pathology
causing upper limb impairment were excluded from the study. As a result of the multi centre
collaboration a greater number of datasets were available, allowing the n value to increase to
72. Data collection and management: Two workshops (one in Europe and one in North
America) were conducted to train the examiners regarding the study protocol and appropriate
use of all study measures. Instruction and demonstration in the administration of the primary
measure, GRASSP and secondary measures were provided. The ISCSCI (Marino, 2000), The
SCIM (Catz et al., 2004), and CUE (Marino et al., 1998) were provided to all examiners
involved in the study. All data were collected by the examiners, de-identified and entered
into excel tables; once entered all data was transferred to the central site for pooling. Data
from all centres were aggregated, cleaned and analyzed using SPSS 17.0 and M-Plus 5.2.
The GRASSP is a multi domain impairment measure specific to the upper limb for
individuals with tetraplegia; details of the development and content are available in a
previous manuscript (Kalsi-Ryan et al., 2009) The ISCSCI is a classification measure which
describes the severity of injury for individuals with SCI; this measure was used to describe
the sample in the results. The SCIM is a global measure of function specific for individuals
with SCI, used to define the global function and independence of the sample in this study.
Inter rater reliability is above 0.8 when assessed by agreement statistics for most SCIM
items, and ICC for the total score is 0.94. Concurrent validity of the SCIM with the FIM is
0.79 (Itzkovich et al., 2007). Within the SCIM, there are three subscales (self care,
respiration and sphincter management, and mobility) and in this analysis the SCIM self care
subscale (SCIM-SS) was used as one of the representations of upper limb function. The
SCIM-SS includes items solely related to the use of the upper limb; therefore, comparisons
between the GRASSP subtests are made with the SCIM-SS, rather than the total SCIM score.
85
Subscales of the SCIM are reliable and useful quantitative representations of the specific
constructs of independence in SCI (Catz et al., 2007). Some researchers have found the
subscales to be more specific for targeted analyses (Rudhe and van Hedel, 2009; van Hedel,
2009).
The CUE is a subjective questionnaire that determines ones own perception of
functional ability. The questions asked are related to one’s perception of how difficult a task
may be. The CUE is embedded with questions that fall into the three components of upper
limb function reaching tasks, prehension tasks, and manipulation tasks, scores for each task
are added for a total CUE score. Psychometric properties of the CUE have been reported as
0.92 for test retest reliability, tested by Cronbach’s alpha and 0.74 for concurrent validity
with the Functional Independence Measure, tested by Pearson correlation coefficient (Marino
et al., 1998). GRASSP, SCIM and CUE results collected during the same visit were extracted
from the dataset for the analysis of impairment and “upper limb function”. All measures used
in the study are available in Appendices 1 and 3.
Analytic Plan: A priori we anticipated there would be a positive relationship between
the impairment domains and upper limb function; specifically, strength would play a stronger
role than sensation in upper limb function. However, the strength of the relationship was
unknown. We anticipated that there was an intermediate relationship where strength and
sensation would influence prehension and as such, prehension would then have an
association with upper limb function. Based on the intermediate role that prehension plays in
the construct of “sensorimotor upper limb function” it was anticipated the core and integrated
elements of impairment would have unique roles in “sensorimotor upper limb function”.
Two methods of analysis were used to explore the relationships between impairment
and function: Multiple linear regression and structural equation modeling. Three
86
hypothesized models were designed for multiple linear regression. The relationship between
several independent variables and a single dependent variable were examined, however, due
to the fact that only one dependent variable per model could be explored, three models were
tested and the collective results were used to define the relationships between impairment
and function. Selection of the independent and dependent variables was based on the
theoretical framework both the domains construct. The independent variables used in the
multiple linear regression were the palmar sensation subtest total score, the strength subtest
total score, and prehension performance subtest total score, right side data only (see Table
17). The SCIM-SS was used as a representation of upper limb function and the CUE was
used as a representation of self perceived upper limb function. In model A the palmar
sensation subtest total and strength subtest total scores were used as the independent
variables with the prehension performance subtest total score as the dependent variable. In
models B and C the palmar sensation subtest total, strength subtest total and prehension
performance subtest total scores were used as the independent variables with the SCIM-SS
and CUE respectively, as the dependent variables. Multiple linear regression was conducted
to determine the role of the impairment variables and their influence on upper limb function.
The second analysis approach examined the relationship between the same variables
and structural equation modeling was used. Structural equation modeling is a general
approach to multivariate data analysis, used to study complex relationships among variables.
Structural equation modeling has become an important data analysis technique, and can be
used as a language to formulate social science theories; and can define the relationships and
pathways between variables. Structural equation modeling is used to describe directed
dependencies among a set of variables and provides an opportunity to test models with
multiple dependent variables and provides a value of both direct and indirect effects of all
87
variables. Some of the variables can be hypothetical or “unobserved”. The “unobserved”
variable is called the latent trait variable. In this case the latent trait was sensorimotor upper
limb function. The SCIM-SS and CUE were the indicators of the latent trait. Essentially
structural equation modeling is a confirmatory technique which confirms a specified model
(the hypothesis). The results define redundancies and pinpoint aspects of conflict with the
data (Wothke, 2010). The fit of the model is confirmed by the analysis, which is known as
the “goodness of fit” (Hooper et al., 2008). Statistics of fit determine how well the specified
model (hypothesis) fits the actual data. A chi-square test is conducted to evaluate the overall
model fit, which assesses the magnitude of discrepancy between the sample and fitted
covariance matrices. A large chi-square with an insignificant value when compared to p<0.05
is indicative of a good fit of the model. The chi-square, although not the most rigorous index
of fit is used and often accompanied by other indices. The root mean square error of
approximation (RMSEA) evaluates how well the model fits the population’s covariance and
is sensitive to the number of estimated parameters in the model. A value less than 0.10 is
indicative of a fair fit (Black, 1993). The RMSEA is used when the number of estimated
parameters is low in the case of this model, only one parameter is estimated. The
standardized root mean square residual (SRMR) is the square root of the difference between
the residuals of the sample covariance matrix and the hypothesized covariance model. A
value of less than 0.09 is indicative of a good fit. The SRMR is used when there are varying
ranges of scales among indicators, which is the case in the model tested. The comparative fit
index (CFI) accounts for the sample size, all latent variables are uncorrelated and compared
to the sample covariance matrix with the null model and the Tucker-Lewis Index (TLI) of fit
is used when a small sample size is being analyzed and can point out a poor fit when other
indices are pointing to a good fit. A value above 0.96 indicates a good fit for both of these
88
indices. The CFI considers the small sample size and the TLI considers simple models. Thus,
the selection of indices was specific to the model hypothesized (Hooper et al., 2008). All
hypotheses were examined simultaneously by specifying one structural equation model
labeled Model D (Figure 7). In this analysis, we incorporated one latent variable of
“sensorimotor upper limb function”, which brought several benefits to the measurement of
variables in this model, hence the accuracy of its structural relations (Tomarken and Waller,
2005).
Figure 7 defines the hypothetical model for which structural equation modeling was
conducted to determine the relationship of impairment to upper limb function. In structural
equation modeling it is necessary to establish a latent trait variable otherwise known as an
unobserved value which is indicated or estimated by observed variables. In the case of our
model, we established our latent trait variable to be “upper limb function”. “Upper limb
function” was indicated by the SCIM-SS and the CUE component (reach, prehension,
manipulation) scores. In structural equation modeling it is more reliable to have at least three
variables estimate the latent trait (Sudano, 2010); therefore, the CUE was split into the three
components (reach, prehension, manipulation) that represent upper limb function (Shumway-
Cook and Woollacott, 2007; Kapandji, 1970).
89
Figure 7: Model D, Hypothetical Model of Impairment and Sensorimotor Upper Limb Function in Tetraplegia to be tested with Structural Equation Modeling.
Legend for Figure 7: Palmar sensation is represented by the palmar sensation subtest score, strength is represented by the strength subtest score, and prehension is represented by the prehension performance subtest score within the GRASSP. SCIM-SS is the self care subscale total, CUE-reach, the responses for reaching tasks totaled, CUE-prehension, the responses for grasp tasks totaled and CUE-manipulation, the responses for manipulation tasks totaled within the CUE. The latent trait variable is sensorimotor upper limb function. This model represents the hypothesized association of palmar sensation, strength and prehension, and the relationship of the three impairment domains to sensorimotor upper limb function in the same model. The strength of associations is not pre-determined.
6.5 Results
Sample - The data used in this analysis included a multi centre/multi national cross
sectional sample. The total sample consisted of 72 individuals with chronic tetraplegia
ranging from 6 months to 20 years post injury. Distribution of the sample according to the
90
ISCSCI is defined in Table 18. Approximately 52.5% of the individuals fall between the C6 -
C7 motor levels while approximately 66% fall between the C4 - C6 sensory levels. The AIS
which defined completeness, classified 39% (n=28) of the sample to be AIS A complete, and
61% (B-25% n=18; C-19% n=14; D-17% n=12) of the sample as AIS B, C or D incomplete.
Table 18: Demographics of the Sample Based on ISCSCI and AIS Classification, n=72 (presented as n values and percentages of sample)
ISCSCI Levels Motor Level
n (%)
Sensory Level
n (%)
AIS Classification
n (%) Right Left Right Left A B C D
C2-C4 10 (14) 14 (19) 29 (40) 29 (40) C5 10 (14) 9 (13) 11 (15) 9 (13) C6 23 (32) 21 (29) 17 (24) 19 (26) C7 17 (24) 16 (22) 8 (11) 6 (8) C8 4 (5) 5 (7) 1 (2) 2 (3)
T1 and below 8 (11) 7 (10) 6 (8) 7 (10)
28 (39)
18 (25)
12 (17)
14 (19)
Legend for Table 18: ISCSCI – International Standards for Neurological Classification of Spinal Cord Injury, AIS – ASIA Impairment Scale
Prior to conducting both analyses all variables were correlated with one another.
Table 19 presents the Pearson correlation coefficients between the measures of upper limb
function and the variables for the impairment domains. The values in table 18 reveal that
sensory, strength and prehension are significantly and positively associated with each other,
and they are all correlated with the measures of upper limb function.
91
Table 19: Correlation Matrix of all Variables Included in the Multiple Linear Regression and Structural Equation Modeling Sensory Strength Prehension SCIM CUE
Total CUE Reach
CUE Prehension
CUE Manip
Sensory 1 - - - - - - - Strength 0.685 1 - - - - - - Prehension 0.656 0.811 1 - - - - - SCIM 0.636 0.692 0.689 1 - - - - CUE Total 0.693 0.705 0.730 0.855 1 - - - CUE Reach 0.578 0.518 0.597 0.785 0.924 1 - - CUE Prehension
0.653 0.731 0.708 0.754 0.922 0.753 1 -
CUE Manipulation
0.692 0.771 0.731 0.769 0.858 0.617 0.839 1
Legend for Table 19: All correlation values are Pearson correlation coefficients with p<0.0001. Sensory represents the palmar sensation subtest score, strength represents the strength subtest score, and prehension represents the prehension performance subtest score within the GRASSP. CUE-reach, the responses for reaching tasks totaled, CUE-prehension, the responses for grasp tasks totaled and CUE-manipulation, the responses for manipulation tasks totaled within the CUE. The CUE Total is the total CUE score.
Multiple linear regression analysis rendered the strength of association between the palmar
sensation, strength and prehension variables to upper limb function. Figure 8 presents the
models and the relationships between the independent and dependent variables; each model
presents different variables to explore the multiple associations. The standardized coefficient
was used to report the strength of association and the significance level is denoted by the
asterix. The SCIM-SS is representative of upper limb function and the CUE is representative
of self perceived upper limb function Table 20 presents the unadjusted values derived from
the multiple linear regression analyses, these values differ from the standardized coefficients
which were used to determine the linear relationships between the independent and
dependent variables.
92
Table 20: Unadjusted Multiple Linear Regression Values for Palmar Sensation, Strength and Prehension when Compared to SCIM-SS and CUE values Prehension SCIM-SS CUE Palmar Sensation 0.499 0.393 2.661 Strength 0.505 0.122 0.420 Prehension - 0.179 1.119 Legend for table 20: Palmar sensation represents the palmar sensation subtest score, strength represents the strength subtest score, and prehension represents the prehension performance subtest score within the GRASSP. SCIM-SS – represents the Spinal Cord Independence Measure selfcare subscale and CUE – represents the Capabilities of the Upper Extremity Questionnaire total score.
Figure 8 presents the results of the multiple linear regression analysis. Model A
defines palmar sensation and strength to have strong and significant associations to
prehension, therefore, palmar sensation and strength of the upper limb play a significant role
in mediating prehension. Model B defines palmar sensation and prehension as having strong
and significant associations, and motor as having no direct association to upper limb function
as represented by the SCIM-SS. Model C defines palmar sensation and prehension as having
strong and significant associations, and motor as having no direct association to upper limb
function as represented by the CUE.
93
Figure 8: Results of Multiple Linear Regression of Impairment Variables and Upper Limb Function.
Legend for Figure 8: Palmar sensation represents the palmar sensation subtest score, strength represents the strength subtest score, and prehension represents the prehension performance subtest score. SCIM-SS is the self care subtest score in the SCIM and the CUE is the total score for the questionnaire. Model A represents the model tested to determine the association of palmar sensation and strength impairment to prehension. Model B represents the model tested to determine the association of palmar sensation, strength and prehension to the SCIM-SS. Model C represents the model tested to determine the association of palmar sensation, strength and prehension to the CUE.
Multiple linear regression was used to determine the relationships of the domains as
they were defined in the theoretical framework through the three hypothesized models in
Figure 8. Multiple linear regression was not conducted as a stepwise process where variables
are eliminated when they are insignificant in a model. Collectively, the three models
94
combined showed that palmar sensation and strength contributed significantly to prehension.
However, only palmar sensation and prehension contributed significantly to upper limb
function as defined by the SCIM-SS and CUE. The role of strength was not confirmed as a
significant contributer upper limb function.
Structural equation modeling rendered the strength of association between
impairment, function and the latent trait variable of sensorimotor upper limb function. Figure
9 presents the structural equation modeling results for model D which defined the effect of
impairment on sensorimotor upper limb function. The structural equation modeling showed a
perfect fit with the data and the goodness of fit indices were well above the accepted
thresholds. Chi-square = 14.3, p = 0.11; Comparative Fix Index = 0.99, Tucker-Lewis Index
= 0.97 and the Root Mean Square Error of Approximation = 0.09, Standardized Root Mean
Square SRMR = 0.02. The model explained 72% of the variance in “sensorimotor upper limb
function”. Prehension has a significant positive effect on upper limb function and strength
and palmar sensation both have a direct and indirect effect through prehension on upper limb
function.
95
Figure 9: Model D, Structural Equation Model for Impairment and Sensorimotor Upper Limb Function after Tetraplegia.
Based on the structural equation modeling, palmar sensation showed a direct and
indirect relationship to upper limb function. The relationship mediated through prehension is
larger (.19+.32) than the direct relationship (.31); but both direct and indirect relationships
are statistically significant. The relationship mediated through prehension is larger (.68+.31)
than the direct relationship to upper limb function, but both direct and indirect relationships
are statistically significant. Therefore, sensorimotor upper limb function can be predicted by
palmar sensation and strength through prehension. As structural equation modeling clarified
the relationship of strength with respect to upper limb function, structural equation modeling
is the preferred model to confirm the hypothetical model (Figure 7). The structural equation
modeling accounted for all variables in one model which demonstrated that strength is
96
related to upper limb function and why in contrast strength was not a significant factor in
upper limb function with multiple linear regression.
6.6 Discussion
It was hypothesized that sensation and strength are mediated through prehension
when affecting function, which was a feature of the theoretical framework for GRASSP
development. The data confirmed this hypothesis. A priori we assumed it was important to
measure all domains of impairment to reflect function accurately. There is little evidence
available to define the inter-relationship between sensation and strength on upper limb
function in tetraplegia. As interventions are more likely to target neurological substrates to
enhance recovery, the most common outcome that should be assessed to determine efficacy
is degree of impairment. Impairment change and how impairment impacts function and
independence is one factor that is used to determine whether an intervention demonstrates
efficacy. To date there is not a strong understanding of how degree of sensory and motor
impairment influence function. Although, some researchers have recognized that poor
functional gains after tendon transfer for increased movement are likely to be related to poor
sensory function (Sollerman and Ejeskar, 1995; McDowell et al., 1986).
The present research provides the evidence to verify that there is a relationship
between sensorimotor impairment and function of the upper limb for individuals with chronic
tetraplegia. It should now be possible to determine if a positive change in sensation results in
improved upper limb function. The way in which sensation and strength are integrated
through prehension and how they directly impact upper limb function is supported by
confirming Model D. The relationships of sensation and strength mediated through
prehension show that strength is a stronger factor. These strong relationships of upper limb
97
function mediated through prehension support the concept of rehabilitation processes
incorporating the use of functional tasks specifically prehension retraining protocols. Using
the structural equation modeling approach across the recovery period could assist in
determining the magnitude of impairment change that will lead to different levels of
functional change.
Sensation is reported to have a significant impact on hand function (Shumway-Cook,
2007; Brand and Hollister, 1991). Recovery of sensation after peripheral hand injury is very
important for the return of function (Moberg 1986; Lundborg and Rosen, 2007). And there is
some evidence to suggest that improvement in sensation during recovery is mediated at the
cortical and tract level (Beekhuizen and Field-Fote, 2008 and Green et al., 1999). The
relationship of hand sensation to upper limb function has not been well established for
individuals with tetraplegia. Structural equation modeling substantiates that sensation is
important, as a component to both prehension and upper limb function. As there is a
significant relationship between sensation and prehension (.19 p<0.05) and then a significant
relationship between prehension and upper limb function (.32 p<001), the role of palmar
sensation is definitive. Therefore, it is of great importance that recovery of sensation be
enhanced after SCI. Targeted sensory retraining of the hand within a functional paradigm
(task specifically-prehension) may be necessary to refine functional ability during the
rehabilitation phase. In this analysis it is noted that having good sensation indirectly impacts
upper limb function, not only prehensile and manipulation tasks. Albeit not specific to the
upper limb, Poynton in 1997 demonstrated that patients with sensory sparing in dermatomes
regain at least a grade 3 in strength for the associated myotome over time, suggesting that the
mere presence of sensation enhances sensorimotor recovery.
98
The recovery of upper limb strength after tetraplegia has been reported to be very
important for the restoration of function (Fisher et al., 2005; Taylor et al., 2002; Rothwell et
al., 2003). The impact that strength has on function in tetraplegia is so well established; that
other factors such as sensation are not always considered to have the same importance in
recovery of function. Strength plays the dominant role in prehension and this is probably true
for many other motor tasks because intact sensation without strength is not functional. The
meaning of the analysis supports the theory that if strength is available it should be within a
functional paradigm to be optimal.
The research findings provide strong evidence to support two aspects of therapeutic
approaches for individuals with tetraplegia. First, interventions should target both the motor
and sensory aspects of recovery. Therapies need to be applied so that the palmar surface of
the hand is stimulated, to promote activity optimizing sensory activity leading to recovery.
The most ideal interventions would provide sensory stimulation while the hand engages in
functional tasks; this method would target the mechanisms that in theory influence cortical
plasticity and reorganization. Secondly, interventions targeted towards recovering motor
function must be in a functional context, with varying degrees of force generation and
sequencing of muscle activation. Reaching transports the hand to the location of the task. The
reach enables the prehension which is modulated by the sensation and strength of the hand.
Once prehension has been engaged, sensibility of the hand and strength of the small muscles
in the hand enable manipulation. Therefore, therapy interventions should incorporate task
oriented protocols which engage the entire upper limb and incorporate sensorimotor
functions.
The GRASSP assesses impairment with acuity in three domains. During the
development of the GRASSP the theoretical framework guided item generation to be
99
anatomically, neurophysiologically and functionally relevant. Sensorimotor upper limb
function has been tested in this research using data from a sample with chronic SCI. The
structural equation modeling confirms the concepts and components of the construct
particularly the domains and their individual and integrative importance.
The robust findings in model D from this analysis were possible because the data
comprehensively represented the impairment of the upper limb in tetraplegia. Quantifying
impairment more precisely has enabled the investigators to establish the strength of the
relationships and integration of palmar sensation, upper limb strength and prehension to
upper limb function. The next steps will be to test the degree of these relationships and
integration during the course of recovery.
100
CHAPTER VII: GENERAL DISCUSSION
7.1 Summary of Findings
Upper limb impairment after tetraplegia was defined as the “phenomenon of interest”
that we intended to measure; subsequently, becoming the purpose of the GRASSP. The
theoretical framework was developed which incorporated the functional, anatomical and
neurophysiological mechanisms of normal upper limb function; and the pathophysiology of
SCI. The framework incorporated concepts and theories of sensorimotor control/
sensorimotor learning (Kandel et al., 2000), and the interactions between the
cognitive/sensory/motor systems that play significant roles in normal upper limb function
(Shumway-Cook and Woollacott, 2007). The core elements of impairment (sensation and
strength) and the integrated element of impairment (prehension) were included as domains of
the construct. The three components of upper limb function (reach, prehension and
manipulation) were represented within the items of the domains of the construct. Therefore,
the information rendered by the GRASSP represents the sensation, strength and prehension at
the periphery. However, how these domains relate to the concepts which define the construct
“sensorimotor upper limb function” and allow us to develop knowledge in how best to
influence the concepts, theories and components that define upper limb function.
The overall objective of this thesis was to complete the development of the GRASSP.
Development of the GRASSP has been a staged process and the findings in this thesis have
finalized the measurement properties and the concepts important for practical use of the
measure. Data was collected over an eight month period with a protocol designed for the
study of reliability and validity. The dataset provided a number of data elements which were
used for multiple analyses. The initial analysis which was not included in this thesis
evaluated the items in the initial GRASSP to finalize inclusion and exclusion of the items in
101
GRASSP Version 1.0 (Appendix 1). The following sections will comment on my
accomplishments of the specific objectives listed in Chapter 2.
7.1.1 Objective One: Development of the Scoring System
The first set of analyses reported in this thesis, indicated how the scoring system was
determined for reporting output from the GRASSP. It was established that assigning a spinal
cord level to numeric values would diminish the true meaning of scores. Therefore, it was
first determined that the items showed a predominant cumulative predictive pattern, thus, the
subtest items could be summed and used as subtest total scores to define impairment of the
upper limb. Theoretical basis of the GRASSP and spinal cord anatomy provided the
definitions for the levels of GRASSP scores. The ISCSCI and AIS did not discriminate the
sample similarly to the GRASSP. The GRASSP discriminative score ranges were not
concordant with the ISCSCI due to the differences in how impairment is defined by both
measures. The GRASSP defines the impairment as a numeric value which incorporates any
amount of neurological intactness between levels, whereas, the ISCSCI defines the
impairment based on the most caudal “normal” spinal cord level. For the purpose of
quantifying impairment; a numeric value which represents all of the impairment is an optimal
method to define upper limb impairment. Essentially, numeric values derived at the periphery
which represent the spinal cord spatially show greater precision in defining impairment.
Based on the subtest scores being meaningful as totals, a method to define all scores in one
place was derived. Because scores could not be added for one global score as there is no
internal consistency among domains, normalizing the subtest total scores and plotting the
values on a bar graph was the method derived to present GRASSP output. Summary of
102
objective one; the scoring system of the GRASSP has been developed to render numeric
values which are meaningful, robust and interpretable.
7.1.2 Objective Two: Psychometric Testing
Psychometric properties were then established to ensure that GRASSP subtest scores
were reliable and valid. Reliability of a measure ensures that the change noted between
assessments can be attributed to a clinical change rather than error in measurement. As
hypothesized inter rater and test retest reliability for all subtests within the GRASSP were
above an ICC of 0.80 (0.83-0.96 and 0.86-0.99 respectively). Therefore, confirming that the
GRASSP can be administered as a repeated measure by the same or multiple examiners to
determine change in a clinical setting or in a clinical trial. This level of reliability would
provide values for a smaller sample size calculation, which is beneficial in SCI research
where incidence and prevalence is low and large samples are difficult to recruit. Construct
validity was confirmed by discordance of the sensibility and strength subtests within the
GRASSP to the ISCSCI sensory and motor items; the subtests showed concurrence with the
SCIM, SCIM-SS and the CUE. Impairment domains had the strongest concurrence with self
perception of function (0.57 - 0.83, p < 0.0001). Summary of objective two; the GRASSP has
demonstrated reliability, construct and concurrent validity for use as a standardized upper
limb impairment measure for individuals with tetraplegia.
7.1.3 Objective Three: Confirmation of Theoretical Framework
With the availability of detailed impairment data regarding the upper limb, some
principles of sensorimotor relationships to upper limb function specific to tetraplegia were
defined. It was determined that sensation and strength played a dual role in mediating upper
103
limb function, one which was mediated through prehension and one which directly affected
upper limb function. It was also determined that in the model of measurement, prehension
was one point where sensation and strength converge, to play an integrative role in upper
limb function. The statistical methodology provided the opportunity to establish the
integrative relationships among impairment and function which substantiate current theories
of neuro-rehabilitation. Summary of objective three; sensation and strength and subsequently
prehension all play unique and indirect and direct roles in upper limb function.
Development begun with a theoretical framework which guided the process of
GRASSP development. In summary a novel impairment measure for the upper limb has been
developed. The data captured by the GRASSP provides relevant information in the form of
numeric values. The output is representative of the degree of intactness at the periphery and
spinal cord. The GRASSP is both reliable and valid, therefore, provides meaningful
impairment data that can be used to follow individuals over time, which measures what it is
intended to measure with a broader range of items with greater response levels. Therefore,
the GRASSP is a measure which can be used clinically and in clinical trials as an outcome.
The impairment data that is rendered by the GRASSP is accurate, precise and describes
impairment under a finer lens, enough to establish relationships between function and
impairment. These relationships between impairment and function are meaningful, during
recovery and for the evaluation of the application of new interventions.
7.2 Impact, Significance and Strengths of the Research
The lack of a precise impairment measure specific for the upper limb after tetraplegia
has made it challenging for researchers to quantify the neurological change of the upper limb
and then relate the change to functional gains. The GRASSP measures upper limb
104
impairment for individuals with tetraplegia during the acute to chronic phases of recovery.
The findings of the measure can be used to make the connection between natural recovery
and interventions and their potential influence on impairment. The GRASSP is a clinical
measure and it can easily be incorporated into multiple clinical practice settings. The most
significant outcome of my PhD work has been the production of the GRASSP Version 1.0.
Table 21 summarizes the attributes of the measure. The contribution of the GRASSP Version
1.0 to the field of SCI research will have a significant impact on the next generation of
clinical trials which include participants with tetraplegia, and studies related to developing
methods to restore upper limb function. The GRASSP will provide an opportunity for
researchers to assess the more subtle changes that occur as a result of interventions and
natural recovery. The GRASSP can be administered earlier in the course of recovery than
most current outcomes measures.
Table 21: Attributes of the GRASSP Version 1.0 Graded Redefined Assessment of Strength Sensibility and Prehension (GRASSP) Purpose of the Measure Designed to measure impairment of the upper limb after
tetraplegia Theoretical Framework Defines the underlying anatomical and neurophysiological
concepts and theories which play a role in upper limb function. Explains why the construct for the measure is “sensorimotor upper limb function”. Defines core and integrated elements of impairment which contribute to upper limb function.
Description of Measure Construct: Sensorimotor Upper Limb Function Domains: Hand Sensation, Strength of Upper Limb and Prehension Tasks. Multi-modality measure, consisting of five subtests.
Scoring Consists of five subtests (dorsal sensation, palmar sensation, strength of upper limb, prehension ability, prehension performance). Each subtest renders a numerical value that represents the impairment manifested at the periphery and spinal cord.
Reliability Inter rater and test retest reliability for all subtests within the measure are above 0.80.
Validity Construct Validity - Sensation and strength domains have greater sensitivity in defining the impairment of the upper
105
limb than the current available measures for SCI. Concurrent Validity – The subtests within the measure hold concurrent validity with measures of function used with the SCI population (SCIM, CUE).
Uses of the Measure The intended use for GRASSP is as a clinical assessment of impairment of the upper limb after tetraplegia. The numerical values rendered by the measure can be used to establish relationships between impairment and function for individuals with the chronic tetraplegia.
The GRASSP characterizes the upper limb for individuals after tetraplegia and can be
used as early as an individual is responsive to being able to participate in clinicial
assessment. This allows impairment information to be extracted as soon as possible after
injury. Data collection early after injury allows one to establish a reliable baseline which can
be used to compare to subsequent assessments done in the same facility or across facilities.
Measuring the three separate domains adds value to upper limb assessment. Sensory
and motor domains define the neurological deficit as it relates to anatomy (dermatomes and
myotomes) and prehension defines how the deficits impact function. The relationships of the
three domains characterize the cause of the functional deficit. The GRASSP was not
designed to measure or evaluate compensation or one’s ability to accomplish a task, rather
how the task is performed and the quality of the performance. The value in the information
gathered by GRASSP allows one to understand more about the neurological integrity of the
upper limb and its impact on function, which may in turn allow clinicians and researchers to
understand how to influence the deficit. The multi domain feature of the GRASSP means the
measure can be used in two ways, one to test new approaches (determine what aspect of
deficit is influenced) and two, to follow the natural recovery process. The GRASSP provides
greater precision or sensory reporting than existing tests because of the use of Semmes
Weinstein Monofilaments, the strength testing provides greater robustness to the motor
106
testing due to the testing of additional muscles. The prehension testing defines the
functionality of the neurological integrity. These features of the GRASSP make it a superior
impairment measure for use in clinical trials assessing neurological change in the upper limb.
Reliability of a measure is an important property in the evaluation of change over
time. As such, the reliability of the GRASSP defines it as a measure for use with the SCI
population to follow natural recovery and the effects of interventions. The construct validity
of the GRASSP ensures that the data collected represents sensation, strength and prehension
domains appropriately. Therefore, the GRASSP can be used in the assessment of impairment
with confidence that the domains assess with accuracy and precision.
Concurrent validity with the SCIM is moderate and positive and indicates as
impairment decreases, function and independence improve. Concurrent validity with the
CUE (self perception of function) was the strongest. This finding is significant because it
indicates at least in a chronic sample that “what” an individual reports about their function is
important and correlates with quantitative measures of impairment. The high level of
concurrence between GRASSP and CUE is significant because it provides credence of
consumer involvement in establishing whether the effect of interventions will be
meaningful..
The GRASSP renders data that characterizes the impairment of the upper limb with
greater detail than any available SCI specific measure to date. This enables researchers to
understand the process of spontaneous recovery, determine the effects of interventions and
make connections between impairment and function with greater accuracy, precision and
meaning.
107
7.3 Limitations of the Project
Although this work presents a novel impairment measure of the upper limb for
individuals with tetraplegia, which has incorporated the many stages of measurement
development and shows substantial promise as a new outcome measure in the field, the work
also has limitations.
Limitations of the GRASSP itself are defined by the purpose of the measure. The
GRASSP has been designed as a clinical measure so implementation and access to the test is
simple and inexpensive. However, a clinical measure no matter how reliable always includes
some degree of human error. Implementation of the measure is clearly defined in the
instruction manual; however, administration of the measure may differ slightly due to the
level of understanding and preference of the clinicians administering the test.
Some of the concepts that were considered, but not included in the development of
the measure were stereognosis and haptics. Sensation after injury is often compensated for by
vision and the GRASSP does not account for visual input and its influence on performance.
Despite the sample size being sufficient for the reliability analysis; when the sample
was sub grouped into cervical levels according to ISCSCI the sub group numbers were not
large enough to confirm definitively some of the results noted. In order to conduct a more
rigorous statistical method such as: Rasch Analysis or Mokken Scale Analysis, to determine
the true predictability of the items and subtests, a much larger sample would be required.
The sample used in this study included individuals with chronic tetraplegia; therefore,
future results of similar analyses may differ with more acute individuals. Individuals with
chronic tetraplegia often develop compensatory patterns and functions otherwise known as
“maladaptative patterns”, which are not always accommodated for by the GRASSP.
Although, the prehension domain is designed to differentiate between maladaptive and
108
normal movement patterns, individuals with chronic tetraplegia have very strong
compensatory behaviours which may not be deciphered from normal behaviours.
Furthermore, abilities to complete tasks in the GRASSP are based on experience through
activities of daily living. For individuals who live in the community many of the tasks are
familiar. Whereas individuals who are still acute would be challenged with the same tasks as
they may not have had the opportunity to practice or to perform them. This difference for
acute and chronic individuals may warrant slight modifications in scoring in the future.
The GRASSP was designed by incorporating many elements of physiology and SCI
pathology, however, being a clinical measure there are limitations with respect to how the
GRASSP measures impairment. The prehension performance subtest of the measure only
includes timing and scoring of tasks; force of prehension during tasks is not measured,
therefore, force generated could enlighten clinicians even more regarding what elements to
target in therapeutic interventions.
7.4 Future Directions
One remaining psychometric property, responsiveness, has not been established for
the GRASSP to date. A longitudinal study that will provide the data to evaluate the
responsiveness of the GRASSP is warranted. Currently, a longitudinal study to establish
responsiveness of the GRASSP is being conducted. The current trial includes serial testing
of individuals from the time of injury to one year post injury. The data will be used to
establish the responsiveness of the GRASSP, define a recovery profile of the upper limb after
tetraplegia, and establish elements of minimally clinical important difference. The trial will
provide data to determine the most appropriate scoring of the GRASSP for individuals during
the acute phase after tetraplegia; the impairment relationships to function throughout the
109
continuum of recovery, which can inform clinical decision making; the temporal association
of recovery of the upper limb and how it influences function; and when the optimal
timeframe for intervention and type of modality might be.
This study focused on the development of GRASSP in traumatic tetraplegia.
However, individuals with non traumatic tetraplegia often present with similar upper limb
impairments. The GRASSP will be applied in two non traumatic SCI longitudinal studies to
examine the impairment changes that occur after surgical and drug interventions.
Modifications to the measure may be required to accommodate the non traumatic group.
7.5 Final Summary
What is currently lacking is an impairment measure with robust measurement
properties that assesses the domains most likely to change with new therapeutic interventions
which are directed towards neural repair and recovery. The GRASSP fills this void.The
GRASSP is a measurement tool used to determine the status of hand function which has the
psychometric properties to be implemented in upcoming studies investigating the
fundamental underlying mechanisms and state of integrated sensorimotor hand impairment.
When the sensitivity to detect change in sensorimotor impairment has been established, the
GRASSP will have the potential to be an important assessment tool to specifically detect
changes in sensorimotor function over the post injury process. In clinical trials where the
primary outcome of sensorimotor integration needs to be decoupled to determine efficacy of
interventions, the GRASSP will be useful to determine the integrated contribution of the
underlying neural substrates of sensation and motor function. This new knowledge gained
has the potential to characterize the effectiveness of sensory and motor therapeutic
interventions.
110
REFERENCES
Alon, G., & McBride, K. (2003). Persons with C5 or C6 Tetraplegia Achieve Selected
Functional Gains Using a Neuroprosthesis. Archives of Physical Medicine and
Rehabilitation, 84, 119-24.
Anderson, K.D. (2004). Targeting recovery: Priorities of the spinal cord-injured population.
Journal of Neurotrauma, 21(10), 1371-1383.
Baptiste, D.C. & Fehlings, M.G. (2007). Update on the treatment of Spinal Cord Injury.
Progress in Brain Research, 161, 217- 234.
Beebe, J.A., & Lang C.E. (2009). Relationships and responsiveness of six upper extremity
Function tests during the first six months of recovery after stroke. Journal of
Neurological Physical Therapy, 33, 96-103.
Beekhuizen, K.S., & Field-Fote, E.C. (2008). Sensory Stimulation Augments the Effects of
Massed Practice Training in Persons With Tetraplegia. Archives of Physical Medicine
and Rehabilitation, 89, 602-8.
Beekhuizen, K.S., & Field-Fote, E.C. (2005). Massed Practice versus massed practice with
Stimulation: Effects on Upper Extremity Function and Cortical Plasticity in
Individuals with Incomplete Cervical Spinal Cord Injury. Neurorehabilitation
and Neural Repair, 19(1), 33-45.
Beninato, M., O’Kane, K.S., & Sullivan, P.E. (2004). Relationship between motor FIM and
Muscle strength in lower cervical-level spinal cord injuries. Spinal Cord, 42(9),
533-542.
Bohannon, R.W., & Smith, M.B. (1993). Interrater reliability of a modified Ashworth scale
of muscle spasticity. Physical Therapy, 67(2), 206–207.
111
Bombardier, C. (2002). Sensibility Assessment Form. Manual for Measurement in Clinical
Research Course, Part 2, 51. Toronto, Canada.
Brand, P.W., & Hollister, A.M. (1991). Clinical Mechanics of the Hand, Third Ed.
St. Louis, Missouri, Mosby.
Brandsma, J.W., Schreurders, T.A., Birke, J.A., Oostendorp, R. (1995). Manual muscle
testing: intraobserver and interobserver reliabilities for the intrinsic muscles of the
hand. Journal of Hand Therapy, 8(3), 185-190.
Bruehlmeier, M., Dietz, V., Leenders, K.L., Roelcke, U., Missimer, J., & Curt, A. (1998).
How does the human brain deal with a spinal cord injury? European Journal of
Neuroscience, 10(12), 3918-3922.
Campaign for Cures Website. (2004). General Information. Multiple sources.
Available: http://www.campaignforcure.org (2010, Mar 1).
Catz, A., & Itzkovich, M. (2007). Spinal Cord Independence Measure: Comprehensive
ability rating scale for the spinal cord lesion patient. Journal of Rehabilitation and
Research Development, 44(1), 65-68.
Catz, A., Greenberg, E., Itzkovich, M., Bluvshtein, V., Ronen, J., & Gelernter, I. (2004).
A new instrument for outcome assessment in rehabilitation medicine: Spinal cord
injury ability realization measurement index. Archives of Physical Medicine and
Rehabiliation, 85(3), 399-404.
Cohen, L.G., Topka, H., Cole, R.A., Hallett M. (1991). Leg paresthesias induced by magnetic
brain stimulation in patients with thoracic spinal cord injury. Neurology, 41(8),
1283-1288.
Corbetta, M., Burton, H., Sinclair, R.J., Conturo, T.E., Akbudak, E., & McDonald, J.W.
(2002). Function reorganization and stability of somatosensory-motor cortical
112
topography in a tetraplegic subject with late recovery. Proceedings of the National
Academy of Sciences USA, 99(26), 17066-17071.
Curt, A., & Dietz, V. (1996). Traumatic Cervical Spinal Cord Injury: Relation Between
Somatosensory Evoked Potentials, Neurological Deficit, and Hand Function.
Archives of Physical Medicine and Rehabilitation, 77(1), 48-53.
Curtin, C.M., Gater, D.R., & Chung, K.C. (2005). Upper Extremity Reconstruction in the
Tetraplegic Population, a National Epidemiologic Study. Journal of Hand Surgery,
30(A), 94-99.
Cuthbert, S.C., & Goodheart, G.J. (2007). On the reliability and validity on manual muscle
testing: A literature review. Chiropractor Osteopath, 15(4).
Dahlberg, A., Kotila, M., Kautianen, H., & Alaranta, H. (2003). Functional independence in
persons with spinal cord injury in Helsinki. Journal of Rehabilitation Medicine,
35(5), 217-220.
Daniels, L., & Worthingham, C. (1995). Daniels and Worthingham’s Muscle Testing:
Techniques of Manual Examination, Sixth Edition. Washington: WB Saunders Co.
DeVivo, M.J. (1997). Causes and costs of spinal cord injury in the United States.
Spinal Cord, 35(12), 809-813.
Ditunno, J.F., Cohen, M.E., Hauck, W.W., Jackson, A.B., & Sipski, M.L. (2000). Recovery
of Upper-Extremity Strength in Complete and Incomplete Tetraplegia: A Multicenter
Study. Archives of Physical Medicine and Rehabilitation, 81(4), 389-393.
Ditunno, J.F., Sipski, M.L., Posuniak, E.A., Chen, Y.T., Staas, W.E., & Herbison,.G.J.
(1987). Wrist Extensor Recovery in Traumatic Quadriplegia. Archives of Physical
Medicine and Rehabilitation, 68(5 Pt 1), 287-290.
Drolet, M., Noreau, L., Vachon, J., & Moffet, H. (1999). Muscle Strength Changes as
113
Measured by Dynamometry Following Functional Rehabilitation in Individuals With
Spinal Cord Injury. Archives of Physical Medicine and Rehabilitation, 80(7),
791-800.
Donner, A., & Eliasziw, M. (1987). Sample Size Requirements for Reliability Studies.
Statistics in Medicine, 5, 441-448.
Dunn, J.A., Sinnott, K.A., Bryden, A.M., Connolly, S.J., & Rothwell, A.G. (2008).
Measurement issues related to upper limb interventions in persons who have
tetraplegia. Hand Clinics, 24(2), 161-168.
Feinstein, A.R. (1990). Clinimetrics, Second Edition. The Theory and Evaluation of
Sensibility. Conneticut: Yale University Press, 141-165.
Filimon, F. (2010). Human cortical control of hand movements: parietofrontal networks for
reaching, grasping and pointing. Neuroscientist, 16(4), 388-407.
Fisher, C.G., Noonan, V.K., Smith, D.E., Wing, P.C., Dvorak, M.F., & Kwon BK. (2005).
Motor recovery, functional status, and health-related quality of life in patients
with complete spinal cord injuries. Spine, 30(19), 2200-2207.
Fugl-Meyer, A.R., Jääskö, L., Leyman, I., Olsson, S., & Steglind, S. (1975). The post
stroke hemiplegic patient. 1. a method for evaluation of physical performance.
Scandinavian Journal of Rehabilitation Medicine, 7(1), 13-31.
Fujiwara, T., Hara, Y., Akaboshi, K., & Chino, N. (1999). Relationship between shoulder
muscle strength and functional independence measure (FIM), score among C6
tetraplegics. Spinal Cord, 37(1), 58-61.
Green, J.B., Sora, E., Bialy, Y., Ricamato, A., Thatcher, R.W. (1999). Cortical motor
reorganization after paraplegia: an EEG study. Neurology, 53(4), 736-743.
Green, J.B., Sora, E., Bialy, Y., Ricamato, A., & Thatcher, R.W. (1998). Cortical
114
Sensorimotor Reorganization after spinal cord injury: an electroencephalographic
study. Neurology, 50(4), 1115-1121.
Guttman, L.A. (1950). The basis of scalogram analysis. In Stouffer, S.A., Guttman, L.A.,
Schuman, E.A. In measurement and Prediction. Volume 4 of Studies in Social
Psychology in World War II. Princeton: Princeton University Press.
Guttman, L.A., (1944). A basis for scaling qualitative data. American Sociological Review,
91, 139-150.
Hoffman, L.R., & Field-Fote, E.C. (2007). Cortical Reorganization Following Bimanual
Training and Somatosensory Stimulation in Cervical Spinal Cord Injury: A Case
Report. Physical Therapy, 87(2), 208-23.
Hooper, D., Coughlan, J., & Mullen, M.R. (2008). Structural Equation Modelling: Guidelines
for Determining Model Fit. Electronic Journal of Business Research Methods, 6(1),
53-60.
Itzkovich, M., Gelernter, I., Biering-Sorensen, F., Weeks, C., Laramee, M.T.,
Craven, B.C., Tonack, M., Hitzig, S.L., Glaser, E., Zeilig, G., Aito, S., Scivoletto, G.,
Mecci, M., Chadwick, R.J., El Masry, W.S., Osman, A., Glass, C.A., Silva, P., Soni,
BM., Gardner, B.P., Savic, G., Bergström, E.M., Bluvshtein, V., Ronen, J., & Catz,
A. (2007). The Spinal Cord Independence Measure (SCIM) version III: reliability and
validity in a multi-center international study. Disability Rehabilitation, 29(24), 926-
933.
Jebsen, R.H., Taylor, N., Trieshmann, R.B., Trotter, M.J., & Howard, L.A. (1969). An
objective and standardized test of hand function. Archives of Physical Medicine and
Rehabilitation, 50(6), 311-319.
Jurkiewicz, M.T., Mikulis, D.J., McIlroy, W.E., Fehlings, M.G., & Verrier, M.C. (2007).
115
Sensorimotor cortical plasticity following spinal cord injury: a longitudinal fMRI
Study. Neurorehabilitation and Neural Repair, 21(6), 527-538.
Kandel, E.R., Schwartz, J.H., & Jessell, T. M. (2000). Principles of Neural Science,
Fourth Ed. New York: Elsevier Science Publishing Co.
Kalsi-Ryan, S., and Verrier, M.C. (2010). A Synthesis of Best Evidence for the Restoration
Of Upper Extremity Function in People with Tetraplegia. Physiotherapy Canada,
In Press
Kalsi-Ryan, S., Curt, A., Fehlings, M.G., & Verrier, M.C. (2009). Assessment of the Hand in
Tetraplegia Using the Graded Redefined Assessment of Strength Sensibility and
Prehension (GRASSP): Impairment versus Function. Topics in Spinal Cord Injury
Rehabilitation, 14(4), 34-46.
Kalsi-Ryan, S. (2006). The Quadriplegia Hand Assessment Tool (Q-HAT). The
Development of a Clinical Assessment Measure of the Hand for the Cervical Spinal
Cord Injured Individual. Graduate Department of Rehabilitation Science, University
of Toronto, Toronto, Canada.
Kapandji, I.A. (1970). The Physiology of the Joints, Volume 1, Second Ed.
Edinburgh, Churchill Livingstone.
Keith, R.A., Granger, C.V., Hamilton, B.B., & Sherwin, F.S. (1987). The functional
independence measure: a new tool for rehabilitation. Advanced Clinical
Rehabilitation, 1, 6-18.
Kendall, F., McCreary, E.K., & Provance, P.G. (1993). Muscles: Testing and Function,
Fourth Edition. Philadelphia: Williams & Wilkins.
Kidd, D., Stewart, G., Baldry, J., Johnson, J., Rossiter, D., Petruckevitch., A.M., &
Thompson, A.J. (1995). The Functional Independence Measure: a comparative
116
validity and reliability study. Disability Rehabilitation, 17, 10-14.
Kirshblum, S.C., & O’Connor, K.C. (2000). Levels of spinal cord injury and predictors of
neurologic recovery. Physical Medicine and Rehabilitation Clinic North
America, 11(1), 1-27
Kohlmeyer, K.M., Hill, J.P., Yarkony, G.M., & Jaegar, R.J. (1996). Electrical Stimulation
and Biofeedback Effect on Recovery of Tenodesis Grasp: A Controlled Study.
Archives of Physical Medicine and Rehabilitation, 77(7), 702-706.
Laffont, I., Hoffmann, G., Dizien, O., Revol, M., Roby-Brami, A. (2007). How do C6/C7
tetraplegic patients grasp balls of different sizes and weights? Impact of surgical
musculo-tendinous transfers. Spinal Cord, 45, 502-12.
Lamb, D.W., & Chan, K.M. (1983). Surgical Reconstruction of the Upper Limb in Traumatic
Tetaplegia. The Journal of Bone and Joint Surgery, 65B(3), 291-298.
Landis, J.R., & Koch, G.G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 33, 159-174.
Levy, W.J. Jr., Amassian, V.E., Traad, M., & Cadwell, J. (1990). Focal magnetic
coil stimulation reveals motor cortical system reorganized in humans after traumatic
quadriplegia. Brain Research, 510(1), 130-134.
Lin, J.H., Hsu, M.J., Sheu, C.F., Wu, T.S., Lin, R.T., Chen, C.H., & Hsieh, C.L. (2009).
Psychometric comparisons of 4 measures for assessing upper-extremity function in
People with stroke. Physical Therapy, 89(8), 840-850.
Link C. The Link Hand Function Test for Patients with a Cervical Spinal Cord Injury: An
Intra-Rater and Inter-Rater Reliability and Expert Opinion Evaluation Study (2004).
Hogeschool van Amsterdam, Institute of Occupational Therapy, The Netherlands;
University College-South, School of Occupational Therapy and Physiotherapy,
117
Denmark; Karolinska Institutet Division of Occupational Therapy, Sweden.
Lo, I.K.Y., Turner, R., Connolly, S., Delaney, G., & Roth, J.H. (1998). The Outcome of
Tendon Transfers for C6-Spared Quadriplegics. Journal of Hand Surgery, 23B(2),
156-161.
Lundborg, G., & Rosen, B. (2007). Hand Function After Nerve Repair. Acta. Physiol.,
189: 207-217.
Lundborg, G. (2000). Brain Plasticity and Hand Surgery: An Overview. Journal of Hand
Surgery, 25B(3), 242-252.
Mackin, E., Callahan, A., Skiver, T., Schneider, L., & Osterman A. (2003). Hunter-Mackin
Callahan Rehabilitation of the Hand and Upper Extremity, Fifth Ed. St. Louis,
Missouri: Mosby.
Marciello, M.A., Herbison, G.J., Ditunno, J.F., Marino, R.J., Cohen, M.E. (1995). Wrist
Strength Measured by Myometry as an Indicator of Functional Independence.
Journal of Neurotrauma, 12(1), 99-106.
Marino, R.J., Barros, T., Biering-Sorensen, F., Burns, S.P., Donovan, W.H., Graves,
D.H., Haak, M., Hudson, L.M., Priebe, M.M., & ASIA Neurological Standards
Committee 2002. (2003). Journal of Spinal Cord Medicine, 26, S50-S56.
Marino, R.J. (2000). American Spinal Injury Association: International Standards of
Neurological Classification of the Spinal Cord Injury; Atlanta, Georgia.
Reprinted 2008.
Marino, R.J., Ditunno, J.F., Donovan, W.H. & Maynard, F. (1999). Neurological recovery
after traumatic spinal cord injury: data from the model spinal cord injury systems.
Archives of Physical Medicine and Rehabilitation, 80(11),1391-1396.
Marino, R.J., & Goin, J.E. (1999). Development of a short-form Quadriplegia Index of
118
Function scale. Spinal Cord, 37(4): 289-296.
Marino, R.J., Shea, J.A., & Stineman, M.G. (1998). The Capabilities of the Upper Extremity
instruement reliability and validity of a measure of functional limitation in
tetraplegia. Archives of Physical Medicine and Rehabilitation, 79(12), 1512-1521.
Marino, R.J., Huang, M., Knight, P., Herbison, G.J., Ditunno, J.F., & Segal, M. (1993).
Assessing selfcare status in quadriplegia: comparison of the quadriplegia index of
function (QIF) and the functional independence measure (FIM). Paraplegia, 31(4),
225-233.
Mathiowetz, V., Weber, K., Volland, G., Kashman, N., Dowe, M., & Rogers, S. (1984).
Reliability and validity of grip and pinch strength evaluations. Journal of Hand
Surgery, 9(2), 69-74.
McDowell, C.L., Moberg, E.A., & House, J.H. (1986). The Second International Conference
of Surgical Rehabilitation of the Upper Limb in Tetraplegia. Journal of Hand
Surgery, 11A(4), 604-608.
McDowell, I., & Jenkinson, C. (1996). Development of standards for health measures.
Journal of Health Service and Research Policy, 1(4), 238-246.
McHorney, C.A. (1999). Health Status Assessment Methods for Adults: Past
Accomplishements and Future Challenges. Annual Review of Public Health, 20,
309-335.
Meiners, T., Abel, R., Lindel, K., & Mesecke, U. (2002). Improvements in activities of daily
living following functional hand surgery for treatment of lesions to the cervical spinal
cord: Self-assessment by patients. Spinal Cord, 40, 574-80.
Miller, R.C., Popovic, M.R., Thrasher, A., & Verrier, M.C. (2008). Functional Electrical
Stimulation Therapy Improves Grasping in Chronic Cervical Spinal Cord Injury: Two
119
Case Studies. Journal of Automatic Control, 80(2), 53-62.
Miller, W.C., Aubut, J.L., Curt, A., Connolly, S.J., Elliott, S., Hsieh, J.T.C., Mortenson,
W.B., Noonan, V.K., Noreau, L., Orenczuk, S.G., Sakakibara, B.M., Sawatzky, B.,
Steeves, J., Wilkinson, S., & Wolfe, D. (2008). SCIRE: Spinal Cord Injury
Rehabilitation Evidence Version 2.0. Chapter 25, Outcome Measures.
Mizukami, M., Kawai, N., Iwasaki, Y., Yamamoto, Y., Yoshida, Y., Koyama, N., Sekiguchi,
S., Kimura, T., & Nihei, R. (1995). Relationship between functional levels and
movement in tetraplegic patients. A retrospective study. Paraplegia, 33(4), 189-194.
Moore, K.L. (1985). Clinically Oriented Anatomy Second Ed. Baltimore, Williams and
Wilkins, 728-744.
Mulcahey, M.J., Smith, B.T., and Betz, R.R. (2004). Psychometric rigor of the Grasp and
Release Test for measuring functional limitation of persons with tetraplegia: a
preliminary analysis. Journal of Spinal Cord Medicine, 27(1), 41-46.
Müslümanoğlu, L., Aki, S., Oztürk, Y., Soy, D., Filiz, M., Karan, A., & Berker, E. (1997).
Motor, sensory and functional recovery in patients with spinal cord lesions.
Spinal Cord, 35(6), 386-389.
Netter, F.H. (1989). Atlas of Human Anatomy, First Edition. Summit, New Jersey,
Acme Printing Company Inc.
Norquist, J.M., Fitzpatrick, R., Dawson, J., & Jenkinson, C. (2004). Comparing Alternative
Rasch-Based Methods vs Raw Scores in Measuring Change in Health. Medical Care,
42(1), 1 -25.
Nunnally, J.C., & Bernstein, I.H. (1996). In Psychometric Theory. Toronto,
McGraw Hill Inc.
Ota, T., Akaboshi, K., Nagata, M., Sonoda, S., Domen, K., Seki, M., & Chino, N. (1996).
120
Functional assessment of patients with spinal cord injury: measured by the motor
score and the Functional Independence Measure. Spinal Cord, 34(8), 531-535.
Patrick, D.L., & Erickson, P. (1993). Health Status and Health Policy Allocation Resource
to Health Care. New York: Oxford University Press, 198-202.
Peckam, H., Keith, M.W., Kilgore, K.L., Grill, J.H., Wuolle, K.S., Thorpe, G.B. (2001).
Efficacy of an Implanted Neuroprothesis for Restoring Hand Grasp in Tetraplegia: A
Multicenter Study. Archives of Physical Medicine and Rehabilitation, 82, 1380-1388.
Popovic, D.P., Stojanovic, A., Pjanovic, A., Radosavljevic, S., Popovic, M., Jovic, S.,
Vulovic, D. (1999). Clinical Evaluation of the Bionic Glove. Archives of Physical
Medicine and Rehabilitation, 80, 299-304.
Popovic, M.R., Thrasher, T.A., Adams, M.A., Takes, V., Zivanovic, V., & Tonack, M.I.
(2006). Functional electrical therapy: retraining grasping in spinal cord injury.
Spinal Cord, 44(3), 143-151
Portney, L.G., Watkin, M.P. (2000). Foundations of Clinical Research: Applications to
Practice, Second Edition. Chapter 26: Statistical Measures of Reliability.
Norwich, Conneticut: Appleton and Lange, 557-586.
Post, M.W., Van Lieshout, G., Seelen, H.A., Snoek, G.J., Ijzerman, M.J., & Pons, C. (2006).
Measurement properties of the short version of the Van Lieshout test for arm/hand
function of persons with tetraplegia after spinal cord injury. Spinal Cord, 44, 763-771.
121
Poynton, A.R., O’Farrell, D.A., Shannon, F., Murray, P., McManus, F., & Walsh, G. (1997).
Sparing of Sensation to Pinprick Predicts Recovery of a Motor Segment after Injury
to the Spinal Cord. Journal of Bone & Joint Surgery, 79-B, 952-954.
Priebe, M., & Waring, W.P. (1991). The interobserver reliability of the revised American
Spinal Injury Association standards for neurological classification of spinal injury
patients. American Journal of Physical Medicine and Rehabilitation, 70(5), 268-270.
Prochazka, A., Gauthier, M., Wieler, M., & Kenwell, Z. (1997). The bionic glove: an
electrical stimulator garment that provides controlled grasp and hand opening in
quadriplegia. Archives of Physical Medicine and Rehabilitation, 78(6), 608-614.
Raineteau, O., & Schwab, M.E. (2001). Plasticity of motor systems after incomplete spinal
cord injury. National Review of Neuroscience, 2(4), 263-73.
Rick Hansen Registry. (2005). Spinal Cord Injury Facts and Statistics, [pdf file].
National Trauma Registry. Available: http://rickhansenregistry.org (2010, Mar 1).
Rosen, B. & Lundborg, G. (2003). A new model instrument for outcome after nerve repair.
Hand Clinic,19(3), 463-470.
Rothwell, A.G., Sinnott, K.A., Mohammed, K.D., Dunn, J.A., & Sinclair, S.W. (2003).
Upper Limb Surgery for Tetraplegia: A 10-Year Re-Review of Hand Function.
Journal of Hand Surgery, 28A, 489-497.
Rowe, B.H., & Oxman, A.D. (1993). An Assessment of the Sensibility of a Quality of Life
Instruement. American Journal of Emergency Medicine, 11, 374-380.
Rudhe, C., & van Hedel, H. (2009). Upper Extremity Function in Persons with Tetraplegia:
122
Relationships Between Strength, Capacity and the Spinal Cord Independence
Measure. Neurorehabilitation and Neural Repair, 23(5), 413-21.
Samejima, F. (1969). Estimation of latent ability using a response pattern of graded scores.
Psychometrika Monograph, No. 17.
Schonherr, M.C., Groothoff, J.W., Mulder, G.A., & Eisma, W.H. (1999). Functional outcome
of patients with spinal cord injury: rehabilitation outcome study. Clinical
Rehabilitation, 13(6), 457-463.
Schwab, J.M., Brechtel, K., Mueller, C.A., Failli, V., Kaps, H.P., Tuli, S.K., & Schulesener,
H.J. (2006). Experimental strategies to promote spinal cord regeneration—an
integrative perspective. Progressive Neurobiology, 78(2), 91-116.
Sear, D.E., and Chung, K.C. (2010). Validity and responsiveness of the Jebsen-Taylor Hand
Function Test. Journal of Hand Surgery of America, 35, 30-37.
Sekhon, L., & Fehlings, M.G. (2001). Epidemiology, Demographics, and Pathophysiology of
Acute Spinal Cord Injury. Spine, 26(24S), S2-S12.
Shumway-Cook, A., Woollacott, M.H. (2007). Motor Control: translating research into
Clinical Practice, Third Edtion. St. Louis: Lippincott, Williams & Wilkins.
Sinnott, K.A., Dunn, J.A., & Rothwell, A.G. (2004). Use of the ICF conceptual framework to
interpret hand function outcomes following tendon transfer surgery for tetraplegia.
Spinal Cord, 42(7), 396-400.
Smith, B.T., Mulcahey, M.J. & Betz, R.R. (1996). Quantitative comparison of grasp
and release abilities with and without functional neuromuscular stimulation in
adolescents with tetraplegia. Paraplegia, 34, 16-23.
Snoek, G.J., Ijzerman, M.J., Post MW, Stiggelbout AM, Roach MJ, Zivold G. Choice-Based
123
Evaluation for the Improvement of Upper-Extremity Function Compared With Other
Impairments in Tetraplegia. (2005). Archives of Physical Medicine and
Rehabilitation, 86(8), 1623-1630.
Snoek, G.J. (2004). Survey of the needs of patients with spinal cord injury: impact and
priority for improvement in hand function in tetraplegics. Spinal Cord, 42(9),
526-532.
Sollerman, C., & Ejeskar, A. (1995). Sollerman hand Function Test: A standardized method
and its use in tetraplegic patients. Scandinavian Journal of Plastic and Reconstructive
Hand Surgery, 29, 167 – 176.
Spooren, A.I., Janssen-Potten, Y.J., Post, M.W., Kerckhofs, E., Nene, A., & Seelen, H.A.
(2006). Measuring change in arm hand skilled performance in persons with a cervical
spinal cord injury: responsiveness of the Van Lieshout Test. Spinal Cord, 44(12),
772-779.
Steeves, J.D., Lammertse, D., Curt, A., Fawcett, J.W., Tuszynski, M.H., Ditunno, J.F.,
Ellaway, P.H., Fehlings, M.G., Guest, J.D., Kleitman, N., Bartlett, P.F., Blight, A.R.,
Dietz, V., Dobkin, B.H., Grossman, R., Short, D., Nakamura, M., Coleman, W.P.,
Gaviria, M., & Privat, A. (2007). Guidelines for the conduct of clinical trials for
spinal cord injury (SCI) as developed by the ICCP panel: clinical trial outcome
measures. Spinal Cord, 45(30), 206-21.
Stevens, S.S. (1946). On the Theory of Scales of Measurement. American Association for the
Advancement of Science, 103(2684), 677-680.
Streiner, D.L., & Norman, G.L. (1995). Health Measurement Scales: A Practical Guide to
Their Development and Use, Second Edition. Reliability, Chapter 8. New York:
Oxford University Press Inc., 104-127.
124
Streiner, D.L., & Norman, G.L. (1995). Health Measurement Scales: A Practical Guide to
Their Development and Use, Second Edition. Validity, Chapter 10. New York:
Oxford University Press Inc., 144-162.
Sudano, J. (2010). Structural Equation Modeling (SEM) for Dummies, [pdf file]
Available: http://www.chrp.org/pdf/HSR050903.pdf; (2003, February 1).
Taylor, P., Esnouf, J., & Hobby, J. (2002). The functional impact of the Freehand System on
Tetraplegic hand function. Clinical results. Spinal Cord, 40(11), 560-566.
Tomarken, A.J., & Waller, N.G. (2005). Structural Equation Modeling: Strengths, Limitation
And Misconceptions. Annual.Review of Clinical Psychology, 1, 31-65.
Tooth, L., McKenna, K., & Geraghty, T. (2003). Rehabilitation outcomes in traumatic spinal
cord injury in Australia: functional status, length of stay and discharge setting.
Spinal Cord, 41(4), 220-230.
Trochim, W.M. (2010). The Research Methods Knowledge Base, Second Edition.
Internet WWW page at URL: <http://www.socialresearchmethods.net/kb/>
(version current as of October 20, 2006).
Vastamaki, M. (2006). Short-Term Versus Long-Term Comparative Results After
Reconstructive Upper-Limb Surgery in Tetraplegic Patients. Journal of Hand
Surgery, 31(A), 1490-1494.
van Abswoude, A.A.H., Verment, J.K., Hemke, B.T., & van der Ark, L.A. (2004). Mokken
Scale Analysis Using Hierarchical Clustering Procedures. Applied Psychological
Measurement, 28(5), 332-354.
125
van den Berg, M.E.L., Castellote, J.M., Mahillo-Fernandez, I., & de Pedro-Cuesta, J.
(2010). Incidence of Spinal Cord Injury Worldwide: A Systematic Review.
Neuroepidemiology, 34, 184-192.
van Schuur, W.H. (2003). Mokken Scale Analysis: Between the Guttman Scale and
Parametric Item Response Theory. Political Analysis, 11, 139-163.
van Tuijl, J.H., Janssen-Potten, Y.J.M., & Seelen, H.A.M. (2002). Evaluation of upper
extremity motor function tests in tetraplegics. Spinal Cord, 40(2), 51-64.
Welch, R.D., Lobley, S.J., O’Sullivan, S.B., & Freed, M.M. (1986). Functional independence
in quadriplegia: critical levels. Archives of Physical Medicine and Rehabilitation,
67(4), 235-240.
Wells, J.E., Hurlbert, R.J., Fehlings, M.G., & Yong, V.W. (2003). Neuroprotection by
minocycline facilitates significant recovery from spinal cord injury in mice.
Brain, 126(Pt7), 1628-1637.
Wilson, I.B., & Cleary, P.D. (1995). Linking Clinical Variables With Health-Related Quality
of Life. Journal of the American Medical Association, 273(1), 59-65.
Wright, J.G., & Feinstein, A.R. (1992). Improving the Reliability of Orthopaedic
Measurements. Journal of Bone and Joint Surgery (Br), 74-B, :287-291.
Wyndaele, M., and J.J. Wyndaele. (2006). Incidence, prevalence and epidemiology of spinal
cord injury: what learns a worldwide literature survey? (2006). Spinal Cord, 44(9),
523-529.
126
Appendix 1
A: Initial GRASSP
B: GRASSP Version 1.0
127
A: Initial GRASSP
GRADED and REDEFINED ASSESSMENT
Of STRENGTH, SENSIBILITY and PREHENSION
(Initial GRASSP)
Collaboration: Armin Curt, Susan Duff, Michael Fehlings Sukhvinder Kalsi-Ryan, Claudia Rudhe-Link, Molly Verrier, Lisa Ann Weurmser
128
TABLE OF CONTENTS
Introduction 2
Purpose of the Measure 2
Design and Development of the GRASSP 2
Modules 2
Timelines for Testing 3
Instructions for implementation of the GRASSP 3
Upper Extremity ASIA 3
Strength 4
Muscle Tone 7
Sensibility 7
Prehension 9
Scoring Sheets 13
Demographics 13
Upper Extremity ASIA 13
Strength and Muscle Tone 14
Sensibility 15
Prehension 16
Summary and Total Scores 16
References 17
129
INTRODUCTION Initiated by the North American Clinical Trials Network (NACTN) and the European Clinical Trials Network (EUCTN), a meeting was held on May 12 and 13, 2006 in Chicago (local organizer Drs. Zev Rymer and Lisa-Ann Wuermser, financially supported by the Christopher Reeves Foundation (CRF)) to discuss the measurement of hand impairment and function in patients suffering from cervical spinal cord injury (cSCI). Members of the networks and independent clinical specialists in hand measurement and therapy, as well as researchers with expertise in engineering and computer technology discussed the development of a comprehensive protocol to assess upper limb impairment and recovery post cSCI. The result of the meeting was a task force to further develop a clinical assessment protocol of hand function by modifying existing tools and introducing new measures that would allow for the quantification of change in hand function for individuals with cSCI. The GRASSP is a combined effort by six clinicians/researchers who have contributed their work (results of tool development through previous research, in some instances graduate studies), skills, and time. The GRASSP is a mosaic of the Link Hand Function Test (LiHFT) (Claudia Rudhe-Link, Zuerich, Switzerland), Testing of Strength & Muscle Tone (Susan Duff, Philadelphia, US), and The Tetraplegia Hand Measure (THM) (Sukhvinder Kalsi-Ryan, Toronto, Canada). PURPOSE OF THE MEASURE The overall objective for the assembly of the GRASSP was to develop a clinical research tool that could capture information on hand function from the cervical (C0-T1) spinal cord injury (SCI) population, obtain integrated sensory and motor impairment data, and discriminate the population according to the level of lesion. The purpose of this project was to design a hand function tool: 1) that was highly responsive (sensitive) to change over time; 2) that could assess the extent of spontaneous (natural) recovery; and 3) be applicable for use in clinical trials to evaluate the effect of novel interventions (pharmacological and surgical). DESIGN AND DEVELOPMENT OF THE GRASSP The GRASSP is a framework that assembles different clinical tools to measure the various aspects of complex sensori-motor hand function. The GRASSP is embedded with currently existing measures of upper limb function (ASIA motor and sensory scores of the upper limb) (Marino et al. 2002). It also includes clinical tools that should provide information useful for comparison to other upper limb testing protocols. For the development of the GRASSP each of the three modules was assigned to one of the measurement developers (Strength/Muscle Tone – Susan Duff, Sensibility- Sukhvinder Kalsi-Ryan, and Prehension – Claudia Rudhe-Link) under the direction of Armin Curt, Molly Verrier, and Lisa Ann Wuermser. Although, individuals were responsible for separate modules all members of the task force made significant contributions to all components of the GRASSP. MODULES The GRASSP is comprised of three separate modules, Strength/Muscle Tone, Sensibility, and Prehension. Multiple modules allow for a comprehensive assessment
130
at multiple time points in the post-injury continuum. Each module can be tested according to the scheduled timeline provided by a trial protocol. TIMELINES OF TESTING The schedule provided below exemplifies how the tests could be sequenced in a given trial (Table 1 shows a potential scheduling of examinations). It’s possible that the very early period following SCI might be monitored by a subset or even different measures than in the advanced and chronic stages of SCI. Table 1: Time of Examination
Modules Impairment Impairment & Capacity Sub-acute – weeks Time of Examination Acute 72hrs-
10 days 4 8 12 24 52ASIA muscles X X X X X X Proximal Arm X* X X X X X Extrinsic Hand Muscles X X X X X Intrinsic Hand Muscles X X X X X
Strength & Muscle Tone
Muscle Tone/Modified Ashworth X X X X X ASIA sensory –LT/PP X X X X X X Semmes Weinstein Monofilaments X X X X X
Sensibility
Static Two Point Discrimination X X X X X Qualitative Grasp X X X X X X Prehension Quantitative Grasp X X X X X
5. Only to be tested if not contradicted by physician Modules – The modules complement the ASIA and functional (SCIM/FIM) testing. ASIA – ASIA items are shown in italic
INSTRUCTIONS FOR IMPLEMENTATION OF THE GRASSP: Consent: Always obtain consent from the subject (patient) and collect the necessary demographic data based on interview with the patient/family and chart review. Positioning the Patient: In acute period the patient is lying supine with both arms exposed to the shoulders and should be tested in this position. During most other testing sessions the subject should be seated in his/her own seating system with his/her appropriate supports. During all testing, the entire upper extremity should be exposed (up to the shoulder). An adjustable table which can move in and out of wheelchair space will be required to perform the assessment. The subject’s hands should be positioned on the table, with approximately 30 degrees of shoulder flexion, 65 to 70 degrees of elbow flexion and the hands and distal half of the forearms supported on the table. This position can be modified slightly to ensure comfort for the individual being tested. The room where the testing will be done should be well lit. Length of the Testing: The time required to complete all of the tests in one session is approximately 45 – 60 minutes (depending on patient ability). It is not recommended to break the testing up into two sessions over two days as individual’s response can vary and recovery can potentially affect the results. For the best outcome it is recommended to complete the testing in one session, however, between tests the individual and the examiner can break and stretch.
131
UPPER EXTREMITY ASIA: Muscle Testing: All of the ASIA muscles are tested within the required GRASSP muscles, therefore, refer to the instructions in the Strength and Tone section and use the scale also. Light Touch and Pinprick: Prepare the patient as above. A safety pin and cotton swab should be applied to test locations as per ASIA standards (Marino et al. 2002). Dermatomes C2 to T1 should be tested for the purpose of this work. The specific locations (dorsal aspect of the hand at position 1, 2 and 3) are amplified in Figure 1. The response is then recorded as intact, impaired or absent as described in Table 2. Pinprick testing is confirmed by performing dull/sharp testing and determining if sharp is intact. If there is uncertainty then dull/sharp should be repeated up to 10 times for one location to be certain of the sensation. Figure 1: ASIA Sensory Test Locations, labeled by dermatome
STRENGTH and MUSCLE TONE
Muscles specific to the upper limb and hand were added to the ASIA repertoire of testing to establish greater sensitivity to potential change post-injury. Strength will be assessed with Manual Muscle Testing (MMT). An isotonic muscle contraction will be required by the subject to grade muscle strength. Specifically, resistance should be given at the distal end of the moving bone while the subject moves the limb through the specific range (Daniels and Worthingham, 1995). Muscle tone will be assessed using the Modified Ashworth Scale (Bohannon and Smith, 1987). The following tables define the scaling for the two tests, the muscles to be tested and the instructions for testing each muscle. Muscle Testing Prior to beginning muscle testing the subject should be oriented to the test by demonstration on an active body part. If the testing will be done in supine the examiner should stand comfortably at the bedside. If the subject is seated, the examiner may choose to stand next to the wheelchair or sit next to/across from them.
Response Normal Feels stimulus but not the same as normal location
No response
Response Label
Normal Impaired Absent
Score 2 1 0
132
During assessment of the distal arm musculature the subjects’ distal forearm should rest on an adjustable table. Begin by testing the muscle for a grade three (range against gravity), ensuring the joints are isolated. If the individual is able to move through full range of motion (ROM) against gravity then the same movement should be tested with resistance for a grade 4 or 5 through full ROM. The examiner will grade the individual according to the scoring key. Resistance is given at the distal end of moving the bone during an isotonic contraction. Table 2 defines the muscles to be tested and the stabilization and resistance required for testing these muscles and the scoring key to be used. Remember that for finger muscles gravity does not have an effect, which defines grade 2 as: movement of the corresponding body part but not through the full range of motion and grade 3 as: movement through full ROM. All MMT scoring should be recorded on the score sheets in the next section. Note: 1) Range of motion for anterior deltoid is considered to be to 90 degrees of elevation and MMT scoring should be based on 90 degrees of elevation as full ROM. 2) For elbow extension if the starting position (full elevation) is not feasible then elbow extension can be tested in 90 degrees of shoulder elevation. Table 3: Strength Testing and Instructions Muscle Action Nerve
Radicular segments
Action Stabilization Starting position
Resistance Point
Anterior Deltoid
Arm elevation for reaching & pointing
Axillary n C5-6
Shoulder abduction – 90°/ flexion in supine
Trunk 00 shoulder adduction/ flexion
Anterior, distal humerus
Elbow extensors
Reaching Radial n C6-C8
Extend elbow
Humerus Elbow flexion, shoulder abduction (hand behind the head, or 90° abd and full inward rotation of the humerus
Distal forearm
Elbow flexors
Hand to face/mouth, bring objects toward trunk
Musculo- Cutaneous C5-6
Flex elbow Humerus Full Elbow extension, shoulder adduction, forearm in supination
Distal, volar forearm
Wrist extensors
Stabilize for prehension, Orienting hand to grasp
Radial n C6-C8
Extend wrist
Forearm Wrist in flexion, forearm in pronation
Distal, dorsal 3rd metacarpal
Extensor Digitorum
Hook grasp, Object release
Radial n C6-C8
Extend MP’s digits 2-5
2-5 Metacarpals
Flexion IP’s / MP’s digits 2-5, forearm in pronation (fingers hanging over
Dorsal Proximal phalanges digits 2-5
133
edge of table)
Finger flexors (3rd FDP)
Pad to pad (3-jaw chuck), Cylindrical Grip, Hook Grasp
Median n C7-T1
Flex DIP joint of 3rd digit
MP, PIP joint 3rd digit
3rd digit DIP extension, 3rd PIP/MP supported in 00 extension on table
3rd volar finger pad
Flexor Pollicis Longus
Spherical Grasp, Lateral Pinch, Tip to Tip Pinch
Median n C6-C8
Flex thumb IP joint
Thumb Proximal phalanx/ metacarpal
00 thumb IP extension, MP supported in 00 extension
Volar thumb pad
Finger abductors (5th)
Spherical Grasp, Cylindrical grasp
Ulnar n C8-T1
Abduct 5th digit
5th metacarpal
5th digit adducted, MP’s extended, forearm in pronation
Distal, ulnar, 5th proximal phalanx
First Dorsal Interossei
Spherical grip, Lateral Pinch
Ulnar n C8-T1
Abduct index
2nd metacarpal
Index adducted next to long finger, MP’s extended, forearm in pronation
Distal, radial, 2nd proximal phalanx
Adductor Pollicis Brevis
Lateral Pinch Ulnar n C8-T1
Adduct toward index
Wrist/ forearm
Thumb abduction, wrist and forearm in supination
Distal, ulnar, thumb proximal phalanx
Opponens Pollicis
Tip to Tip, Pad to Pad, Spherical grasp, Cylindrical grasp
Median n C6-C7
Rotate 1st metacarpal toward 5th digit pad
wrist/ 2-5 metacarpals
Thumb in a resting posture against 5th metacarpal, wrist and forearm in supination
Distal, volar proximal phalanx with derotating pressure
ASIA Muscles in italics (Daniels and Worthingham, 1995; Kendall, McCreary, and Provance, 1993) Table 4: Scoring Key for Manual Muscle Testing
0 Absent – No palpable muscle contraction 1 Trace – Palpable Muscle Contraction 2 Poor – Moves full ROM with gravity eliminated 3 Fair – Moves full ROM against gravity without added resistance 4 Good – Moves through full ROM against gravity against moderate resistance 5 Normal – Moves through full ROM against gravity against maximal resistance
134
MUSCLE TONE Muscle tone will be assessed during passive range of motion (PROM) of the limbs and evaluated using the Modified Ashworth Scale as per the following instructions (Bohannon and Table 5: Muscle Tone Testing, Instructions and Scoring Key Muscle Action Test Regime Elbow Flexors Reaching to mouth Passively extend the elbow Finger Flexors Grasp Passively extend the fingers
Scoring for Muscle Tone 5 No increase in muscle tone 4 Slight increase in muscle tone, manifested by a catch and release or minimal
resistance at the end of movement (ROM) when the affected part is moved in flexion or extension
3.5 Slight increase in muscle tone, manifested by a catch following resistance throughout the remainder (less than half of ROM)
3 More marked increase in muscle tone through most of the ROM but affected part is easily moved
2 Considerable increase in muscle tone, passive movement difficult 1 Affected part rigid in flexion or extension 0 Unable to test
SENSIBILITY Sensory testing should always be conducted in a room that is at a comfortable temperature (as close to room temperature as possible). When applying the stimulus to the hands the examiner must ensure that he/she does not touch the hand as this can alter the individual’s ability to sensate accurately. Prior to beginning the testing the subject should be oriented to the test by demonstration on an area of intact sensation such as the face. The examiner will be standing beside the bed or seated across from the subject. The test is performed with the subjects eyes closed or occluded. The forearm and hand should be supported in supination or pronation with a towel or a pillow (not with the examiners hands). A circumferential 2 inch �elcro strap may be used to secure the hand to the pillow allowing for access to the palm and finger tips during the second part of the testing. Semmes Weinstein Monofilaments (SWM) – The monofilaments should be applied to the 6 points distal to proximal (test points 1 to 6 in Figure 2). The filament should be applied until it bends: applying for 1.5 seconds, holding for 1.5 seconds, and removing for 1.5 seconds. Filament 3.61 is to be applied three times at all test locations, 2/3 positive responses indicates intact sensibility of that force. The assessor should determine if the participant has sensation by asking “do you feel a touch?” and following by “where do you feel the touch?”. The remaining three filaments are applied once. The first filament which is applied three times, all dorsal test locations (points 1-3) can be tested once and then repeated in the same order, and then repeated a third time, after that, the hand can be turned and the palmar side of the hand can be tested in the same way. After recording all three applications of the filament 3.61, it can be determined if the test location is positive to that filament. Delayed responses more than three seconds are abnormal. The 3.61 filament is first
135
used in all test locations. If the patient feels the filament in all areas the examination is complete. It will not be necessary to use the other filaments. If the patient does not respond to the 3.61 filament the next heavier filament is used. Only test locations which do not respond to the previous filament need to be tested with the next filament. The exam continues until the patient recognizes a force in all test locations or until it is established that he/she does not feel even the heaviest filament. When the response is positive for a particular filament a check can be put in the associated box. When all the test locations have been tested the filament force should be written into the final box labeled g/F followed by the associated score in the score box (see Table 6). Figure 2: Diagram for Sensibility Test Locations
Table 6: Scoring of Pressure Sensibility with the Semmes Weinstein Monofilaments
Static Two Point Discrimination (S2PD) – The examiner should begin with the 5mm stimulus and apply it to each point (test locations 7 to 9, Figure 2) and record whether the individual detects 1 or 2 points, by asking the participant if they feel one or two points. A lateral hold of the digit might be necessary, which should be applied at the IP joint of the digit. The examiner records nothing in the box if the individual does not detect two points and a check mark if he/she does. If the subject feels two points then move to the next test location and proceed. If the subject feels only one point then the examiner should apply the next level of stimulus to the same test location and repeat this pattern until there is a positive response for the location or there is no response at the last level of stimuli. Complete one point before moving to the next point. At the end of the testing record the associated score in the appropriate box as per Table 7. Table 7: Scoring of the Static Two Point Discrimination Discriminator Distance 5mm 8mm 12mm 15mm No response Score 4 3 2 1 0
Filament Label
3.61 4.31
4.56
6.65 No Respon
se Filament Force
0.217
g/F
2.35
g/F
4.19
g/F
279.4
g/F
No Respon
se Score 4 3 2 1 0
Semmes Weinstein Monofilaments Static Two Point Discrimination
136
PREHENSION Prehension is assessed both qualitatively and quantitatively. Qualitative Prehension The aim of this subsection is to ensure that the early movement is captured before an individual may be ready for a seated assessment. No specific positioning of the patient is required, but appropriate positioning of the hand for movement should be ensured. The patient is asked to form the following prehension patterns with each hand separately. The requested movement and grasp patterns can be demonstrated by the examiner. The purpose of this testing is to establish which components of the finger-hand-forearm can be actively or passively positioned and directed to allow a grasp function and if the movement is wrist dominant. The intent is to establish whether the participant can perform a limited movement that does or does not include the components to develop an active grasp. The assessor should be looking to isolate, wrist, fingers and thumb. The basic pattern for grasping might be visible although the patient yet can not quite grasp. Table 8: Qualitative Grasp, Instructions and Scoring Qualitative Grasps
Muscles Myotome Task
Cylindrical Grasp FDS, FDP, OPPONENS
C7-T1 Neutral wrist position and finger movements performed with gravity eliminated
Lateral Key Pinch
ADP, FPL, Index FDS
C6-T1 Neutral wrist position and finger movements performed with gravity eliminated
Tip to Tip Pinch Index FDP, OPPONENS
C7-T1 Neutral wrist position and finger movements performed with gravity eliminated
Scoring: 0 – Subject is not able to position the wrist or fingers in any specific pattern for the requested grasp. 1 – Subject is able to move the wrist actively and fingers passively assume the requested prehension pattern or the subject is able to begin a potential grasp with use of the wrist and no subsequent finger movement. 2 – Subject is able to partially or completely move the fingers actively into the requested prehension pattern (combination of wrist movement and intrinsic hand muscle activation) but fails to generate force because the grasp is acquired through passive positioning from the activating the wrist. 3 – Subject is able to actively position the fingers and/or thumb into the requested prehension pattern with normal wrist movement for a grasp, touching the opposite finger(s) or the palm with some noticeable active force. 4 – Subject is able to perform the grasp with normal strength (like in a normal shaking hand)
Quantitative Prehension Patient positioning · The patient is positioned in a sitting position symmetrically in front of a table. Additional support for trunk stability is allowed. This includes for example the use of a belt but also sitting in bed supported by the back rest or using the bed side table to set up the test.
137
· A change in position, concerning the person’s angle to the table, is not allowed during the standardised test administration. · The patient is asked to use his favourite grasping form when performing the tasks. · The test is conducted twice, for each hand individually. · The stabilisation of the objects/test board, if necessary, is done by the examiner. · The test board is placed parallel to the edge of the table, in front of the patient. Moving the board in a parallel line to the table’s edge is permitted. Turning or rotating of the board is not permitted (see picture). · All other items are placed on the table in front of the patient. Procedure The required material for the different tasks is placed on the table in front of the patient just prior to the performance of each task. Prior to the first test administration, the patient is allowed to perform each task once as a rehearsal, without being scored. This will allow familiarization with the task and reduce the learning effect. The rehearsal time is limited to 1 minute for each task. The precise administration procedure for each task can be found in the table below. The examiner clocks the time required to accomplish each task. The clocking starts at a clear signal “start” and ends when the task is fully completed. The material can only be touched or grasped after the “start” signal. The initiation of each task is defined by clear activity, such as moving pegs, lifting up coins, manipulating the bottle of water in the hand. To score a 1 at least one part of the task must be done i.e. lifting up a coin, grasping and/or moving a coin, holding/lifting the bottle. Only moving the hand is not regarded as “done part of the activity”, neither is just placing the hand on the test equipment. The examiner observes task performance focussing on the form of the grasp. The time required for task performance is recorded on the score sheet and the task is scored according to the scoring key in Table 9. Grasp performance leads to a score with an associated time that is recorded separately. The maximum score for each task is 5 points with a maximal total score of 60 points. To judge the quality of the performance, the examiner must refer to the description of the “expected performance”. This description defines the typical form of grasp used and performance with an unaffected hand (see Table 9). One minute and 15 seconds is allowed for the completion of each task, if the individual is unable to complete the task within 1 minute and 15 seconds score accordingly and move on to the next task. Dropping of objects: If a patient drops an object and it falls onto the table, still reachable for the patient to retrieve, the task is continued without stopping the clock (In North America the drops are counted and the number is entered in the # of drops column). If the object falls onto the floor or the lap of the patient and cannot be reached by the patient, the clock is stopped. The examiner can pick up the object and the task may be repeated. If the drop lands again on the floor or lap of the patient during the repeated execution, the task is judged as “not conducted” (0 points) and comments are noted.
138
Table 9: Quantitative Grasp, Instructions and Scoring Task / Instruction “Expected Normal” Performance
5. Take the bottle and pour the water into the cup, approx. ¾ full.
(lower rim of the black plastic ring)
Cylindrical grasp
The filled, already opened bottle (0.5L) and the cup are placed onto the table in front of the patient. ¾ filled corresponds to the lower rim of the black plastic ring around the cup. The task is completed if the water is poured in the cup and cup and bottle are put back down onto the table. The cup is stabilized by the examiner. 2. Unscrew the lids of the 2 jars and place them on the table.
Spherical grasp
The two jam jars are placed onto the table in front of the patient. The lids should sit tightly on the jar but should not require much strength to be opened (check before task administration!). The task is completed when both lids and jars lay on the table. The jars are stabilized by the examiner. 3. Pull the 9 pegs, one by one, out of the block and place them back into the markings on the opposite side.
Pad to Pad Prehension / Tripod pinch
The peg board is placed on the table in front of the patient. It does not matter if the pegs are moved from right to left or left to right; the patient can choose his preference. The task is completed when all 9 pegs are placed in the opposite board. The board is stabilized by the examiner or use of dycem. 4. Take the key from the table, insert it in the lock and turn it 90°.
Picking up the key from the table or sliding it over the table’s edge and putting it into the slot using Lateral Pad to Pad Prehension
The test board is placed on the table, parallel to the table’s edge, in front of the patient. The distance from the table’s edge to the test board is not important. The key is put on the table in front of the patient. The turning direction of the key is of no importance. The task is completed when the key was rotated 90°. The board is stabilized by the examiner or use of dycem.
5. Pick up the 4 coins, one by one, from the table and put them through the slot.
Picking up the coins from the table or sliding them over the table’s edge and putting them into the slot using Pad to Pad Prehension
The position of the test board is as described for task no. 4. The coins are placed side by side onto the table in front of the patient. The task is completed when all coins are dropped into the slot. The board is stabilized by the examiner or use of dycem. 6. Pick up the 4 nuts, one by one, from the table and screw them onto the matching screws.
Picking up nuts from table or sliding them over the table’s edge and screwing using Pad to Pad Prehension or Tip to Tip Prehension
139
The position of the test board remains the same as described for task no.4 and 5. The 4 nuts are placed side by side onto the table in front of the patient. The task is completed when all nuts are screwed onto the matching screws, the screw’s-end aligned with the nuts. Rotating the test board is not permitted in this task either. The board is stabilized by the examiner or by dycem. Scoring (a maximum of 1 minute and 15 seconds is allowed for each task) 0-the task can not be conducted at all 1 – the task can not be completed, (less than 50% of the task) and the expected grasp is not used 2 – the task is not completed, (50% or more of the task) and the expected grasp is not used 3 – the task is conducted (completed) using tenodesis or an alternative grasp other than the expected grasp 4 – the task is conducted using the expected grasp with difficulty (lack of smooth movement or difficult slow movement) 5 – the task is conducted without difficulties using the expected grasping pattern and unaffected hand function Note: 50% of task 1 is when the participant has begun to pour the water, 50% or task 4 is when the participant is able to get the key to insertion point.
140
SCORING SHEETS 1 – Demographics Subject ID: Assessment Number (circle)
1 2 3
Assessor ID: (circle) 1 2 Date Patient Name Age Gender Hand Dominance Pre-injury Post-injury Injury Date Injury Type Brief Description
Surgery and Date
Comments
ASIA Classification Motor Level R
Motor Level L
Sensory Level R
Sensory Level L
Motor Score R
Motor Score L
Sensory Score R
Sensory Score L
CUE Score SCIM Score 2 – ASIA Upper Extremity Scoring Myotome/ Dermatome Myotome LT-Dermatome PP-Dermatome
ASIA Level Right Left Right Left Right Left C2 C3 C4
C5 elbow flexor C6 wrist extensor
C7 Elbow extensor C8 Finger flexor
T1 Finger abductors Totals
141
3 – Strength – score 0 to 5 as per instructions; enter 0 to 5 in each box
Muscle Tone – score 0 to 5 as per Modified Ashworth Scale, Enter 0 to 5 in each box.
Right Muscle Group Left (T1-R) Elbow flexors (T1-L)(T2-R) Finger flexors (T2-L)(TT-R) Total out of 10 per arm (TT-L)
Manual Muscle Testing RIGHT MUSCLE GROUP LEFT
(ST1-R) Anterior Deltoid C5/6 (ST1-L)(ST2-R) Elbow Extensors C7 (ST2-L)(ST3-R) Elbow Flexors C5 (ST3-L) Proximal Arm (ST4-R) Wrist Extensors (ST4-L)(ST5-R) Extensor Digitorum (DIII) (ST5-L)(ST6-R) Finger Flexors (DIII) (ST6-L)(ST7-R) Flexor Pollicis Longus (ST7-L) Extrinsic Hand Muscles (ST8-R) Finger Abductors (ST8-L)(ST9-R) First Dorsal Interossei (ST9-L)(ST10-R) Adductor Pollicis Brevis (ST10-L)(ST11-R) Opponens Pollicis (ST11-L) Intrinsic Hand Muscles (Stotal-R) Total out of 55 per arm (Stotal-L)
142
4 – SENSIBILITY Monofilament and Two Point Discrimination Test Scores 4.1 SWM Threshold Sensed Right Hand AREA 3.61(4) 3.61(4) 3.61(4) 4.31(3) 4.56(2) 6.65(0) NR(0) Score
1 (S1-R)
2 (S2-R)
3 (S3-R)
4 (S4-R)
5 (S5-R)
6 (S6-R)
Total out of 24 (SWMT-R)
4.2 SWM Threshold Sensed Left Hand AREA 3.61(4) 3.61(4) 3.61(4) 4.31(3) 4.56(2) 6.65(0) NR(0) Score
1 (S1-L)
2 (S2-L)
3 (S3-L)
4 (S4-L)
5 (S5-L)
6 (S6-L)
Total out of 24 (SWMT-L)
4.3 Static Two Point Discrimination Sensed Right Hand
AREA 5mm (4) 8mm (3) 12mm (2)
15mm (1)
NR (0) Score
7 (S7-R)
8 (S8-R)
9 (S9-R)
Total out of 12 (2PDT-R)
4.4 Static Two Point Discrimination Sensed Left Hand
AREA 5mm (4) 8mm (3) 12mm (2)
15mm (1)
NR (0) Score
7 (S7-L)
8 (S8-L)
9 (S9-L)
Total out of 12 (2PDT-L)
143
5 – PREHENSION 5.1 – Qualitative Grasp
Right Qualitative Grasps Left (QLG1-R) Cylindrical Grasp (QLG1-L)(QLG2-R) Lateral Key Pinch (QLG2-L)(QLG3-R) Tip to Tip Pinch (QLG3-L)(QLT-R) Total out of 12 (QLT-L) 5.2 – Quantitative Grasp Task/ Instruction Expected
Prehension Time Right
Score Right
Drops Right
Time Left
Score Left
Drops Left
1. Take the bottle and pour the water into the cup, approx. ¾ full.(lower rim of the black plastic ring)
Cylindrical grasp
(TR-1) (QN1-R) (TL-1) (QN1-L)
2. Unscrew the 2 lids of the jam jars and put them onto the table.
Spherical grasp
(TR-2) (QN2-R) (TL-2) (QN2-L)
3. Pull the 9 pegs, one by one, out of the foam and stick them back into the markings on the opposite side.
Tip to Tip Prehension
(TR-3) (QN3-R) (TL-3) (QN3-L)
4. Take the key from the table, insert it in the lock and turn it 90°.
Lateral Pad to Pad Prehension
(TR-4) (QN4-R) (TL-4) (QN4-L)
5. Pick up the 4 coins, one by one, from the table and put them through the slot.
Pad to Pad Prehension.
(TR-5) (QN5-R) (TL-5) (QN5-L)
6. Pick up the 4 nuts, one by one, from the table and screw them on the matching screws.
Pad to Pad Prehension or Tip to Tip Prehension
(TR-6) (N6-R) (TL-6) (QN6-L)
144
6 SUMMARY AND TOTAL SCORES Right Left
STRENGTH-Upper limb (55/55) MUSCLE TONE – Elbow/Finger Flexors (10/10) SENSIBILITY ASIA – LT (16/16) SENSIBILITY ASIA – PP (16/16) SENSIBILITY Pressure – Monofilaments (24/24) SENSIBILITY Two Point Discrimination (12/12) PREHENSION – Qualitative Grasp (12/12) PREHENSION – Quantitative Grasp (30/30) TOTAL SCORE-(175/175)
REFERENCES 1. Marino, R.J., Barros, T., Biering-Sorensen, F., Burns, S.P., Donavan, W.H., Graves, D.E., Haak, M., Hudson, L.M., and Priebe, M.M. International Standards for Neurological Classification of Spinal Cord Injury. Sixth Edition 2002. 2. Mackin, E., Callahan, A., Skiver, T., Schneider, L. & Osterman, A. (2002) Hunter-Mackin-Callahan Rehabilitation of the Hand and Upper Extremity, Fifth Edition. St. Louis, Missouri: Mosby. 3. Daniels L, Worthingham C. (1995). Daniels and Worthingham’s Muscle Testing: Techniques of Manual Examination, 6th Ed. Philadelphia: WB Saunders Co. 4. Bohannon RW, and Smith MB. (1987). Interrater reliability of a modified Ashworth scale of muscle spasticity. Phys. Ther. 67(2): 206-7. 5. Kendall F, McCreary EK, Provance PG. (1993). Muscles: Testing and Function, 4th Ed. Philadelphia: Williams & Wilkins.
145
B: GRASSP Version 1.0
GRADED and REDEFINED ASSESSMENT
Of STRENGTH, SENSIBILITY and PREHENSION Version 1.0
(GRASSP Version 1.0)
146
International GRASSP Research and Design Team: Sukhvinder Kalsi-Ryan University of Toronto, Graduate Department of Rehabilitation Science Toronto Western Hospital Toronto Rehabilitation Institute Armin Curt University of British Columbia Vancouver Coastal Health ICORD Susan Duff Thomas Jefferson University Michael Fehlings Toronto Western Hospital, Krembil Neuroscience Program University of Toronto Claudia Rudhe Balgrist University Hospital, Zürich Molly Verrier University of Toronto, Department of Physical Therapy Toronto Rehabilitation Institute Funding: Christopher and Dana Reeve Foundation, Rick Hansen Foundation, Toronto Rehabilitation Institute Student Scholarship Fund Supporting Organizations: North American Clinical Trials Network, European Clinical Trials Network, ICORD, Krembil Neurosciences Centre
147
TABLE OF CONTENTS
GRASSP Design Team 1
Introduction 3
Purpose of the Measure 3
Design and Development of the GRASSP 3
Modules 4
Instructions for implementation of the GRASSP 4
Strength 4
Sensibility 7
Prehension 8
Scoring Sheets 13
Demographics 13
Strength 13
Sensibility 14
Prehension 15
Summary and Total Scores 16
References 16
148
INTRODUCTION Initiated by the North American Clinical Trials Network (NACTN) and the European Clinical Trials Network (EUCTN), a meeting was held on May 12 and 13, 2006 in Chicago (local organizer Drs. Zev Rymer and Lisa-Ann Wuermser, financially supported by the Christopher and Dana Reeve Foundation) to discuss the measurement of hand impairment and function in patients suffering from cervical spinal cord injury (cSCI). Members of the networks and independent clinical specialists in hand measurement and therapy, as well as researchers with expertise in engineering and computer technology discussed the development of a comprehensive protocol to assess upper limb impairment and recovery post cSCI. The result of the meeting was a task force to further develop a clinical assessment protocol of hand function by modifying existing tools and introducing new measures that would allow for the quantification of change in hand function for individuals with cSCI. The GRASSP is a combined effort by six clinicians/researchers who have contributed their work (results of tool development through previous research, in some instances graduate studies), skills, and time. The GRASSP is a mosaic of the Link Hand Function Test (Link, 2004) and The Tetraplegia Hand Measure (Kalsi-Ryan et al. 2004). PURPOSE OF THE MEASURE The overall objective for the assembly of the GRASSP was to develop a clinical research tool that could capture information on hand impairment from the cervical (C0-T1) spinal cord injury (SCI) population, obtain integrated sensory and motor impairment data, and discriminate the population according to the level of lesion. The purpose of this project was to design a hand impairment tool: 1) that was highly responsive (sensitive) to change over time; 2) that could assess the extent of spontaneous (natural) recovery; and 3) be applicable for use in clinical trials to evaluate the effect of novel interventions (pharmacological and surgical). The GRASSP is recommended for use in the very early acute phases out to approximately one year post injury. Use of the GRASSP is recommended when a change in neurological status is being assessed. DESIGN AND DEVELOPMENT OF THE GRASSP The GRASSP is a framework that assembles different clinical tools to measure the various aspects of complex sensori-motor hand function. The GRASSP is embedded with currently existing measures of upper limb function. For the development of the GRASSP each of the three modules was assigned to one of the measurement developers (Strength – Susan Duff, Sensibility- Sukhvinder Kalsi-Ryan, and Prehension – Claudia Link-Rudhe) under the direction of Armin Curt and Molly Verrier. Although, individuals were responsible for separate modules all members of the task force made significant contributions to all components of the GRASSP. MODULES The GRASSP is comprised of three separate modules, Strength, Sensibility, and Prehension. Multiple modules allow for a comprehensive assessment at multiple time points in the post-injury continuum. Each module can be tested according to the scheduled timeline provided by a trial protocol. INSTRUCTIONS FOR IMPLEMENTATION OF THE GRASSP: Consent: Always obtain informed consent from the subject (patient) and collect the necessary demographic data based on interview with the patient/family and chart review. Positioning the Patient: In the acute period the patient is lying supine with both arms exposed to the shoulders and should be tested in this position. During other test sessions the subject should be seated in his/her own seating system with his/her appropriate supports. During all testing, the entire upper extremity
149
should be exposed (up to the shoulder). An adjustable table which can move in and out of wheelchair space will be required to perform the assessment. The subject’s hands should be positioned on the table, with approximately 30 degrees of shoulder flexion, 65 degrees of elbow flexion and the hands and distal half of the forearms supported on the table. This position can be modified slightly to ensure comfort for the individual being tested. The room where the testing will be done should be well lit. Length of the Testing: The time required to complete all of the tests in one session is approximately 30 - 45 minutes (depending on patient ability). It is not recommended to break the testing up into two sessions over two days as an individual’s response can vary and recovery can potentially affect the results. For the best outcome it is recommended to complete the testing in one session, however, between sub-tests the individual and the examiner can break and stretch. In the early phases of injury (0-21 days) it is recommended to only perform the partial GRASSP which consists of the sensory, strength and qualitative prehension portions (15-20 minutes) of the test. After the early phase the full GRASSP is recommended. STRENGTH Muscles specific to the upper limb and hand were added to the ASIA (Marino et al. 2004) repertoire of testing to establish greater sensitivity to potential change post-injury. Strength will be assessed with Manual Muscle Testing (MMT) (Daniels & Worthington, 1995). An isotonic muscle contraction will be required by the subject to grade muscle strength. Specifically, resistance should be given at the distal end of the moving bone while the subject moves the limb through the specific range (Daniels and Worthingham, 1995). The following table defines the scaling for the muscle testing and the instructions for testing each muscle. Muscle Testing Prior to beginning muscle testing the subject should be oriented to the test by demonstration on an active body part. If the testing will be done in supine the examiner should stand comfortably at the bedside. If the subject is seated, the examiner may choose to stand next to the wheelchair or sit next to/across from them. During assessment of the distal arm musculature the subjects’ forearm should rest on an adjustable table. Begin by testing the muscle for a grade three (range against gravity), ensuring the joints are isolated. If the individual is able to move through full range of motion (ROM) against gravity then the same movement should be tested with resistance for a grade 4 or 5 through full ROM. The examiner will grade the individual according to the scoring key. Resistance is given at the distal end of the moving bone during an isotonic contraction. Table 1 defines the muscles to be tested, the starting position, the stabilization and resistance required for testing these muscles and the scoring key to be used. Remember that for finger muscles gravity does not have an effect, which defines grade 2 as: movement of the corresponding body part but not through the full range of motion and grade 3 as: movement through full ROM. All MMT scoring should be recorded in the scoring sheets section. Note: 1) Full range of motion for anterior deltoid should be established and then measured based on available range (available should be considered full range). 2) For elbow extension if the starting position (full elevation) is not feasible then elbow extension can be tested in 90 degrees of shoulder elevation.
150
Table1: Strength Testing and Instructions
ASIA Muscles in Italics (Daniels and Worthingham, 1995; Kendall, McCleary and Provance, 1993)
Muscle Action Stabilization Starting position Resistance Anterior/Mid Deltoid C5-6
Shoulder abduction 90°/ flexion in supine
Trunk 00 shoulder abduction/ flexion Anterior, distal humerus
Elbow flexors (Biceps) C5-6
Flex elbow Humerus Full Elbow extension, shoulder adduction, forearm in supination
Distal, volar forearm
Elbow extensors (Triceps) C6-C8
Extend elbow Humerus Elbow flexion, shoulder abduction (hand behind the head, or 90° abd and full inward rotation of the humerus
Distal forearm
Wrist extensors C6-C8
Extend wrist Forearm Wrist in flexion, forearm in pronation
Distal, dorsal 3rd metacarpal
Extensor Digitorum C6-C8
Extend MP's digits 2-5
2-5 Metacarpals
Flexion IP's / MP's digits 2-5, forearm in pronation (fingers hanging over edge of table)
Dorsal Proximal phalanges digits 2-5
Opponens Pollicis C6-C7
Rotate 1st metacarpal toward 5th digit pad
wrist/ 2-5 metacarpals
Thumb in a resting posture next to 2nd metacarpal, slightly abducted, forearm in supination
Volar proximal phalanx with derotating pressure
Flexor Pollicis Longus C6-C8
Flex thumb IP joint
Thumb Proximal phalanx/ metacarpal
00 thumb IP extension, MP supported in 00 extension
Volar thumb pad
Finger flexors (3rd FDP) C7-T1
Flex DIP joint of 3rd digit
MP, PIP joint 3rd digit
3rd digit DIP extension, 3rd PIP/MP supported in 00 extension on table
3rd volar finger pad
Finger abductors (5th) C8-T1
Abduct 5th digit
5th metacarpal 5th digit adducted, MP’s extended, forearm in pronation. (Position hand on sheet of paper to reduce friction)
Ulnar side of inter-phalangeal joint of 5th digit
First Dorsal Interossei C8-T1
Abduct index 2nd metacarpal Index adducted next to long finger, MP’s extended, forearm in pronation (Position hand on sheet of paper to reduce friction on the table)
Radial side of inter-phalangeal joint of 2nd phalanx
151
Table 2: Scoring for Manual Muscle Testing
SENSIBILITY Sensory testing should always be conducted in a room that is at a comfortable temperature (as close to room temperature as possible). When applying the stimulus to the hands the examiner must ensure that he/she does not touch the hand as this can alter the individual’s ability to sensate accurately. Prior to beginning the testing the subject should be oriented to the test by demonstration on an area of intact sensation such as the face. The examiner will be standing beside the bed or seated across from the subject. The test is performed with the subjects eyes closed or occluded. The forearm and hand should be supported in supination or pronation with a towel or a pillow (not with the examiners hands). A circumferential 2 inch Velcro strap may be used to secure the hand to the pillow allowing for access to the palm and finger tips during the testing. Semmes Weinstein Monofilament Testing (SWM) The monofilaments should be applied to all 6 points (test points 1 to 6 in Figure 1). The filament should be applied until it bends: applying for 1.5 seconds, holding for 1.5 seconds, and removing for 1.5 seconds. Filament 3.61 is to be applied three times at all test locations, 2/3 positive responses indicates intact sensibility of that force. The assessor should determine if the participant has sensation by asking “do you feel a touch?” and following by “where do you feel the touch?” It the patient is not able to adequately localize the stimulus then he/she is not feeling the applied stimulus. The remaining three filaments are applied once. The test is started on the dorsal side of the hand. The first filament (3.61) is applied three times; all dorsal test locations (points 1-3) can be tested before moving to the palmar test locations (4-6). Delayed responses of more than three seconds are abnormal. If the patient feels the first filament in all areas the examination is complete. It will not be necessary to use the other filaments. If the patient does not respond to the 3.61 filament the next heavier filament is used. Only test locations which do not respond to the previous filament need to be tested with the next filament. The exam continues until the patient recognizes a force in all test locations or until it is established that he/she does not feel even the heaviest filament. When the response is positive for a particular filament a check can be put in the associated box. When all the test locations have been tested the filament force should be scored appropriately into the final box score. Table 3 defines the score associated to the log label of the monofilament (Mackin et al. 2002).
0 Absent – No palpable muscle contraction 1 Trace – Palpable Muscle Contraction 2 Poor – Moves full ROM with gravity eliminated 3 Fair – Moves full ROM against gravity without added resistance 4 Good – Moves through full ROM against gravity against moderate resistance 5 Normal – Moves through full ROM against gravity against maximal resistance
152
Figure 1: Diagram for Sensibility Test Locations
Table 3: Scoring of Pressure Sensibility with Semmes Weinstein Monofilaments
PREHENSION Prehension is assessed both qualitatively and quantitatively. A. Qualitative Prehension Testing The aim of this sub-section is to ensure that the early movement is captured before an individual may be ready for a seated assessment. No specific positioning of the patient is required, but appropriate positioning of the hand for movement should be ensured. The patient is asked to form three prehension patterns with each hand separately. The requested movement and grasp patterns can be demonstrated by the examiner. The purpose of this testing is to establish which components of the finger-hand-forearm can be actively or passively positioned and directed to allow a grasp function and if the movement is wrist dominant. The intent is to establish whether the participant can perform a limited movement that does or does not include the components to develop an active grasp. The assessor should be looking to isolate, wrist, fingers and thumb. The basic pattern for grasping might be visible although the patient yet can not quite grasp. In the very early stages a patient will require the assessor to support the hand so that the patient can see it. This may require providing the neutral position of the wrist as well. Table 4 defines the three grips to be tested and the associated scoring. Table 4: A. Qualitative Prehension, Instructions and Scoring Qualitative Prehension Task Cylindrical Grasp Neutral wrist position and finger movements performed with gravity
eliminated Lateral Key Pinch Neutral wrist position and finger movements performed with gravity
eliminated Tip to Tip Pinch (thumb and index finger)
Neutral wrist position and finger movements performed with gravity eliminated
Filament Label 3.61 4.31 4.56 6.65 No Response Filament Force in g/F 0.217 2.35 4.19 279.4 No Response Score 4 3 2 1 0
SWM Test Locations
153
Scoring 0 - Subject is not able to position the wrist or fingers in any specific pattern for the requested grasp. 1 - Subject is able to move the wrist actively and fingers passively assume the requested prehension pattern (able to begin a grasp using the wrist and no finger movement) 2 - Subject is able to partially or completely move the fingers actively into the requested prehension pattern (combination of wrist movement and intrinsic hand muscle activation) but fails to generate force because the grasp is acquired through passive positioning from the activating the wrist. 3 - Subject is able to actively position the fingers and/or thumb into the requested prehension pattern with normal wrist movement for a grasp, touching the opposite finger(s) or the palm with some noticeable active force. 4 - Subject is able to perform the grasp with normal strength (in a normal shaking hand).
B. Quantitative Prehension Testing The patient is positioned in a sitting position symmetrically in front of a table. Additional support for trunk stability is allowed. This includes for example the use of a belt but also sitting in bed supported by the back rest or using the bed side table to set up the test. · A change in position, concerning the person’s angle to the table, is not allowed during the standardised test administration. · The test is conducted twice, once for the right hand and then once for the left hand. · The stabilisation of the objects/test board, if necessary, is done by the examiner. · The test board is placed parallel to the edge of the table, in front of the patient. Moving the board in a parallel line to the table’s edge is permitted. Turning or rotating of the board is not permitted (see picture). · All other items are placed on the table in front of the patient.
Procedure The required material for the different tasks is placed on the table in front of the patient just prior to the performance of each task. Prior to the first test administration, the patient is allowed to perform each task once as a rehearsal, without being scored. This will allow familiarization with the task and reduce the learning effect. The rehearsal time is limited to 1 minute for each task. The precise administration procedure for each task can be found in the table below.
The examiner times each task. The timing starts at a clear signal “start” by the examiner and ends when the task is fully completed. The material can only be touched or grasped after the “start” signal by the examiner. Table 5 defines the instructions to the examiner and the patient. The initiation of each task is defined by clear activity, such as moving pegs, lifting up coins, manipulating the bottle of water in the hand. To score a 1 at least one part of the task must be done (i.e. lifting up a coin, grasping and/or moving a coin, holding/lifting the bottle). Moving the hand alone is not regarded as "done part of the activity"; neither is placing the hand on the test equipment. The examiner observes task performance focussing on the form of the grasp. The time required for task performance is recorded on the score sheet and the task is scored according to the scoring key in Table 6. Quantitative prehension performance leads to a score with an associated time that is recorded separately. The maximum score for each task is 5 points with a maximal total score of 30 points per hand. To judge the quality of the performance, the examiner must refer to the description of the “expected performance”. This description defines the typical form of grasp used and performance with an unaffected hand (see Table 5). One minute and 15 seconds is allowed for the completion of each task, if the individual is unable to complete the task within 1 minute
154
and 15 seconds score accordingly and move on to the next task (Sollerman and Ejeskar, 1995). There is no specific order to the tasks and they appear in order of simplicity (least difficult to more difficult). Dropping of objects: If a patient drops an object and it falls onto the table, still reachable for the patient to retrieve, the task is continued without stopping the clock the drops are counted and the number is entered in the # of drops column. If the object falls onto the floor or the lap of the patient and cannot be reached by the patient, the clock is stopped. The examiner can pick the object up and the task may be repeated. If the drop lands on the floor or lap of the patient again, during the repeated execution, the task is judged as “not conducted" (0 points) and comments are noted. Table 5: Quantitative Prehension, Instructions and Scoring Task and Instructions to Examiner Expected Prehension Pattern for Task Task Instructions for Subject 1. The filled, already opened bottle (0.5L) and the cup are placed onto the table in front of the patient. The task is completed if the water is poured in the cup and cup and bottle are put back down onto the table. The cup is stabilized by the examiner. 50% of task 1 is when the participant has begun to pour the water. Cylindrical grasp 1. Take the bottle and pour the water into the cup,
approx. ¾ full. 2. The two jam jars are placed onto the table in front of the patient. The lids should sit tightly on the jar but should not require much strength to be opened (check before task administration). The task is completed when both lids and jars lay on the table. The jars are stabilized by the examiner. 50% of task 2 is when the participant has removed one lid. Spherical grasp 2. Unscrew the lids of the 2 jars and place them on the
table. 3. The peg board is placed on the table in front of the patient. It does not matter if the pegs are moved from right to left or left to right; the patient can choose his preference. The task is completed when all 9 pegs are placed in the opposite board. The board is stabilized by the examiner or use of dycem. 50% of task 3 is when the participant has inserted 4 pegs. Tip to Tip pinch (thumb and index finger) / Tripod pinch
3. Pull the 9 pegs, one by one, out of the block and place them back into the markings on the opposite side.
4. The test board is placed on the table, parallel to the table’s edge, in front of the patient. The distance from the table’s edge to the test board is not important. The key is put on the table in front of the patient. The turning direction of the key is of no importance. The task is completed when the key was rotated 90°. The board is stabilized by the examiner or use of dycem. 50% of task 4 is when the participant is able to get the key to insertion point. Lateral Key pinch 4. Take the key from the table, insert it in the lock and
turn it 90°. 5. The position of the test board is as described for task no. 4. The coins are placed in a row on the table in front of the patient. The task is completed when all coins are dropped into the slot. The board is stabilized by the examiner or use of dycem. 50% of task 5 is when the participant is able to insert 2 coins Tip to Tip Pinch (thumb and index finger) 5. Pick up the 4 coins, one by one, from the table and
drop them through the slot.
155
6. The position of the test board remains the same as described for task no.4 and 5. The 4 nuts are placed in a row onto the table in front of the patient. The task is completed when all nuts are screwed onto the matching screws, the screw’s-end aligned with the nuts. Rotating the test board is not permitted in this task either. The board is stabilized by the examiner or by dycem. 50% of task 6 is when the participant is able to screw two nuts on Tip to Tip pinch (thumb and index finger) and/or Tripod pinch
6. Pick up the 4 nuts, one by one, from the table and screw them onto the matching screws.
Table 6: Scoring for the Quantitative Prehension Scoring (a maximum of 1 minute and 15 seconds is allowed for each task) 0 - the task can not be conducted at all 1 - the task can not be completed, (less than 50% of the task) 2 - the task is not completed, (50% or more of the task) 3 - the task is conducted (completed) using tenodesis or an alternative grasp other than the expected grasp 4 - the task is conducted using the expected grasp with difficulty (lack of smooth movement or difficult slow movement) 5 - the task is conducted without difficulties using the expected grasping pattern and unaffected hand function.
156
SCORING SHEETS 1 - Demographics Patient Name Examiner Assessment Number 1 2 3 4 5 6 Date of Assessment DOB Gender Hand Dominance Pre-injury Post-injury Injury Date Injury Type Brief Description
Surgery/Intervention and Date
Comments
2 - Strength - score 0 to 5 as per instructions in each box, then sum for each side
Right Muscles Tested for MMT Left Anterior Deltoid Elbow Flexors Elbow Extensors Wrist Extensors Extensor Digitorum (DIII) Opponens Pollicis Flexor Pollicis Longus Finger Flexors (DIII) Finger Abductors First Dorsal Interossei
/50 Total out of 50 for each side /50
157
3 – Sensibility
SWM Threshold Scores
Right Hand Left Hand
3.61 (4) 3.61 (4) 3.61 (4) 4.31 (3) 4.56 (2) 6.65 (1) NR (0)
Score Area 3.61 (4) 3.61 (4) 3.61 (4) 4.31 (3) 4.56 (2) 6.65 (1) NR (0)
Score
1
2
3
Dorsal Total /12 Dorsal Total /12
4
5
6
Palmar Total /12 Palmar Total /12
Dorsal Total+Palmar Total=Total SWM /24 Dorsal Total+Palmar Total=Total SWM /24
158
4 - Prehension A - Qualitative Prehension
Right Qualitative Grasps Left Cylindrical Grasp Lateral Key Pinch Tip to Tip Pinch
/12 Total out of 12 /12
B - Quantitative Prehension Right Task/ Instruction
Expected Prehension Left
Time Score Drops Time Score Drops 1. Take the bottle and pour the water
into the cup, approx. ¾ full. Cylindrical grasp
2. Unscrew the 2 lids of the jam jars and put them onto the table. Spherical grasp
3. Pull the 9 pegs, one by one, out of the foam and stick them back into the markings on the opposite side. Tip to Tip pinch
4. Take the key from the table, insert it in the lock and turn it 90°. Lateral Key pinch
5. Pick up the 4 coins, one by one, from the table and put them through the slot. Tip to Tip Pinch
6. Pick up the 4 nuts, one by one, from the table and screw them on the matching screws. Tip to Tip pinch and/or Tripod pinch
Total Score /30 5 – Summary and Total Scores
Right Left STRENGTH-Upper limb (50/50) SWM –DORSAL (12/12) SWM-PALMAR (12/12) PREHENSION – Qualitative (12/12) PREHENSION – Quantitative (30/30) TOTALS /116
159
REFERENCES 1. Link C. The Link Hand Function Test for Patients with a Cervical Spinal Cord Injury: An Intra-Rater and Inter-Rater Reliability and Expert Opinion Evaluation Study (2004). Hogeschool van Amsterdam, Institute of Occupational Therapy, The Netherlands; University College-South, School of Occupational Therapy and Physiotherapy, Denmark; Karolinska Institutet Division of Occupational Therapy, Sweden. 2. Kalsi-Ryan S, Beaton D, McIlroy W, Fehlings M, Verrier M. The Development of the Quadriplegia Hand Assessment Tool (Q-HAT) - A Discriminative and Evaluative Approach. Journal of Spinal Cord Medicine. 2004:27(2);164. 3. Marino RJ, Barros T, Biering-Sorensen F, et al. International Standards for Neurological Classification of Spinal Cord Injury. Sixth Edition 2004. 4. Daniels L, Worthingham C. Daniels and Worthingham’s Muscle Testing: Techniques of Manual Examination. Washington: WB Saunders Co. Sixth Edition 1995. 5. Kendall F, McCreary EK, Provance PG. Muscles: Testing and Function, Philadelphia: Williams & Wilkins. Fourth Edition 1993. 6. Mackin E, Callahan A, Skiver T, Schneider L & Osterman A. Hunter-Mackin-Callahan Rehabilitation of the Hand and Upper Extremity. St. Louis, Missouri: Mosby. Second Edition 2002. 7. Sollerman C. & Ejeskar A. Sollerman Hand Function Test: A Standardized Method and its Use in Tetraplegic Patients. Scandinavian Journal of Plastic and Reconstructive Hand Surgery 1995:29;167 – 176.
160
Appendix 2
A: Sensibility Questionnaire
161
A: Sensibility Questionnaire compiled by Claire Bombardier, University of Toronto, Institute for Work and Health
SENSIBILITY ASSESSMENT WORKSHEET PUPOSE, POPULATION, SETTING
1. Is the purpose clearly specified? What concept or attribute is being measured? What will it be used for (ie. function: description, prognosis, evaluation)? Why is this measurement needed?
2. Is the population clearly specified (ie. Framework)? 3. Is the setting clearly specified (ie. Framework)?
CONTENT VALIDITY
4. Are all relevant domains and exclusions clearly specified?
Are there important omissions? Are there inappropriate inclusions?
5. Is the breakdown of domains and /or categories appropriate, considering the purpose?
Are they mutually exclusive? 6. What was the method of selecting items for inclusions?
FACE VALIDITY
7. Is each element phrased in a suitable way? 8. Are the response categories for each element appropriate? 9. If a global rating is calculated, is the method of aggregation appropriate?
FEASIBILTY
10. Is it easy to understand? Are the items, their scaling and the aggregate score simple?
11. Is it easy to use? Does the data collection sheet conform to basic principles of questionnaire design? Are there instructions and definitions provided? Are procedures standardized?
12. Is is acceptable to the patient and to the observer? 13. Is the format for administration appropriate for your purpose? Does it require
special tests or special skills? 14. Is the administration time suitable?
162
Appendix 3
Study Design and Protocol
Figure 1: North American Data Collection
Figure 2: European Data Collection
163
Study Design Used to Collect Data: The general methodology for this study was based on establishing psychometric properties. Therefore, a stable sample was targeted and repeated administration of the GRASSP was conducted along with the administration of the comparator measures (SCIM, CUE, ISCSCI). Figures 1 and 2 define the visits, number of GRASSP administrations, raters and the use of data for analysis. Each Figure defines how the study design was similar and different in North America and Europe. Figure 1: North American Data Collection
Figure 1 refers to the method of data collection in the North American centres (n=4) and how the data was used in the analysis.
Visit I
TEST 1
TEST 2
Rater 1
Visit II
TEST 3
SCIM CUE ISCSCI
Inter rater Reliability n=45
Test Retest Reliability n=45
Construct Validity n=45
Rater 1
Rater 2
Rater 2
164
Figure 2: European Data Collection
Figure 2 refers to the method of data collection in the European centres (n=3) and how the data was used in the analysis.
Visit I
TEST 1 Rater 1
Visit II
TEST 2
SCIM CUE ISCSCI
Inter rater Reliability n=27
Construct Validity n=27
Rater 2
Rater 1
165
Appendix 4
A: Capabilities of Upper Extremity Questionnaire (CUE) B: International Standards of Neurological Classification of Spinal Cord Injury (ISCSCI)
C: Spinal Cord Independence Measure (SCIM)
166
A: CUE
The CUE is a 32 item questionnaire developed to assess difficulty in performing certain
activities with the upper extremities. The CUE is a self perceived measure of upper limb
function which incorporates components of reach, grasp and manipulation. The scoring is
based on degree of difficultly in performing tasks which is important in defining functional
limitation (Marino et al., 1998). The questions asked are related to one’s perception of how
difficult a task may be. Psychometric properties of the CUE have been reported as 0.92 for
test retest reliability, tested by Cronbach’s alpha and 0.74 for concurrent validity with the
Functional Independence Measure, tested by Pearson correlation coefficient (Marino et al.,
1998). The CUE was selected as a comparator measure to establish validity and determine
relationships between impairment and self perceived function. Although, the CUE is a
measure of function it incorporates the three aspects of upper limb function (reach, grasp and
manipulation) which make it appropriate for comparisons with the GRASSP and the concepts
that have contribute to it’s design.
167
CAPABILITIES OF UPPER EXTREMITY QUESTIONNAIRE
This questionnaire is designed to find out how well you are able to use your arms
and hands. I will ask you about a number of actions which some people with spinal cord
injury have difficulties or limitations performing. Please consider whether, on an average
day, you have difficulties or limitations performing these actions. By this I mean difficulty
doing the action, or trouble doing it as often as you would like or need in order to complete
everyday activities. Consider only the specific part of your arm or hand asked about in each
question. For example, if asked about pulling something with your arm, do not worry about
whether or not you can grab it with your hand.
Answer each question on a scale of 0 to 4, where 4 is the best – you have no
difficulty or limitation doing the action, and 0 is the worst – you are totally limited and can’t
do it at all.
Score
Description
4 No Difficulty
3 Mild Difficulty
2 Moderate Difficulty
1 Severe Difficulty
0 Unable/ Complete difficulty
168
CUE 2.1 Unable/
Complete Difficulty
Severe Difficulty
Moderate Difficulty
Mild
Difficulty No
Difficulty
THE FOLLOWING QUESTIONS ARE ABOUT
YOUR ABILITY TO REACH OR LIFT
1. Think about reaching out with your arm to touch something directly in front of you that is at shoulder level:
….how difficult is it to do this using your RIGHT ARM 0 1 2 3 4
...how difficult is it to do this using your LEFT ARM? 0 1 2 3 4
2. Think about raising your arm directly over your head, with your arm straight:
...how difficult is it to do this motion using
your RIGHT ARM?0 1 2 3 4
...how difficult is it to do this motion using
your LEFT ARM?0 1 2 3 4
3. Think about reaching down to touch the floor and sitting back up straight, without hooking with your other arm or using it to pull yourself up:
...how difficult is it to do this with your RIGHT HAND? 0 1 2 3 4
...how difficult is it to do this with your LEFT HAND? 0 1 2 3 4
4. Think about raising a 5-pound object like a heavy blanket over your head using both arms. (Don't worry about whether you could grab it with your hands, just if you could raise something
169
that heavy over your head.):
...how difficult is it to do this using BOTH ARMS? 0 1 2 3 4
CUE 2.1 Unable/
Complete Difficulty
Severe Difficulty
Moderate Difficulty
Mild
Difficulty No
Difficulty
THE FOLLOWING QUESTIONS ARE ABOUT YOUR ABILITY TO PULL AND PUSH WITH YOUR
ARMS
5. Think about pulling or sliding (without grasping) a light object such as a can of soda, that is on a table, towards you:
...how difficult is it to do this kind of thing using
your RIGHT ARM?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT ARM?0 1 2 3 4
6. Think about pulling or sliding (without grasping) a heavy object (up to 10 lbs.), that is on a table, towards you:
...how difficult is it to do this kind of thing using
your RIGHT ARM?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT ARM?0 1 2 3 4
7. Think about pushing a light object such as a can of soda on a table, away from you:
...how difficult is it to do this kind of thing using
your RIGHT ARM?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT ARM?0 1 2 3 4
8. Think about pushing a heavy object
170
(up to 10 lbs.) on a table, away from you:
...how difficult is it to do this kind of thing using
your RIGHT ARM?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT ARM?0 1 2 3 4
9. Think about pushing down with both arms into your chair enough to lift your buttocks (both sides) off the seat (do a push-up weight shift):
...how difficult is it to do this? 0 1 2 3 4
CUE 2.1 Unable/
Complete Difficulty
Severe Difficulty
Moderate Difficulty
Mild
Difficulty No
Difficulty
THE FOLLOWING QUESTIONS ARE ABOUT MOVING AND
POSITIONING YOUR ARM AND WRIST
10. With your hand on your lap palm down, think about curling your wrist upwards, keeping your arm on your lap:
...how difficult is it to do this motion using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this motion using
your LEFT HAND?0 1 2 3 4
11. Think about turning your hand over - from your palm facing up to facing the floor, keeping your elbow bent at your side (the arm motion someone would make when turning a doorknob or a dial):
...how difficult is it to do this motion using
your RIGHT ARM?0 1 2 3 4
171
...how difficult is it to do this motion using
your LEFT ARM?0 1 2 3 4
THE FOLLOWING QUESTIONS ARE ABOUT USING YOUR
HANDS AND FINGERS
12. Think about grasping and holding an object like a hammer with your hand:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
13. Think about picking up a small object such as a paper clip or the cap of a tube of toothpaste with the tips of your thumb and first two fingers:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
CUE 2.1 Unable/
Complete Difficulty
Severe Difficulty
Moderate Difficulty
Mild
Difficulty No
Difficulty
14. Think about pinching and holding an object between your thumb and the side of your index finger, such as holding a key:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
15. Think about grasping a large object like the lid of a 2 pound jar of
172
mayonnaise with the tips of the fingers hard enough to pick the jar up or open the lid:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
16. Think about using your fingers to manipulate objects, such as holding a coin and turning it over and over with your fingers:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
17. Think about pressing something with the tip of your index finger (not knuckle) such as dialing a touch-tone phone or ringing a doorbell:
...how difficult is it to do this kind of thing using
your RIGHT HAND?0 1 2 3 4
...how difficult is it to do this kind of thing using
your LEFT HAND?0 1 2 3 4
CUE v2.1 January, 2006
173
B: ISCSCI
The ISCSCI renders a sensory and motor level for each side which is based on the most
normal caudal spinal cord level that is represented by the dermatomes and myotomes tested.
A sensory or motor neurological level is derived when the most caudal side is used to express
the spinal level. A SCI can also be classified according to the ASIA impairment scale as A, B
C, D or E. Also, four different syndromes for SCI can also be defined. And finally a zone of
partial preservation is derived from the partial sensory and motor integrity below the
assigned ISCSCI levels. Reliability is 35% to 93% consistent among raters across sensory
and motor testing (Priebe and Waring, 1991). The ISCSCI was selected for use in the study
to define the sample according to an international classification method and to define the
severity of injury for individuals involved in the study. The ISCSCI is the most widely used
measure to define sensory and motor levels in SCI, therefore, considered by some to be a
“gold standard” in the discrimination of SCI. The ISCSCI was therefore, used as a
comparator to establish validity with GRASSP.
174
175
176
C: SCIM
The SCIM is a global measure of function specific for individuals with SCI (Catz et al.,
2004), used to define the function and independence of the sample in this study. The SCIM is
a scale that has been specifically developed to evaluate the functional outcomes of patients
with traumatic and non-traumatic SCI. The SCIM assesses function in three core areas: 1)
Self-care, which includes feeding, bathing, dressing and grooming, and is scored between a
range of 0 to 20; 2) Respiration and sphincter management are scored between a range of 0 to
40; and lastly 3) Mobility, also scored between a range of 0 to 40 (Catz et al., 2007). Inter
rater reliability is above 0.8 when assessed by agreement statistics for most SCIM items, and
ICC for the total score is 0.94. Concurrent validity of the SCIM with the FIM is 0.79
(Itzkovich et al., 2007). In this analysis the SCIM self care subscale (SCIM-SS) was used as
one of the representations of upper limb function. The SCIM-SS includes items solely related
to the use of the upper limb; therefore, comparisons between the GRASSP subtests are made
with the SCIM-SS, rather than the total SCIM score. Subscales of the SCIM are reliable and
useful quantitative representations of the specific constructs of independence in SCI (Catz et
al., 2007). Some researchers have found the subscales to be more specific for targeted
analyses, such as: the mobility subscale for comparison with other measures of walking and
the selfcare subscale for comparison with other measures of the upper limb (Rudhe and van
Hedel, 2009; van Hedel, 2009).
Function for individuals with tetraplegia is often considered to represent overall
global function. Therefore, the SCIM was used as a comparator between impairment and
function.
177
Spinal Cord Independence Measure
Patient ID Number: _______ Examiner: _________________________ Date: __________ Enter the score for each function in the adjacent square Self-Care 1. Feeding (cutting, opening containers, bringing food to mouth, holding cup with fluid) 0. Needs parenteral, gastrostomy or fully assisted oral feeding 1. Needs partial assistance for eating and/or drinking, or for wearing adaptive devices 2. Eats independently; needs adaptive devices or assistance only for cutting food and/or pouring and/or opening containers 3. Eats and drinks independently; does not require assistance or adaptive devices 2. Bathing (soaping, manipulating water tap, washing). A-upper body; B-lower body A. 0. Requires total assistance 1. Requires partial assistance 2. Washes independently with adaptive devices or in a specific setting 3. Washes independently; does not require adaptive devices or a specific setting B. 0. Requires total assistance 1. Requires partial assistance 2. Washes independently with adaptive devices or in a specific setting 3. Washes independently; does not require adaptive devices or a specific setting 3. Dressing (preparing clothes, dressing, undressing). A-upper body; B-lower body A. 0. Requires total assistance 1. Requires partial assistance with clothes without buttons, zippers or laces (cwobzl) 2. Independent with cwobzl; requires adaptive devices and/or in a specific setting (adss) 3. Independent with cwobzl; does not require adss; requires assistance or adss only for bzl 4. Dresses (any cloth) independently; does not require adaptive devices or specific setting B. 0. Requires total assistance 1. Requires partial assistance with clothes without buttons, zippers or laces (cwobzl) 2. Independent with cwobzl; requires adaptive devices and/or in a specific setting (adss) 3. Independent with cwobzl; does not require adss; requires assistance or adss only for bzl 4. Dresses (any cloth) independently; does not require adaptive devices or specific setting 4. Grooming (washing hands and face, brushing teeth, combing hair, shaving, applying makeup) 0. Requires total assistance 1. Requires partial assistance 2. Grooms independently with adaptive devices 3. Grooms independently without adaptive devices S u b t o t a l (0 - 20) Respiration and Sphincter Management 5. Respiration 0. Requires tracheal tube (TT) and permanent or intermittent assisted ventilation (IAV) 2. Breathes independently TT; requires oxygen, much assistance in coughing or TT management 4. Breathes independently with TT; requires little assistance in coughing and TT management 6. Breathes independently without TT; requires oxygen, much assistance in coughing, a mask (e.g. peep) or IAV (bipap) 8. Breathes independently without TT; requires little assistance or stimulation for coughing 10. Breathes independently without assistance or device 6. Sphincter Management D Bladder 0. Indwelling catheter 3. Residual urine volume (RUV) > 100 cc; no regular catheterization or assisted intermittent catheterization 6. RUV < 100 cc or intermittent self-catheterization; needs assistance for applying drainage instrument 9. Intermittent self-catheterization; uses external drainage instrument; does not need assistance for applying 11. Intermittent self-catheterization; continent between catheterizations; does not use external drainage instrument 13. RUV < 100cc; needs only external urine drainage; no assistance is required for drainage 15. RUV < 100cc; continent; does not use external drainage instrument 7. Sphincter Management D Bowel 0. Irregular timing or very low frequency (less than once in 3 days) of bowel movements 5. Regular timing, but requires assistance (e.g., for applying suppository); rare accidents
178
(less than twice a month) 8. Regular bowel movements, without assistance; rare accidents (less than twice a month) 10. Regular bowel movements, without assistance; no accidents 8. Use of Toilet (perineal hygiene, clothes adjustment before after, use of napkins or diapers) 0. Requires total assistance 1. Requires partial assistance; does not clean self 2. Requires partial assistance; cleans self independently 4. Uses toilet independently in all tasks but needs adaptive devices or special setting (e.g., bars) 5. Uses toilet independently; does not require adaptive devices or special setting (e.g., bars)
S u b t o t a l (0 - 40) Mobility (room and toilet) 9. Mobility in Bed and Action to Prevent Pressure Sores 0. Needs assistance in all activities; turning upper body in bed, turning lower body in bed, sitting up in bed, doing push-ups in wheelchair, with or without adaptive devices, but not with electric aids 2. Performs one the activities without assistance 4. Performs two or three of the activities without assistance 6. Performs all the bed mobility and pressure release activities 10. Transfers: bed-wheelchair (locking wheelchair, lifting footrests, removing and adjusting arm rests, transferring, lifting feet) 0. Requires total assistance 1. Needs partial assistance and/or supervision and/or adaptive devices (e.g., sliding board) 2. Independent (or does not require wheelchair) 11. Transfers: wheelchair-toilet-tub (if uses toilet wheelchair; transfers to and from; if uses regular wheelchair; locking wheelchair, lifting footrests, removing and adjusting armrests, transferring, lifting feet) 0. Requires total assistance 1. Needs partial assistance and/or supervision, and/or adaptive devices (e.g., grab-bars) 2. Independent (or does not require wheelchair) Mobility (indoors and outdoors) 12. Mobility Indoors 0. Requires total assistance 1. Needs electric wheelchair or partial assistance to operate manual wheelchair 2. Moves independently in manual wheelchair 3. Requires supervision while walking (with or without devices) 4. Walks with a walking frame or crutches (swing) 5. Walks with crutches or two canes (reciprocal walking) 6. Walks with one cane 7. Needs leg orthosis only 8. Walks without aids 13. Mobility for Moderate Distances (10 - 100 metres) 0. Requires total assistance 1. Needs electric wheelchair or partial assistance to operate manual wheelchair 2. Moves independently in manual wheelchair 3. Requires supervision while walking (with or without devices) 4. Walks with a walking frame or crutches (swing) 5. Walks with crutches or two canes (reciprocal walking) 6. Walks with one cane 7. Needs leg orthosis only 8. Walks without aids 14. Mobility Outdoors (more than 100 meters) 0. Requires total assistance 1. Needs electric wheelchair or partial assistance to operate manual wheelchair 2. Moves independently in manual wheelchair 3. Requires supervision while walking (with or without devices) 4. Walks with a walking frame or crutches (swing) 5. Walks with crutches or two canes (reciprocal walking) 6. Walks with one cane 7. Needs leg orthosis only 8. Walks without aids 15. Stair Management 0. Unable to climb or descend stairs 1. Climbs and descends at least 3 steps with support or supervision of another person 2. Climbs and descends at least 3 steps with support of handrail and/or crutch or cane 3. Climbs and descends at least 3 steps without any support or supervision
179
16. Transfers: wheelchair-car (approaching car, locking wheelchair, removing arm and foot rests, transferring to and from car, bringing wheelchair into and out of car) 0. Requires total assistance 1. Needs partial assistance and/or supervision and/or adaptive devices 2. Transfers independent; does not require adaptive devices (or does not require wheelchair) 17. Transfers: ground-wheelchair 0. Requires assistance 1. Transfers independent with or without adaptive devices (or does not require wheelchair)
S u b t o t a l (0± 40)
TOTAL SCIM SCORE (0± 100)
180
Appendix 5 A: Guttman Scaling Procedure
B: Matrices used to rank-order all items within subtests
181
Guttman Scaling: Guttman scaling is a procedure used to determine whether a set of items can be rank-ordered on a unidimensional scale. Guttman scaling is also sometimes known as cumulative scaling or scalogram analysis. The purpose of Guttman scaling is to establish a one-dimensional continuum for a concept you wish to measure (Guttman 1944; Guttman 1950, Trochim, 2010). In the case of the items within subtests of GRASSP the order was defined by sensory and motor spinal cord levels moving rostral to caudal. The steps that were conducted to perform the manual Guttman scaling are as follow:
1. First the items for each subtest are reorganized to best represent spinal nerve root order example moving left to right items would be ordered to represent C6, C7 and C8 and so on and so forth. In the case of prehension where myotomes and dermatomes are not so clearly defined, innervation of muscles for specified grips were used to determine order of items.
2. Each set of ratings for all items would then be ordered according to responses, more and greater responses to each item were moved towards the bottom of the list and fewer and lower responses moved to the top. Main point is to look for a cumulative order, for example that if item #2 is scored at a 3 that item #1 would be scored at a 3 or more and so and so forth for all items.
3. In the strength scores, 0 and 1 were collapsed and 4 and 5 were collapsed. In Prehension performance 4 and 5 were collapsed.
4. Any set of ratings that did not show a cumulative item response pattern were placed in the exceptions and were not considered to show a cumulative predictive item response pattern.
5. Since a large part of the sample did show a cumulative predictive item response pattern it was concluded that items could be summated for subtest total scores.
182
Guttman Matrices of Dorsal Sensation before and after ordering: Item Item
Table 1 illustrates the Dorsal Sensation, items and total scores for the whole sample. C6 represents Test Location 1, C7 represents Test Location 2, and C8 represents Test Location 3. The table on the left side presents the data as it was entered and the table on the right side presents the data ordered by cumulative predictability of the item scores, in descending order of less impairment. The bottom shaded group of data is the individuals with unpredictable item responses.
183
Guttman Matrices of Palmar Sensation before and after ordering: Item Item
Table 2 illustrates the Palmar Sensation items and total scores for the whole sample. C6 represents Test Location 4, C7 represents Test Location 5 and C8 represents Test Location 6. The table on the left side presents the data as it was entered and the table on the right side presents the data ordered by cumulative predictability of the item scores, in descending order of less impairment. The shaded group of data is the individuals with unpredictable item responses.
184
Guttman Matrices of Strength before and after ordering: Item Item
Table 3 illustrates the Strength items and total scores for the whole sample. M1 represents Anterior Deltoid (C5), M2 represents Biceps (C5), M3 represents Triceps (C7), M4 represents Wrist Extensors (C6), M5 represents Extensor Digitorum (C8), M6 represents Opponens Policis (C7), M7 represents Finger Flexors (C8), M9 represents Finger Abductor (DV) (T1) and M10 represents Dorsal Interossei (T1). Items are ordered according to myotomal order. The table on the left side presents the data entered and the table on the right side presents the sample ordered by cumulative predictability of the item scores, in descending order of less impairment. The shaded group of data is individuals with unpredictable item responses.
185
Guttman Matrices of Prehension Ability before and after ordering: Item Item
Table 4 illustrates the Prehension Ability items and total scores ordered according to muscle innervation which represent myotomal order. Grasp 1 represents cylindrical grasp, Grasp 2 represents lateral key pinch and Grasp 3 represents tip to tip pinch. The table on the left side presents the data as entered and the table on the right side presents the sample ordered by cumulative predictability of the item scores, in descending order of less impairment. The shaded group of the sample are the individuals with unpredictable item responses.
186
Guttman Matrices of Prehension Performance before and after ordering: Item Item
Table 5 illustrates the Prehension Performance items ordered according to muscle innervation that represents myotomal order. Task 1 through 6 represent the tasks defined in the GRASSP. The table on the left side presents the data as entered and the table on the right side presents the sample ordered by cumulative predictability of the item scores, in descending order of less impairment. The shaded group of the sample are the individuals with unpredictable item response.