GMFM (GMFM-66 & GMFM-88) User's Manual, 2nd edition

By Dianne Russell, Peter L Rosenbaum and Marilyn Wright

The Gross Motor Function Measure (GMFM) has become the best evaluative measure of motor function designed for quantifying change in the gross motor abilities of children with cerebral palsy. The new version of the scoring programme has now been released, and includes two abbreviated methods of estimating GMFM-66 scores using the GMFM-66-Item sets and the GMFM-66-Basal & Ceiling.

• Language: English
• Publisher: Mac Keith Press
• Release date: Oct 7, 2013
• ISBN: 9781908316882

Dianne Russell

Dianne J Russell is Research and Knowledge Exchange Specialist with CanChild Centre for Childhood Disability Research. She has been a health services researcher for over 20 years and a key individual in the development, evaluation and dissemination of clinical outcome measures such as the Gross Motor Function Measure (GMFM). Her current focus is on facilitating the use of research evidence in practice by engaging with research users throughout the research process and by making research results easily accessible in multiple formats to families, service providers, and policy decision-makers.

Book preview

Clinics in Developmental Medicine

GROSS MOTOR FUNCTION MEASURE

(GMFM-66 & GMFM-88)

USER’S MANUAL

SECOND EDITION

Clinics in Developmental Medicine

Gross Motor Function Measure (GMFM-66 & GMFM-88) User’s Manual

Second Edition

DIANNE J RUSSELL

PETER L ROSENBAUM

MARILYN WRIGHT

LISA M AVERY

CanChild Centre for Childhood Disability Research

McMaster University

Hamilton, Ontario

Canada

2013

Mac Keith Press

© 2013 McMaster University

Editor: Hilary M Hart

Managing Director: Ann-Marie Halligan

Production Manager: Udoka Ohuonu

Project Manager: Ushadevi Medapati

The views and opinions expressed herein are those of the authors and do not necessarily represent those of the publisher.

Dianne J Russell, Peter L Rosenbaum, Marilyn Wright, Lisa M Avery, Niina Kolehmainen, Mary Lane, Doreen Bartlett and Laura Brunton have asserted their rights under the Canadian Copyright Act to be identified as the authors of this work.

First published 2002.

Reprinted 2011.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN 978-1-908316-88-2

Typeset by Datamatics Global Services Limited

Printed by Latimer Trend and Company, Plymouth, Devon, UK

Mac Keith Press is supported by Scope

CONTENTS

AUTHORS’ APPOINTMENTS

PREFACE TO THE SECOND EDITION

PREFACE TO THE FIRST EDITION

ACKNOWLEDGEMENTS TO THE FIRST EDITION

ACKNOWLEDGEMENTS TO THE SECOND EDITION

DISCLAIMER

1. OVERVIEW OF THE GROSS MOTOR FUNCTION MEASURE (GMFM)

2. CONCEPTUAL BACKGROUND

3. DEVELOPMENT AND VALIDATION OF THE GMFM-88

with Niina Kolehmainen

4. DEVELOPMENT AND VALIDATION OF THE GMFM-66

with Niina Kolehmainen

5. GMFM-66: TWO SHORT FORMS OF ADMINISTRATION

6. ADMINISTRATION AND SCORING GUIDELINES FOR THE GMFM-88 AND GMFM-66

with Mary Lane

7. INTERPRETATION AND USES OF THE GMFM-88 AND GMFM-66

8. APPLICATIONS OF THE GMFM: WHAT HAVE WE LEARNED AND WHERE CAN WE GO NEXT?

REFERENCES

GLOSSARY OF TERMS

APPENDIX 1.METHODS OF DISPLAYING ITEM DIFFICULTY

APPENDIX 2.DISPLAY OF ITEM DIFFICULTIES USING THURSTONE THRESHOLDS

APPENDIX 3.GROSS MOTOR ABILITY ESTIMATOR-2 (GMAE-2) SCORING PROGRAM TUTORIAL FOR THE GMFM-66

APPENDIX 4.CASE SCENARIOS OF TWO CHILDREN WHO MISFIT THE GMFM-66 ITEM DIFFICULTY MODEL

APPENDIX 5.GMFM-66 ITEM SET SCORE SHEET

APPENDIX 6.GMFM-66 BASAL & CEILING SCORE SHEET

APPENDIX 7.GROSS MOTOR FUNCTION CLASSIFICATION SYSTEM-EXPANDED & REVISED (GMFCS-E&R)

APPENDIX 8.GMFM-88 AND GMFM-66 SCORE SHEET

APPENDIX 9.GMFM-66 AND GMFM-88 CROSS-SECTIONAL AND CHANGE SCORES

APPENDIX 10. STANDARD ERROR OF MEASUREMENT

APPENDIX 11. CASE SCENARIO OF TREVOR

APPENDIX 12. ADDITIONAL EXAMPLES OF THE USE OF THE MOTOR MEASURES WITH CHILDREN WITH CEREBRAL PALSY IN PHYSICAL THERAPY PRACTICE 270

Doreen J Bartlett and Laura K Brunton

APPENDIX 13. ONTARIO MOTOR GROWTH CURVES

INDEX

AUTHORS’ APPOINTMENTS

Dianne J Russell

Research and Knowledge Exchange Consultant with CanChild Centre for Childhood Disability Research; Associate Clinical Professor, School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada

Peter L Rosenbaum

Professor of Pediatrics, Canada Research Chair in Childhood Disability; Co-Founder, CanChild Centre for Childhood Disability Research, McMaster University, Hamilton, Ontario, Canada

Marilyn Wright

Physiotherapist, McMaster Children’s Hospital; Assistant Clinical Professor, McMaster University, Hamilton, Ontario, Canada

Lisa M Avery

Statistical Consultant, Avery Information Systems, Orillia, Ontario, Canada

CONTRIBUTORS

Niina Kolehmainen

MRC Population Health Scientist, Health Services Research Unit, University of Aberdeen, Scotland, UK

Doreen Bartlett

Professor, School of Physical Therapy, Western University, London, Ontario; Scientist, CanChild Centre for Childhood Disability Research, McMaster University, Hamilton, Ontario, Canada

Laura Brunton

PhD Candidate in Health and Rehabilitation Sciences Program, Western University, London, CanChild Centre for Childhood Disability Research, McMaster University, Hamilton, Ontario, Canada

Mary Lane

Pediatric Physiotherapist (Retired), Kingston, Ontario, Canada

PREFACE TO THE SECOND EDITION

As we noted in the Preface to the First Edition of the Gross Motor Function Measure (GMFM-66 & GMFM-88) User’s Manual (Russell et al. 2002), the work we reported owed a considerable debt of gratitude to many people (named in the acknowledgements to that First Edition). This is equally true as we complete this Second Edition. We particularly want to recognize the many insightful colleagues, supportive friends, thoughtful critics, and users from all over the world who have communicated with us about the GMFM. It has been their (your!) observations and questions to us over the past decade that have kept us honest in considering how to improve the measure and to make it as accessible as possible for clinical and research colleagues everywhere.

This updated version of the User’s Manual describes both our own further development and applications of the measure and the work of colleagues who have kindly shared their experiences with us and with you. Thus, for example, there are reports here about two shorter approaches to administering and scoring the GMFM-66, the Item Set approach by Russell et al. (2010), and the Basal & Ceiling approach of Brunton and Bartlett (2011). A paper in Developmental Medicine and Child Neurology by Avery et al. (2013) contrasted the ‘performance’ of these shorter versions against the full GMFM-66 using a common data set, and the findings are reported in detail in Chapter 5 of this book, with the kind permission of Mac Keith Press.

We try wherever possible to share our ideas, experience and opinions when asked, and hope that people can benefit from both our experience and from the lessons we have learned during this long and fascinating journey. Any changes to the measure need to be considered carefully and should always undergo rigorous reliability and validity studies to determine whether the measurement properties remain intact. We have briefly updated what we have learned about the use of the GMFM-88 and GMFM-66 from the literature; however, we encourage readers to do their own critical evaluation of the research as evidence is continually emerging.

One major change that we trust will represent a significant improvement is the creation of the Gross Motor Ability Estimator-2 (GMAE-2) data management and scoring program. The original GMAE was developed to provide GMFM users with an easy method to enter and score assessment data from either the 88- or 66-item version of the measure. It worked well in the early years after the manual was published. However, as computer software became more sophisticated many of the newer systems could no longer read the old GMAE. Efforts to keep up with technology by developing patches proved challenging and eventually we recognized that this was a losing proposition.

The solution we have arrived at, that we hope will work in the long term, was to redevelop the program as GMAE-2. Doing things this way will allow us to make adjustments to the program relatively easily, and hopefully to keep up with developments in computing software so that people using the program will not experience the frustrations we know have happened for reasons beyond our control. The new GMAE-2 will allow users to (1) download score sheets for the GMFM-88, GMFM-66, GMFM-66-IS (Item Sets) and GMFM-66-B&C (Basal & Ceiling); (2) enter and calculate scores for the GMFM-88 and the GMFM-66 using several formats including GMFM-66-IS and GMFM-66-B&C items; (3) import data from the original GMAE program; and (4) plot a child’s GMFM-66 assessment on the centile curves appropriate for the child’s GMFCS-E & R levels. It is important to have the manual in addition to the software to ensure that users are knowledgeable about which version of the GMFM to use and how to administer, score and interpret the measure.

This second edition has been a long time in development. We hope that it will provide interesting information for all readers, and be a useful improvement for long-time users. For people new to the GMFM, we trust that the ideas in this book will be accessible and applicable to your work. As always, the authors are interested in hearing from those who have ideas for us, or questions or concerns.

Online Resources

The GMAE-2 can be downloaded from the CanChild website (http://motorgrowth.canchild.ca/en/GMFM/gmae.asp. The score sheets for the GMFM-88, GMFM-66, GMFM-66-IS (item sets) and the GMFM-66-B&C (Basal & Ceiling) can be downloaded from http://motorgrowth.canchild.ca/en/GMFM/gmfmscoresheets.asp.

PREFACE TO THE FIRST EDITION

After a great deal of work by many people, a paper describing the validation of the Gross Motor Function Measure (GMFM) was published in Developmental Medicine and Child Neurology in June 1989. The first edition of the GMFM manual was published by the authors in 1990 and subsequently revised in 1993 to include improved administration and scoring guidelines along with updated information on research done with the GMFM. At the time of the second edition of the manual we identified that future work would be aimed at applying a relatively new methodology for constructing measures, called Rasch analysis, to determine its usefulness for the GMFM. Rasch analysis promised many advantages that would improve the scaling and interpretation of the GMFM provided the assumptions of the Rasch model could be met.

This new manual reflects the work done to improve the original 88-item GMFM which for purposes of clarity we will refer to in this manual as the GMFM-88. This has been accomplished by (1) reducing the number of items to 66; (2) converting the ordinal scale to a scale with interval properties; (3) determining the hierarchical structure of the items, making it possible to create an item map which illustrates how difficult each item is for children with cerebral palsy relative to the other items; and (4) developing a computer program to score the GMFM-66.

Many decisions needed to be made when applying the Rasch methodology to the GMFM. There were sometimes trade-offs between clinical feasibility and rigorous methodology. With the GMFM being used widely as an outcome measure for many intervention trials, we felt that the over-riding decision was always to make the GMFM the most rigorous measure possible for research purposes. This led us to decide against a table of scores in the manual but rather to apply a more precise scoring system using a computer program. We recognized that this might limit the clinical applicability of the GMFM-66 for some people, but would ultimately provide more accurate information about changes in gross motor function in children with cerebral palsy. In the light of these decisions, we have tried to maintain or increase the clinical usefulness as much as possible but have left the original 88-item version for those who prefer it or have difficulty accessing the scoring program.

ACKNOWLEDGEMENTS TO THE FIRST EDITION

The first Gross Motor Function Measure (GMFM) manual was in fact the third (and we hope much improved) version of work that had been ongoing since the late 1980s. The original manual was prepared in 1990 and a second edition of that manual appeared in 1993. In both versions of the manual we acknowledged people whose efforts had contributed to the work. These people included, of course, the authors of the original publication describing the GMFM (Carolyn Gowland, Susan Hardy, Nancy Plews, Heather McGavin, David Cadman and Sheila Jarvis). We wish to thank Kate O’Connor Steel who shared with us her Motor Control Assessment where many of the items for the GMFM originated. In addition, we recognized clinical colleagues and a number of students and support staff who had participated in the creation of the work then available. They number too many to cite individually, but we are very grateful that they supported our early efforts to create a new clinical assessment tool.

A number of McMaster University faculty colleagues provided signal input into the early conception of the GMFM through either conceptual or technical support. They include Drs Charlie Goldsmith, Gordon Guyatt, David Streiner and Christal Woodward, all acknowledged experts in clinical epidemiology and measurement development. Each gave freely of their time and ideas as the GMFM took shape in the mid-1980s. The recent Rasch analysis work would not have been possible without the statistical and methodological expertise of Drs Stephen Walter and Parminder Raina, who spent many meetings poring over Rasch outputs and discussing issues as we worked our way through this relatively new approach to measurement. Dr Bob Palisano provided invaluable advice and feedback on many aspects of this work, as did Dr Steve Hanna.

A project to create and validate a clinical measure requires the efforts of literally dozens of dedicated clinical colleagues who toil in the trenches to collect the raw data that are shaped by the research team. These people—mainly clinical physiotherapists in the Children’s Treatment Centres associated with the Ontario Association of Children’s Rehabilitation Services (OACRS)—have participated in the Ontario Motor Growth Curve study since 1996. Between them they have done over 3000 GMFMs that provided the basic materials for the work reported here! They participated in the GMFM training program and tolerated repeated reassessments of their reliability with patience and enthusiasm. We are forever in their debt, because this work could clearly not have been done without them. We also recognise that clinical research is an ‘add-on’ for many OACRS centres, and thank the Executive Directors for their partnership in research ventures like this one.

CanChild Centre for Childhood Disability Research has been supported since 1989 by the Health System-Linked Research Units program of the Research Branch of the Ontario Ministry of Health and Long-Term Care. Work such as that reported here would be difficult if not impossible without the continuing support of this program, for which we are most grateful. The research data collection activities of the therapists were made possible through the generous grant support of both the Medical Research Council of Canada (now the Canadian Institutes of Health Research) and the National Center for Medical Rehabilitation Research of the National Institute of Child Health and Human Development at NIH (grant R01-HD-34947). Barbara Galuppi coordinated all aspects of this multi-site 5-year study with admirable skill and efficiency, and in this way has made an enormous contribution to all the work of our research group. Of course were it not for the unique contribution of time and effort of hundreds of children with cerebral palsy (CP) and their families, completing questionnaires and permitting videotaping of their activities, this work would have been impossible.

Colleagues from around the world have participated in GMFM training workshops and helped us in our work through a combination of perceptive suggestions, insightful questions, the use of the GMFM in their clinical and research efforts, and generous sharing of their experiences and often their data. At the risk of missing some people, we wish to recognize in particular Kristie Bjornson at Children’s Hospital and Regional Medical Centre, Seattle and Marjolijn Ketelaar at Utrecht University. Dr Suzann Campbell at the University of Illinois at Chicago and Dr Steve Haley of Boston University were very helpful in the early phases of our project, sharing perspectives from their Rasch analysis experiences with the Test of Infant Motor Performance (TIMP) and the Pediatric Evaluation of Disability Inventory (PEDI), respectively. The work reported here, thus, has a truly international flavour, and we appreciate the support of these friends.

Having created a Gross Motor Ability Estimator (GMAE) computer program to score the GMFM-66, we wished to assess whether the program could be applied and interpreted by clinical therapists outside of a research environment. We are indebted to the many therapists in OACRS programmes who helped to pilot test the several versions of the program as we worked to make it accessible. Drs Doreen Bartlett, Lisa Rivard, Marilyn Wright and Virginia Wright were particularly helpful in offering insightful feedback, which we have used to improve the program.

Thanks to Eric Bosch and Graham Passmore for their programming expertise in turning our GMFM training videotapes into the interactive GMFM CD-ROM teaching tool that is now available for self-training in the use of the GMFM.

We wish finally to acknowledge the encouragement and help we receive every day from our colleagues at CanChild, who have tolerated our apparently endless efforts to complete this work, perhaps to the neglect of other responsibilities! Particular thanks go to Betsy Spencer, Pat Abernathy and Kamal Mangat who have been unofficial members of the Motor Measures Group, and whose behind-the-scenes contributions we know better than most people.

At the end of the day, of course, with all the help and support we have received from people named and unnamed, any shortcomings in this work are the responsibilities of the authors.

ACKNOWLEDGMENTS TO THE SECOND EDITION

For this second edition of the manual we would like to thank Mary Lane, who wrote the original administration and scoring guidelines for the previous edition, for her continued support and sage advice.

Justin Turco-Gwozdowski is a software engineer who, as a recent university graduate, took on the programming of the new GMAE as a volunteer experience. His patience in addressing each programming challenge is much appreciated!

Special thanks to Bob Palisano and Virginia Wright for their thoughtful comments on Chapter 8.

A great many therapists helped Justin and the authors by pilot testing successive versions of the GMAE-2 and provided feedback to help improve it. We are indebted to them for their thoughtful feedback.

Finally, our CanChild colleagues and our families have put up with the time demands that have distracted us from other work and responsibilities! We cannot thank them enough.

DISCLAIMER

The Gross Motor Function Measure (GMFM) is a clinical assessment tool developed to evaluate motor function, and change in the function over time, in children and youth with cerebral palsy (CP). All contents, materials and scoring programs provided in this text [Gross Motor Function Measure (GMFM-66 and GMFM-88) User’s Manual 2nd Edition] are provided ‘as is’ without warranty of any kind. McMaster University and the publisher make no warranties, express or implied, that any of the contents or materials provided in this book are free of error, or are consistent with any particular standard of merchantability, or that they will meet people’s requirements for any particular application. The authors and publisher disclaim all liability for direct, incidental or consequential damages resulting from your use of this book.

1

OVERVIEW OF THE GROSS MOTOR FUNCTION MEASURE (GMFM)

What is the GMFM?

The Gross Motor Function Measure (GMFM) is a clinical measure designed and validated to evaluate changes in the gross motor function in children with cerebral palsy (CP).

There are two versions of the GMFM: the original 88-item measure (GMFM-88) and the 66-item measure (GMFM-66). These items are administered in the same way—the difference between the two versions simply concerns which of the items (from the full pool of 88 possible items) are included in GMFM-66.

There are three methods of administering the GMFM to obtain a GMFM-66 score. All three methods require the Gross Motor Ability Estimator (GMAE-2) scoring software to convert the item data into a GMFM-66 score. Administration methods include the following: (1) GMFM-66: administer all 66 items; (2) GMFM-66-Item Sets (GMFM-66-IS): use a scoring algorithm to identify a subset of items to administer; and (3) GMFM-66 Basal & Ceiling (GMFM-66-B&C): use a basal/ceiling approach to identify a subset of items.

Items on the GMFM-88 span aspects of what the World Health Organization’s International Classification of Functioning, Disability and Health refers to as gross motor ‘activity’, ranging from activities such as lying and rolling to walking, running and jumping skills. The GMFM-66 consists of a subset of the 88 items that has been shown to be unidimensional.

Which version of the GMFM should I use?

The GMFM version you should use depends upon the purpose of the assessment and the population. The GMFM-88 could be used when you want a detailed description of a child’s current motor abilities, for which more items may be useful. This version might be the best choice for children with CP who are very young (functioning primarily in prone and supine) or are functioning in the Gross Motor Function Classification System (GMFCS) Level V. The GMFM-88 must be used if the child tested is wearing shoes or orthoses or using a mobility aid. This is the most suitable version (with appropriate validation) for children who may have motor difficulties resulting from other motor impairments (e.g. Down syndrome and acquired brain injury [ABI]).

The GMFM-66 provides a GMFM score for children with CP by using fewer items. Because it was based specifically and exclusively on patterns of motor development observed in a large population of children with CP, and because the items are weighted according to difficulty (given the interval properties of the GMFM-66), it is a more valid measure to assess changes over time in children with CP. It is not recommended to be used for children other than those with CP, as the item weights are likely to vary by population.

If there is a need for a briefer method to obtain a GMFM-66 score, then either the GMFM-66-IS or the GMFM-66-B&C approach can be used. However, the GMFM-66-IS is preferred over the GMFM-66-B&C for children with unilateral CP when assessing changes. These brief methods are discussed in detail in Chapter 5.

Whatever version one chooses to administer, the same version should be applied consistently over time to a particular child. It is strongly recommended not to alternate between methods, as this may increase the variability in scores and make the detection of ‘real’ change more difficult.

How is the GMFM administered?

The GMFM requires the child to demonstrate various gross motor skills, as outlined in the GMFM administration and scoring guidelines in Chapter 6.

Prior to administering the GMFM, it is important to know which administration method is being used and to ensure that the appropriate score sheets, administration and scoring guidelines, and equipment are available.

Who is the GMFM appropriate for?

While the measure was designed and validated for children with CP, there is evidence that the GMFM-88 version of the measure is also valid for use with children with Down syndrome and ABI.

Because the GMFM samples motor skills that are typical of normal developmental milestones, it may be useful to assess current motor activities for children other than for those with whom it has been validated; however, because the detailed measurement properties of the GMFM have not been established with other populations than children with CP, Down syndrome or ABI, the reliability and validity of the GMFM-88 should be established prior to using it with other groups of children.

The original GMFM validation sample included children from 5 months to 16 years old. The GMFM would be appropriate for children whose motor skills were at or below those of a 5-year-old child without any motor disability.

How is the GMFM scored?

There is a common 4-point scoring system for each item in the GMFM. Specific descriptors for each item are detailed in the administration and scoring guidelines for the GMFM-66 and GMFM-88 (see Chapter 6). The item scoring is the same for the GMFM-88 and GMFM-66. It is not recommended to report changes in individual item scores as these are not as reliable as total scores.

The GMFM-88 item scores can be summed to calculate raw and percent scores for each of the five GMFM dimensions and a total GMFM-88 score. The five dimensions are A: Lying and Rolling; B: Sitting; C: Crawling and Kneeling; D: Standing; and E: Walking, Running and Jumping. Goal total scores have been used to target goal areas and may be useful when assessing children when the GMFM-66 is not appropriate (e.g. assessing children using aids/orthoses). However, with the development of the GMFM-66, the need for goal areas is no longer necessary for children with CP who are tested barefoot.

The GMFM-66 requires a user-friendly computer program (called the Gross Motor Ability Estimator [GMAE]) to enter individual item scores and convert them to an interval-level total score. The original GMAE program was updated in 2013 with the GMAE-2 scoring software available for downloading from the CanChild website (www.motorgrowth.canchild.ca/en/GMFM/gmae.asp). This URL will also give the user access to all the required score sheets, the GMFCS and the GMFCS Expanded and Revised version.

New features of GMAE-2 allow the user to import data from the original GMAE software, download score sheets, and enter and calculate scores for the GMFM-88, GMFM-66, GMFM-66-IS and GMFM-66-B&C. It is also possible to plot GMFM-66 scores on centile curves by the GMFCS level and export data to csv files.

How long does it take?

Administering the GMFM-88 may take 45 to 60 minutes for someone familiar with the measure, depending on the skill of the assessor, the ability level of the child, and the child’s level of cooperation and understanding.

The GMFM-66 takes less time to administer, as there are fewer items. The GMFM-66-IS and the GMFM-66-B&C take on an average 20 to 30 minutes to administer.

Where should it be administered?

The GMFM should be administered in an environment that is comfortable for the child and is large enough to hold the necessary equipment and allow the child to move freely (e.g. one item requires the child to run 4.5m [15ft] and return). The floor should be a smooth, firm surface.

Because the GMFM was designed to measure changes over time, it is important to keep the environment and assessment conditions as consistent as possible for each successive assessment of the same child.

What equipment is needed to administer the GMFM?

The equipment required is described in detail Chapter 6. Most of what is needed is standard equipment in a physiotherapy gym (e.g. mat, bench, toys). Access to stairs (with at least five steps) is also necessary.

What qualifications are required to administer and score the GMFM?

The GMFM was designed for use by pediatric therapists who are familiar with assessing motor skills in children.

Users should familiarize themselves with the GMFM guidelines and score sheet before assessing children. It may be helpful to practise on several children with and without motor disabilities prior to using it for clinical assessments.

It is recommended that users assess their reliability with the GMFM prior to using it.

2

CONCEPTUAL BACKGROUND

Cerebral palsy

Cerebral palsy (CP) was first described in the medical literature about 150