Gross Motor Function Measure

Abstract

The Gross Motor Function Measure (GMFM) is a criterion-referenced observational clinical assessment tool designed to evaluate longitudinal changes in gross motor function in children with cerebral palsy (CP), as well as individuals with Down syndrome, pediatric acquired brain injury, and spinal muscular atrophy. Developed by Dianne J. Russell, Peter L. Rosenbaum, and their colleagues at the CanChild Centre for Childhood Disability Research at McMaster University, the original 88-item iteration (GMFM-88) evaluates motor capacity across five ontogenetic dimensions: Dimension A: Lying and Rolling (17 items); Dimension B: Sitting (20 items); Dimension C: Crawling and Kneeling (14 items); Dimension D: Standing (20 items); and Dimension E: Walking, Running and Jumping (17 items). Each item is scored on a four-point ordinal scale ranging from 0 (does not initiate) to 3 (completes task), with an option to record “Not Tested” (NT). Psychometric evaluations demonstrate exceptional inter-rater reliability (intraclass correlation coefficients [ICC] typically exceeding .98) and test-retest stability (ICC ≥ .95). Unlike norm-referenced motor developmental tests that evaluate chronological age equivalents, the GMFM evaluates the magnitude of motor change achieved regardless of developmental delays, demonstrating profound sensitivity to change and evaluative responsiveness following interventions such as physical therapy, selective dorsal rhizotomy, intrathecal baclofen pump implantation, botulinum neurotoxin-A administration, and orthopedic reconstructions. Subsequent psychometric refinement utilizing Rasch item response theory led to the GMFM-66, which establishes a unidimensional, interval-scaled hierarchy of motor difficulty. The GMFM remains the undisputed gold standard outcome measurement instrument in pediatric neurorehabilitation worldwide.

Keywords

Gross Motor Function Measure, GMFM-88, GMFM-66, cerebral palsy, pediatric neurorehabilitation, gross motor capacity, Rasch analysis, psychometrics, outcome measurement, motor development, physical therapy, CanChild

Authors

The Gross Motor Function Measure was originally conceived, developed, and standardized by an interdisciplinary consortium of researchers and clinicians at McMaster University in Hamilton, Ontario, Canada, affiliated with CanChild Centre for Childhood Disability Research:

• Dianne J. Russell, MSc – CanChild Centre for Childhood Disability Research, School of Rehabilitation Science, McMaster University.
• Peter L. Rosenbaum, MD, FRCPC – Professor of Pediatrics, Co-Founder of CanChild, McMaster University.
• Lisa M. Avery, MSc – Avery Information Services; Research Biostatistician, CanChild, McMaster University.
• Mary Lane, PT – Pediatric Physiotherapist and Clinical Researcher, McMaster University.
• John Gowland, MSc – Methodologist, McMaster University.
• David L. Streiner, PhD – Department of Psychiatry and Behavioural Neurosciences, McMaster University.
• Stephen D. Walter, PhD – Department of Health Research Methods, Evidence, and Impact, McMaster University.

The official Dutch cross-cultural adaptation and psychometric validation of the GMFM (GMFM Dutch Version) was directed by:

• Marjolijn Ketelaar, PhD – Center of Excellence for Rehabilitation Medicine, UMC Utrecht Brain Center, University Medical Center Utrecht, and De Hoogstraat Rehabilitation, Utrecht, the Netherlands.
• Adri Vermeer, PhD – Department of Educational Sciences, Utrecht University, Utrecht, the Netherlands.
• Paul J. M. Helders, PhD, PT – Department of Pediatric Physical Therapy, Wilhelmina Children’s Hospital, University Medical Center Utrecht, Utrecht, the Netherlands.

Purpose

The Gross Motor Function Measure was engineered specifically to bridge an urgent psychometric and clinical gap in pediatric neurology and neurorehabilitation. Historically, pediatric clinicians evaluated motor progress using norm-referenced developmental indices, such as the Bayley Scales of Infant and Toddler Development or the Gesell Developmental Schedules. While norm-referenced tools effectively identify developmental delays by comparing an infant’s or child’s trajectory against typically developing populations, they fundamentally fail when deployed as evaluative measures for children with non-progressive neuromuscular disorders like cerebral palsy. Because children with cerebral palsy acquire gross motor skills at an altered rate, plateau at different stages, and exhibit qualitatively distinct movement kinematics, comparing them to unimpaired peers invariably misclassifies functional clinical improvements as regression or developmental divergence. Clinicians required an evaluative measure that could document longitudinal change within the child across chronological time, irrespective of whether the child’s movement patterns mimicked typical developmental milestones.

The primary clinical purpose of the GMFM is to measure how much of a gross motor activity a child can perform (motor capacity) rather than how well or how typically the movement is executed (movement quality). The instrument evaluates performance across the entire spectrum of antigravity postures and mobility domains, beginning with baseline head alignment and supine or prone postural control, advancing through independent floor transitions and sitting balance, and culminating in advanced dynamic equilibrium, reciprocal stair navigation, hopping, and sprinting. The clinical utility of the GMFM extends across diagnostic planning, goal setting, longitudinal outcome surveillance, and empirical assessment in comparative effectiveness trials.

In contemporary pediatric practice, the GMFM serves as the universally accepted baseline and post-intervention endpoint in studies examining pharmacological, surgical, and therapeutic regimens. It assesses the precise motor gains induced by botulinum toxin type A injections for focal spasticity, selective dorsal rhizotomy for lower limb spasticity reduction, orthopedic corrections (e.g., femoral derotation osteotomies, tendon lengthenings, foot stabilization), intrathecal baclofen for dyskinesia and generalized hypertonia, and intensive targeted neurorehabilitation (e.g., constraint-induced movement therapy, body-weight-supported treadmill training, functional electrical stimulation). In parallel, the GMFM provides parents, physical therapists, and interdisciplinary healthcare teams with a criterion-based roadmap to formulate realistic, incremental, and functionally meaningful gross motor goals, mapped systematically to the International Classification of Functioning, Disability and Health (ICF) framework under the domain of “Activities.”

Psychological Construct

The central construct quantified by the GMFM is gross motor capacity within a standardized testing environment. Psychometrically, gross motor capacity must be distinguished from motor performance in habitual, everyday contexts. In accordance with the World Health Organization’s ICF model, capacity captures the highest probable level of functional functioning an individual can execute in a standardized, controlled “clinical laboratory” setting, whereas performance captures what the individual spontaneously executes in their ambient, unstandardized daily living environment (such as at home, school, or the playground). The GMFM focuses exclusively on capacity: the child is prompted and encouraged to demonstrate their maximal functional execution across standardized environmental conditions.

The construct encompasses five foundational developmental motor dimensions, structured according to ontogenetic movement progression:

• Dimension A: Lying and Rolling (17 items): Quantifies foundational cephalocaudal and proximal motor control, primitive reflex integration, dynamic antigravity cervical and trunk flexor/extensor activation, and axial rotational movement patterns. Tasks range from basic symmetrical head alignment in the supine position, full active hip and knee flexion against gravity, and purposeful unilateral crossing of midline to reach for objects, to rolling from supine to prone and vice versa, prone head elevation with extended forearms, prone pivoting, and controlled transitions to sitting. These items assess whether the child possesses the prerequisite postural trunk stability and segmental differentiation to manipulate the body across a support surface.
• Dimension B: Sitting (20 items): Focuses on postural equilibrium, head-on-trunk stabilization, dynamic trunk balance, and pelvic base-of-support stability in the seated posture. Items range from supported sitting on a mat with therapist trunk stabilization, progressing to independent unsupported sitting with arms free, functional reaching across and behind the body in transverse planes without loss of balance, side-sitting stabilization, controlled descent to the prone position, pelvic disassociation during transitional pivots, and independent sitting on low and high benches.
• Dimension C: Crawling and Kneeling (14 items): Assesses prone locomotion, quadrupedal antigravity weight-bearing, reciprocal coordination of upper and lower extremities, pelvic-femoral disassociation, and knee-weight-bearing balance. Items systematically assess prone creeping, sustaining four-point quadrupod kneeling, quadrupedal reach above shoulder level, reciprocal crawling across distances, ascending and descending four steps on hands and knees/feet, and transitioning into high-kneeling and half-kneeling postures with the arms completely free from supportive contact.
• Dimension D: Standing (20 items): Evaluates lower extremity axial loading, vertical postural alignment, closed-kinetic-chain ankle-knee-hip stability, and autonomous balance maintenance against gravitational torque. Tasks begin with pulling to standing at a bench and maintaining independent stance for 3 to 20 seconds, and advance to single-leg balance (holding a bench and with arms free), unassisted floor-to-stand transitions through half-kneeling, controlled squatting, picking up objects from the floor without falling, and lateral cruising along furniture.
• Dimension E: Walking, Running and Jumping (17 items): Measures dynamic bipedal ambulation, gait initiation and termination, directional changes, high-level motor coordination, and explosive lower extremity propulsion. Items evaluate assisted and unassisted linear walking over 10 steps, 180-degree directional turns, walking backward, carrying large objects during ambulation, walking within bounded narrow trajectories (20-cm corridors and 2-cm lines), stepping over obstacles elevated to knee height, sprinting forward and returning, kicking a ball with either foot, dual-foot jumping across horizontal and vertical space, single-leg hopping within a constrained perimeter, and alternating reciprocal stair ascent and descent holding a single handrail.

Theoretical Framework

The design, structure, and clinical logic of the GMFM are grounded in the theoretical paradigm shift that revolutionized motor development and pediatric neurology during the late 20th century: the movement away from rigid hierarchical neuromaturational theories toward Dynamic Systems Theory and the Neuronal Group Selection Theory (NGST).

Under the historical neuromaturational framework pioneered by Arnold Gesell and Myrtle McGraw, motor development was conceptualized as an invariant, linear unfolding of the central nervous system. In this outdated view, higher cortical centers progressively inhibited primitive subcortical reflexes, dictating a strict, predetermined sequence of developmental milestones governed solely by biological maturation. Clinical therapy grounded in this perspective (such as classical neurodevelopmental treatment approaches) prioritized movement “normality” and reflexive inhibition, positing that atypical movement kinematics precluded functional developmental progression.

The CanChild research group rejected this linear paradigm when engineering the GMFM. Instead, they adopted dynamic systems principles articulated by Nicolai Bernstein and expanded by Esther Thelen. Dynamic Systems Theory conceptualizes motor behavior not as the rigid output of a hardwired central nervous system program, but as an emergent, self-organizing phenomenon generated by the continuous, non-linear interaction of multiple subsystems: organismic/biomechanical constraints (musculoskeletal geometry, muscle strength, body mass, tone, joint range of motion), environmental constraints (surfaces, assistive devices, physical context), and task demands (functional intention, reaching, navigating barriers). Within this framework, a child with cerebral palsy exhibits atypical movement patterns not merely as an expression of brain pathology, but as an adaptive, compensatory solution to the constraints imposed by their damaged central nervous system, muscle spasticity, and altered biomechanics.

Consequently, the GMFM operationalizes gross motor function by categorizing what functional motor tasks the child can self-organize and achieve, rather than penalizing the child for utilizing non-normative biomechanical strategies. If a child accomplishes reciprocal locomotion or shifts from sitting to standing using alternative muscular synergies, compensatory trunk flexion, or broad-based foot placement, the GMFM credits the accomplishment of the functional task. This conceptual framework aligns directly with modern task-oriented neurorehabilitation, neuroplasticity, and goal-directed functional motor learning.

Validity

The psychometric validity of the Gross Motor Function Measure has been scrutinized and confirmed across hundreds of empirical peer-reviewed investigations spanning more than three decades.

Content and Construct Validity

During its initial development, content validity was established through an exhaustive Delphi process involving expert panels of pediatric physiotherapists, developmental pediatricians, physiatrists, and biostatisticians. The items were selected to represent developmental milestones spanning from birth to five years of age in typically developing children, while ensuring clinical applicability to individuals with physical impairments across all functional severity levels.

Construct validity was demonstrated by establishing that GMFM total and dimension scores correlate significantly with clinician ratings of motor severity and clinical staging systems. When the CanChild group developed the Gross Motor Function Classification System (GMFCS), GMFM scores demonstrated profound construct validity by stratifying distinct, stable gross motor development curves across GMFCS Levels I through V. Children classified in GMFCS Level I achieve near-ceiling scores on Dimensions A through E, reaching predicted GMFM-66 maximum plateaus of approximately 85–90 points, whereas children categorized in GMFCS Level V plateau at an average GMFM-66 score of roughly 20–25 points, primarily manifesting capacity limited to Dimensions A and partial B. Longitudinal validation studies tracking thousands of children across Canada, Europe, Australia, and North America have confirmed these distinct developmental trajectories.

Criterion and Convergent Validity

Because no preexisting evaluative gold standard existed at the time of the GMFM’s creation, criterion-related validity was initially demonstrated against subjective, blinded global clinician ratings of change. Physiotherapists scored video recordings of children before and after therapeutic intervals, rating them as “much worse,” “no change,” or “much better.” Changes in GMFM scores correlated strongly with these expert global clinical impressions (Spearman ρ typically ranging from .65 to .82, p < .001). Convergent validity has been repeatedly substantiated through strong statistical associations with concurrent pediatric functional instruments, including the Pediatric Evaluation of Disability Inventory (PEDI; r = .75 to .88 for functional skills mobility scales), the Functional Independence Measure for Children (WeeFIM; r = .70 to .85 for motor domains), and instrumented gait laboratory parameters (including stride velocity, cadence, and single-limb stance percentage).

Discriminant Validity

Discriminant validity is evidenced by the scale’s divergence from non-motor constructs. GMFM scores demonstrate low to negligible correlations with cognitive function indices, receptive language scores, and psychosocial behavioral measures (r < .25), confirming that the instrument isolates gross motor capacity rather than cognitive comprehension or linguistic competence.

Evaluative Validity and Responsiveness to Change

The critical psychometric virtue of the GMFM is its evaluative responsiveness. Using Guyatt’s Responsiveness Index, standardized response means (SRM), and Cohen’s effect sizes, the GMFM has proven exceptionally sensitive to small but functionally vital increments in motor development resulting from physical therapy, selective dorsal rhizotomy (SDR), lower extremity orthopedic surgery, and botulinum neurotoxin type A injections. In comparative clinical trials, the GMFM consistently detects measurable changes over 6- to 12-month intervention periods with large effect sizes (SRM > 0.80) in responsive patient populations, where classical norm-referenced instruments failed to demonstrate statistically significant changes.

Reliability

The Gross Motor Function Measure demonstrates robust reliability across inter-rater, intra-rater, and test-retest paradigms, verified across diverse clinical environments and cross-cultural adaptations.

Inter-Rater and Intra-Rater Reliability

In the seminal standardization trials conducted by Russell and colleagues (1989, 1993, 2002), inter-rater reliability was evaluated by having pairs of trained pediatric physical therapists independently and simultaneously score live clinical sessions, as well as scoring randomized videotaped administrations. For the GMFM-88 total score, the inter-rater intraclass correlation coefficient (ICC) consistently surpassed .98 (with 95% confidence intervals typically between .96 and .99). Across individual dimensions, inter-rater reliability remained remarkably elevated:

• Dimension A (Lying and Rolling): ICC = .95 to .99
• Dimension B (Sitting): ICC = .97 to .99
• Dimension C (Crawling and Kneeling): ICC = .96 to .99
• Dimension D (Standing): ICC = .96 to .99
• Dimension E (Walking, Running and Jumping): ICC = .97 to .99

Intra-rater reliability, evaluated by having identical raters re-evaluate masked videotaped administrations separated by a minimum two-week latency, yielded ICC values exceeding .99 for the total score and ranging from .96 to .99 across individual dimensions.

Test-Retest Stability

Test-retest reliability examines the stability of the instrument over brief time intervals (typically one to two weeks) during which clinical change in the child is theoretically absent. In clinically stable children with cerebral palsy, test-retest ICCs for the GMFM-88 total score range from .95 to .98. The Dutch cross-cultural validation conducted by Ketelaar, Vermeer, and Helders (1999) corroborated these findings, reporting test-retest ICCs of .99 for total scores and .92 to .98 for individual dimensions, demonstrating that the instrument is robust against administrator differences across linguistic adaptations.

Measurement Error and Minimal Detectable Change

The Standard Error of Measurement (SEM) for the GMFM-88 total score is estimated at approximately 1.2 to 2.1 percentage points. Consequently, the Minimal Detectable Change at the 95% confidence level (MDC95) ranges between 3.3% and 5.8% for the GMFM-88 total score. In the Rasch-based GMFM-66, the SEM is calculated across the difficulty continuum, with an overall MDC95 generally recognized as approximately 1.58 to 3.0 points depending on the child’s baseline motor ability level, providing clinicians with statistical thresholds to determine whether observed progress reflects authentic motor gain rather than measurement error.

Factor Analysis

The underlying dimensionality and psychometric architecture of the GMFM have been explored through classical exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and advanced Item Response Theory (IRT) models, most notably Rasch rating scale analysis.

Exploratory and Confirmatory Dimensionality

Originally, the 88 items were clinically clustered into five distinct motor dimensions based on anatomical postures and developmental emergence. Early exploratory factor analyses supported a primary general factor of gross motor capacity explaining over 70% to 80% of common variance, accompanied by sub-factors that mapped onto postural stability versus dynamic ambulation. Confirmatory factor analysis demonstrated that while a single second-order “gross motor capacity” construct accounts for the overarching variance, the five first-order latent dimensions (Lying/Rolling, Sitting, Crawling/Kneeling, Standing, Walking/Running/Jumping) demonstrate distinct domain variance, confirming the utility of generating individual dimension percentage profiles alongside the composite score.

Rasch Measurement Model and the Evolution to GMFM-66

Despite the high reliability of the GMFM-88, methodologists recognized a critical psychometric limitation in the original version: the GMFM-88 is an ordinal-scale instrument that calculates sum scores and equal-weighted averages. Summing ordinal scores presumes that each item difficulty step is equal and that progressing from 0 to 1 on an item represents an identical amount of motor ability as progressing from 2 to 3. Furthermore, it assumes that mastering an item in Dimension A (e.g., rolling prone to supine) requires the same latent gross motor capacity as mastering an item in Dimension E (e.g., hopping on one foot).

To overcome these limitations, Russell, Avery, and colleagues implemented Rasch analysis using the Partial Credit Model. Rasch modeling evaluates whether items conform to a mathematical function linking an individual’s latent motor ability level (θ) to the probability of endorsing specific ordinal item categories, thereby transforming raw ordinal responses into true, interval-level logit scales. The Rasch calibration of the GMFM revealed that:

• A subset of 22 items in the GMFM-88 exhibited item misfit, differential item functioning (DIF), or statistical redundancy within the unidimensional latent trait.
• By removing these 22 items, the authors created the GMFM-66, which satisfies the stringent requirements of unidimensionality and local independence (Goodness-of-Fit mean square statistics typically between 0.70 and 1.30; root mean square error of approximation [RMSEA] < .05).
• The GMFM-66 ranks all 66 items along a continuous, linear difficulty continuum, allowing for computerized scoring through the Gross Motor Ability Estimator (GMAE) software.

While the GMFM-66 offers interval-level scaling and handles missing data via maximum-likelihood estimation, the original GMFM-88 remains widely utilized in clinical practice when assessing children with severe impairments (GMFCS Level V) or children with Down syndrome, because the GMFM-88 contains specific foundational items in Dimensions A and B that provide nuanced granularity for very young or severely affected populations.

Instrument / Measurement Tool

• Instrument Name: Gross Motor Function Measure (GMFM); GMFM-88.
• Construct Assessed: Evaluative gross motor function and motor capacity across developmental motor domains.
• Administration Format: Direct, criterion-referenced observational clinical examination conducted by a trained pediatric professional (e.g., physical therapist, occupational therapist, physiatrist).
• Target Population: Pediatric populations with cerebral palsy (aged 5 months to 18 years), Down syndrome, traumatic brain injury, and spinal muscular atrophy.
• Testing Environment & Equipment:
  • Even, firm, non-slip floor or mat.
  • Smooth mat table/plinth.
  • Small wooden bench (approx. 20 cm height) and large bench (approx. 45 cm height).
  • Standard flight of stairs (at least 4 steps with at least one handrail).
  • Stopwatch, measuring tape, masking tape for floor markings (parallel lines 20 cm apart; straight line 2 cm wide; 60 cm circular perimeter).
  • Standardized red wooden/plastic toy, ball (soccer/basketball size), and a horizontal wooden stick (for jumping/stepping hurdles).
• Total Number of Items: 88 items grouped into five distinct dimensions:
  • Dimension A: Lying and Rolling: 17 items (Items 1 to 17; maximum raw score = 51)
  • Dimension B: Sitting: 20 items (Items 18 to 37; maximum raw score = 60)
  • Dimension C: Crawling and Kneeling: 14 items (Items 38 to 51; maximum raw score = 42)
  • Dimension D: Standing: 20 items (Items 52 to 71; maximum raw score = 60)
  • Dimension E: Walking, Running and Jumping: 17 items (Items 72 to 88; maximum raw score = 51)
• Authentic Response Scale: 4-point ordinal scale:
  • 0 = does not initiate
  • 1 = initiates (< 10% of task)
  • 2 = partially completes (10% to < 100% of task)
  • 3 = completes (100% of task)
  • NT = not tested
• Administration Guidelines:
  • The child is tested in minimal clothing (shorts and T-shirt, barefoot, except when assessing specific orthotic/footwear condition protocols).
  • A child may be given up to three attempts for each item; the highest observed score across the three trials is recorded.
  • Verbal encouragement and demonstration by the examiner are permitted; physical assistance is prohibited unless specifically stipulated by the item description (e.g., Item 18, 21, 22).
  • Testing duration typically ranges from 45 to 60 minutes. Testing may be split across two sessions within a 7-day interval if fatigue intervenes.
• Scoring and Computational Rules:
  • Calculate the sum of raw scores for each dimension.
  • Compute Dimension Percentage Score: $$\text{Dimension % Score} = \left( \frac{\text{Sum of Item Scores in Dimension}}{\text{Ma\ximum Possible Dimension Score}} \right) \times 100$$
  • Calculate Total GMFM-88 Percentage Score: $$\text{Total Score} = \frac{\text{Sum of all 5 Dimension % Scores}}{5}$$
  • If items are marked “Not Tested” (NT) in the GMFM-88, specific scoring handling algorithms apply (e.g., scoring as 0, or utilizing the Rasch-based GMAE software algorithm to compute the GMFM-66 score, which accommodates missing data without arbitrarily penalizing the individual).
  • Goal Total Score calculation: When specific dimensions are designated as priority therapeutic goals, an average of only those selected dimensions can be computed.

Permissions & Fee and Test Year

The Gross Motor Function Measure (GMFM-88) was initially released in 1989, followed by the definitive first edition manual published in 1993, the second edition in 2002, and the third edition in 2021 by Mac Keith Press. The copyright is held by McMaster University and CanChild Centre for Childhood Disability Research.

The administration manual and scoring sheets are commercial clinical and research instruments. The published manual, Gross Motor Function Measure (GMFM-66 & GMFM-88) User’s Manual (3rd Edition, 2021), along with the Gross Motor Ability Estimator (GMAE) software, is published and distributed worldwide by Mac Keith Press (London, UK) and John Wiley & Sons. Formal user fees, manual purchase charges, and licensing conditions apply for institutional clinical and commercial research trial deployments. CanChild provides access to official scoring sheets, translation permissions, and computerized scoring software tools through their institutional portal (CanChild McMaster University). To ensure assessment fidelity, examiners are strongly encouraged to complete self-instructional training criteria, calibration video exams, or official GMFM training workshops prior to collecting research or clinical outcome data.

References

• Ketelaar, M., Vermeer, A., & Helders, P. J. M. (1998). Functional motor abilities of children with cerebral palsy: A systematic literature review of assessment measures. Clinical Rehabilitation, 12(5), 369–380. https://doi.org/10.1191/026921598675367850
• Ketelaar, M., Vermeer, A., & Helders, P. J. M. (1999). Gross Motor Function Measure (GMFM): Handleiding Nederlandse versie. Utrecht University & Wilhelmina Children’s Hospital.
• Ketelaar, M., Vermeer, A., Hart, H., van Petegem-van Beek, E., & Helders, P. J. M. (2001). Effects of a functional therapy program on motor abilities of children with cerebral palsy. Physical Therapy, 81(9), 1534–1545. https://doi.org/10.1093/ptj/81.9.1534
• Palisano, R., Rosenbaum, P., Walter, S., Russell, D., Wood, E., & Galuppi, B. (1997). Development and reliability of a system to classify gross motor function in children with cerebral palsy. Developmental Medicine & Child Neurology, 39(4), 214–223. https://doi.org/10.1111/j.1469-8749.1997.tb07414.x
• Rosenbaum, P. L., Walter, S. D., Hanna, S. E., Palisano, R. J., Russell, D. J., Raina, P., Wood, E., Bartlett, D. J., & Galuppi, B. E. (2002). Prognosis for gross motor function in children with cerebral palsy: Creation of motor development curves. JAMA, 288(11), 1357–1363. https://doi.org/10.1001/jama.288.11.1357
• Russell, D. J., Rosenbaum, P. L., Cadman, D. T., Gowland, C., Hardy, S., & Jarvis, S. (1989). The gross motor function measure: A means to evaluate the effects of physical therapy. Developmental Medicine & Child Neurology, 31(3), 341–352. https://doi.org/10.1111/j.1469-8749.1989.tb04003.x
• Russell, D. J., Rosenbaum, P. L., Gowland, C., Hardy, S., Lane, M., Plews, N., McGavin, H., Cadman, D., & Jarvis, S. (1993). Gross Motor Function Measure Manual (2nd ed.). McMaster University.
• Russell, D. J., Rosenbaum, P. L., Avery, L. M., & Lane, M. (2002). Gross Motor Function Measure (GMFM-66 & GMFM-88) User’s Manual. Clinics in Developmental Medicine No. 159. Mac Keith Press.
• Russell, D. J., Avery, L. M., Rosenbaum, P. L., Raina, P. S., Walter, S. D., & Palisano, R. J. (2000). Improved scaling of the Gross Motor Function Measure for children with cerebral palsy: Evidence of reliability and validity. Physical Therapy, 80(9), 873–885. https://doi.org/10.1093/ptj/80.9.873
• Russell, D. J., Rosenbaum, P. L., Wright, M., & Avery, L. M. (2021). Gross Motor Function Measure (GMFM-66 & GMFM-88) User’s Manual (3rd ed.). Mac Keith Press.

Items of the Scale