INTRODUCTION AND PURPOSE
The Gross Motor Function Classification System (GMFCS) was originally designed and published in 1997 to assist researchers and clinicians in classifying children with cerebral palsy (CP) based on their gross motor function using 5 levels.1 Level I, the highest functional level in the GMFCS, describes children with CP who have the ability to participate in community settings with minimal functional deficits. On the opposite end of the spectrum, children in level V are typically fully dependent, present great difficulty with voluntary movements and are transported in wheelchairs.1
The validity and reliability of the original GMFCS have been studied extensively and are well established.2–5 Subsequently, in a review by Morris and Bartlett,6 a high level of uptake of this classification system internationally was reported; however, at that time there was a general lack of information on how the GMFCS was used in clinical practice and education. Since 2004, a large number of publications making use of the GMFCS have continued. Thus, a research question was proposed for this article: “How has the GMFCS been used in the literature since the previous review?” Using the time frame of July 2003 to December 2008, a scoping review of publications was performed to gain further insight on the impact and utility of this classification system. Identification of gaps will suggest areas for future investigations. Clarification of misuses of the GMFCS will refine the clinical utility of the system.
To present the relevant literature, a scoping review was used.7 This type of review focuses on a topic of interest that is wide-ranging in both purpose and study design.7 The 5 stages of a scoping review leads to an organized list of the literature that is comprehensive yet summarizes the main findings in a user-friendly way. It also does not require critiquing the quality of all relevant literature; thus, it focuses on breadth of information not depth. This is useful in that it gives the reader a variety of resources to find articles of interest so they can critique the information provided to them, and helps to identify gaps in the literature for both research and clinical decision-making.7
A thorough literature search was performed in OVID, MEDLINE, and CINAHL with the following key words: “Cerebral Palsy” and “GMFCS” or “Gross Motor Function Classification System.” Additional limits were applied specifically to locate literature published in English from July 2003 to December 2008, as the previous review ended with June 2003.6 To further identify any potential publications for inclusion in this scoping review, a citation search on Morris and Bartlett6 was also performed using both search engines.
Inclusion and Exclusion Criteria
Specific exclusion criteria were created, and prescreening was performed on all search results before initiating the scoping review. Search results that contained no abstracts or abstracts that specified “comment” only were excluded. Doctoral dissertations and reviews were excluded. Studies that misused the GMFCS were also excluded but are commented on in the discussion. All other publications were subsequently included in this scoping review according to the classification system that was used.
Literature Classification Process
The selected publications were classified into various categories according to the uses of the GMFCS. The 2 main classification themes were research studies and clinical practice. Categories under research studies included research on the measurement properties of the original GMFCS, research on the validity of other measurement tools with the GMFCS, as well as the use of the GMFCS in observational and experimental research studies.
For research on the measurement properties of the original GMFCS, the research studies that evaluated interrater reliability, validity, and stability of the GMFCS or translation of the GMFCS into another language were examined. For research on the validity of other measurement tools, there was an interest in publications that included the GMFCS in the validation process of specific measurement tools. We further classified the different types of measurement tools according to the components proposed by the International Classification of Functioning Disability and Health (ICF): Body structure and function, activity, participation, personal factors, and environmental factors.8 Quality of life was also added as the last subcategory in addition to the ICF components.
For observational and experimental research studies, there was interest in examining how researchers made use of the GMFCS in research. Therefore, an investigation was undertaken into whether there were any statistical tests of associations between the GMFCS levels and the outcomes of interest, or whether the levels were used only to describe the participants. More specifically, for experimental research studies, there was interest in the type of intervention being examined and whether the GMFCS was used for sample stratification prior to random allocation. Quasi-experimental studies, as defined by Cook and Campbell,9 were also considered as a type of experimental studies in this scoping review.
In addition, we were also interested in reviewing publications that commented on the use of the GMFCS in clinical practice. These publications were classified as either primary or secondary. Publications were classified as primary clinical practice studies when the sole purpose of the publications was to investigate the use of the GMFCS in clinical settings. On the other hand, when researchers made specific clinical practice recommendations based on research findings and participants' GMFCS levels, these publications were considered as secondary clinical practice studies. Clinical practice recommendations made by researchers were subsequently classified into subcategories according to the Guide to Physical Therapist Practice: examination, prognosis, intervention, and outcome.10 Furthermore, the use of the GMFCS in family-centered practice was also examined. Publications were reviewed that commented on parent-report versions or family-assessed instruments in the context of the GMFCS.
A total of 228 unique publications were retrieved from both searches. In addition, 10 publications were retrieved from the citation search of Morris and Bartlett.6 Out of the total of 238 results, 26 were excluded: 3 of them were doctoral dissertations; 2 did not contain an abstract; 8 were reviews; and 13 misused of the GMFCS. There were 212 publications included, of which 12 of them fit into 2 separate categories for a total reported frequency count of 224 as illustrated in Figure 1.
Summary tables illustrate: the classifications and subclassifications of the research studies on measurement properties of the original GMFCS (Table 1), the validity of other measurement tools with the GMFCS (Table 2), observational (Table 3) and experimental (Table 4) research studies making use of the GMFCS, as well as studies on the use of the GMFCS in clinical practice (Table 5). In addition to the summary tables, 4 publications that examined the use of the GMFCS in family-centered practice were identified. To summarize the findings, overall frequency counts of publications in each category as well as frequency counts of all publications that made use of the GMFCS each year since its introduction in 1997 are presented in Figures 1 and 2, respectively.
To present the findings clearly and concisely, only summary tables and figures, as described earlier, are presented in the main context of the current publication. To assist readers in locating further information and references, a detailed alphabetized reference list of all publications and detailed charts containing all classification information are available as electronic appendices. The detailed alphabetized reference list is located in Appendix A, http://links.lww.com/PPT/A10, followed by 5 detailed charts in Appendix B, http://links.lww.com/PPT/A11.
The detailed charts contain references that are first arranged chronologically according to the year published year and then alphabetically according to the authors' names for those published in the same year. They contain detailed information on references and subclassifications of individual publications. The information contained in Appendix B (http://links.lww.com/PPT/A11), Detailed Charts 1 to 5, respectively, correspond to Tables 1 to 5 in this article. Selected reference numbers can be identified in the Discussion section at the end of sentences in square brackets preceded by the letter “A,” which stands for appendix.
Publications on the various uses of the GMFCS were reviewed. Selected publications on the measurement properties of the original GMFCS, validity of other measurement tools with the GMFCS, observational and experimental research studies, as well as on clinical practice are discussed to illustrate the major findings of the scoping review.
Research on the Measurement Properties of the Original GMFCS
The current review suggested that the research on the measurement properties of the original GMFCS is continuous in process so that the validity, reliability, and stability are further examined in different settings. Two publications involved the determination of interrater reliability of the GMFCS, with kappa values ranging from 0.57 to 0.75.[A-50,109] Emphasis was on consistently better agreement with children older than 2 years.[A-109] Also, a few publications explored the other aspects of the GMFCS. Content validity was examined on the expanded and revised version of the GMFCS using group consensus methods for the 6- to 12-year and 12- to 18-year age bands.[A-140] Another study by the original developers further supported the stability of the GMFCS, concluding that children at levels I and V are least likely to be reclassified.[A-138] Moreover, researchers also translated the GMFCS into a Greek version and found it to be reliable.[A-141]
The above findings further supported the use of the GMFCS in research as well as in clinical settings. The GMFCS has also demonstrated good reliability, validity, and stability for children with CP between the ages of 2 and 12 years, as well as validity for youth between the ages of 12 and 18 years.11 However, investigation is needed to further assess the reliability and stability of the GMFCS for the 12- to 18-year age band.11
Validity of Other Measurement Tools With the GMFCS
The GMFCS has been used to examine the validity of other measurement tools for children with CP. Most studies focused on measures that quantify activity as well as body structure and function, with a few of them evaluating participation (Table 2). Measures that quantify activity limitations included the Timed Up and Go test,[A-51,203] Berg Balance Scale,[A-51] Child Health Questionnaire[A-189], and Shuttle Run tests.[A-195] With respect to body structure and function, measures included the Spinal Alignment and Range of Motion Measure,[A-10] the Pediatric Reach Test,[A-9] the Functional Reach Test,[A-51] and limb distribution and type of motor impairment,[A-60] as well as a measure of overall health for preschool children.[A-157] Measures of quality of life were the CP Quality of Life for Children measurement,[A-201] the Caregiver Priorities and Child Health Index of Life With Disabilities,[A-124] and the Pediatric Quality of Life Inventory.[A-192]
Notably, no studies made use of the GMFCS to validate measurement tools on personal or environmental factors. This could indicate a potential area of future research, as personal and environmental factors are integral parts of the ICF and should be examined further.8
Observational Research Studies
Observational studies continued to be the major type of research studies that made use of the GMFCS. These studies examined associations with diagnostic tools, prognostic and physical factors, as well as overall quality of life experienced by children with CP. Although numerous studies used only GMFCS levels to describe their participants, many studies made use of the GMFCS in their data analyses, attempting to establish a correlation between the GMFCS levels and their outcomes of interest.
Physical factors including severity of visual and motor deficits,[A-56] presence of drooling and saliva production,[A-164] prevalence of obesity in ambulatory children,[A-150] and the use and impact of assistive devices[A-135,137,155] were also correlated with the GMFCS levels of their participants. Other physical characteristics studied included hip displacement/dislocation,[A-63,119,171] body mass index,[A-46,52,79], energy efficiency in gait,[A-83,86], gait velocity,[A-52] muscle thickness,[A-133] dysphagia,[A-21] and nutritional support.[A-69,173,193] Some observational studies attempted to establish a relationship between overall quality of life of children with CP and their GMFCS levels. In these studies, GMFCS levels were correlated with health-related quality of life[A-14,153,165,186,191] as well as with the quality of life of mothers who take care of children with CP.[A-43,134]
Some studies also used the GMFCS as the outcome of interest, thus looking into factors that could potentially predict GMFCS levels. For example, da Costa and colleagues[A-31] found that there was a high positive correlation between visual acuity loss and GMFCS levels, leading to more questions about the potential cause of the vision problem. Furthermore, Roze and associates[A-154] identified that a highly extended form of periventricular hemorrhagic infarction was associated with the development of CP, but not with the severity of it. Kulak and Sobaniec[A-93] reported a negative correlation between corpus callosum surface area and the GMFCS levels of children with CP and further noted that magnetic resonance imaging findings and GMFCS classifications were significantly correlated in children with diplegic and tetraplegic CP.[A-94] The same researchers also found that children with spastic diplegia were more likely to be at levels I or II, while those with spastic tetraplegia were more likely to be at levels IV or V. Therefore, the reported GMFCS levels could reflect the severity of functional limitation.[A-95] A large multicenter study also reached similar conclusions, as they found that the children with more limb involvement had higher GMFCS levels, but indicated that the GMFCS was a better indicator of gross motor functional impairment than the traditional indicator of the number of limbs involved.[A-198]
Experimental or Quasi-experimental Research Studies
Experimental studies mainly focused on examining the effect of specific interventions on children with CP, and the GMFCS primarily played a role in participant descriptions or stratification prior to random allocation. Experimental studies mainly examined the application of medication, general exercise training, rehabilitational services (occupational therapy or physical therapy), adjunctive and recreational activities, assistive devices as well as orthopedic/neurological surgeries. Medical interventions for CP identified in this review included botulinum toxin type A injection,[A-4,12,13,35,58,100,101,107,127,158,159,160,210] intrathecal Baclofen,[A-16,77,120,161,163] hyoscine skin patches for drooling,[A-156] and muscle release surgery.[A-91] Assistive devices mainly included dynamic ankle foot orthoses for gait and spasticity management.[A-112,125,188] Studies also examined the benefits of intermittent versus continuous physiotherapy for children with CP.[A-27] General exercises included hippotherapy,[A-64] robotic-assisted locomotor training,[A-112] a static bicycling program for nonambulant CP,[A-204] partial body-weight–supported treadmill training,[A-26,38,43] reach performance,[A-125] as well as pediatric endurance and limb strengthening.[A-49] Adjunctive recreational training included conductive education (effects on hand motor functions),[A-17] massage on mechanical behavior of muscles,[A-102] therapeutic taping,[A-47] application of a novel pointing-device apparatus,[A-25] gastrostomy feeding,[A-142,178] comparison of efficacy of Adeli suit and neurodevelopmental treatments,[A-6] functional therapy,[A-1] balance training on muscle activity and stability,[A-168, 207] a school-based conductive education program,[A-209] and neurodevelopmental treatment.[A-6,185]
Similar to the findings in some observational studies, associations were identified among the GMFCS levels and the outcomes of interest in some studies. For example, Macgregor and colleagues[A-102] found that children can have sustained improvements in GMFM-66 scores postmassage of calf muscles for children who were in GMFCS levels I or II only. Bar-Haim and associates[A-6] found that Adeli suit treatment may improve mechanical efficiency in children with GMFCS levels II and III but not IV and V. One research group also looked into parents' perceived perception of 2 intensive training programs, and it was found that the higher the GMFCS level, the lower the probability of fulfillment of needs and functional improvements.[A-129]
The GMFCS is becoming increasingly important in clinical practice and the clinical decision-making process. Two publications primarily investigated the use of the GMFCS in clinical practice. Hanna et al[A-65] developed cross-sectional reference percentiles for the Gross Motor Function Measure-66 within the GMFCS levels to improve normative interpretation of scores, thus improving clinical utility. Also, Ketelaar et al[A-87] examined the challenges and possible strategies to incorporating evidence-based measures like the GMFCS into clinical practice in the Netherlands.
In addition, specific clinical practice recommendations were made in research studies secondarily to assist clinicians in examination, as well as determining prognosis, intervention, and outcome. Recommendations from these studies were condition-specific and provided clinicians with insights on decision making in their clinical practice. For example, it was found that the GMFCS and Manual Ability Classification System worked well together in clinical practice.[A-23] Also, a correlation was found between hip dislocation and GMFCS levels; therefore, a hip surveillance program with x-rays for children with CP was recommended on the basis of GMFCS levels.[A-63] GMFCS levels were also found to be important for assessing the risk of hip displacement.[A-171] Moreover, children with GMFCS levels of lower gross motor function may be predisposed to worsening of hip subluxation after selective dorsal rhizotomy.[A-71] For self-care and mobility, it was found that children with GMFCS levels II to V will need increased physical assistance during the day. Furthermore, the classification has implications on the role of physiotherapy in transition planning.[A-139] Selective muscle release surgery appears to be the most beneficial for children at GMFCS levels III and IV based on research findings.[A-91] In addition, GMFCS levels are also helpful in determining the energy costs or the metabolic demands, which are important in the treatment selection process.[A-83]
Some studies were done on a parent-report version of the GMFCS to examine the possibility of expanding its use in clinical practice; however, further research is still needed. It was found that these parent-report instruments are useful for classifying children's activities and participation.[A-116] However, discrepancies still existed between families and professionals for GMFCS level classifications, and these discrepancies could be a result of variations seen in children's performance in different environments.[A-118] McDowell et al[A-109] also found that agreement with parent report increased with the therapists' experience with the GMFCS, as well as how knowledgeable the therapist was about the child. Regardless, reliability of parent-report versions of the GMFCS was found to be high.[A-118] There was also an excellent agreement between family reports of the GMFCS and GMFCS classifications from professionals.[A-115] Family assessment tools are indicated for future use in research and clinical practice.[A-115,118]
Misuses of GMFCS
It is important to examine researchers' knowledge on the use of the GMFCS in this scoping review. Thirteen studies were found to have misused the GMFCS. Three studies inappropriately used the GMFCS in adults with CP and were thus excluded. The GMFCS was designed to classify children on the basis of their gross motor functional level in childhood, originally younger than 12 years and now up to 18 years of age.11 One study used the GMFCS to classify gross motor function of children with traumatic brain injuries, and another with children with Down syndrome; however, the GMFCS has been validated for use only in children with CP. Many researchers used the GMFCS as an outcome measure with the inappropriate interpretation that GMFCS levels would change in response to interventions over time. This does not reflect the true intention of the measure, as the GMFCS level typically does not change. Stability of the GMFCS was previously studied and found to be satisfactory, despite the fact that there were still some children being reclassified over time.[A-138] These reclassifications could reflect situations such as a child appearing to be a low-functioning level III at one particular time and a high level II at another point.
On the basis of this scoping review, it has been determined that the GMFCS has had a variety of uses in both research settings and clinical practice from July 2003 to December 2008, as illustrated in Figure 1. Although the GMFCS is primarily used in observational and experimental research studies, there is an emerging use of the GMFCS in clinical and family-centered practice. In addition, the current findings, as well as the findings from Morris and Bartlett,6 indicate that GMFCS use is steadily increasing over time, as shown in Figure 2. This observation is consistent with the previous findings that the GMFCS continues to have a positive impact in research as well as a high potential for expanding its use in clinical practice.6 Importantly, therapists should be aware that the system is not appropriate for use with adults, or for those without a diagnosis of CP. Finally, it is appropriately used as a descriptor or a determinant, but not as an outcome.
Direction for Future Research
Research on the utility of the GMFCS in clinical practice and with the use of parent-report tools12 should be continued to further explore its use in promoting effective, efficient, and family-centered practice. At present, no publication on the use of the GMFCS in clinical education has been identified, reflecting a need for further examination and documentation of this use. Anecdotally, one therapist's report indicates that the GMFCS training DVD is very useful in assisting physical therapy students in understanding the wide range of functional ability levels of children with CP.
This scoping review demonstrates that the GMFCS is clearly established as a classification system to describe research participants and to investigate the variation in gross motor functional level for a variety of research purposes. Most recently, the system has been expanded to include youth with CP aged 12 to 18 years.11 It appears that the use of the GMFCS is increasing in family-centered practice and education, but this remains to be documented. It is anticipated that the use of the GMFCS will further expand over time, including its use in clinical practice, family-centered practice, and education.
This work was completed in fulfillment of the requirements of the MPT degree for Ms Gray and Ms Ng at the University of Western Ontario.