This article provides an overview of outcome tools commonly used to assess ambulatory children with cerebral palsy, research findings from a recent large multi-center study, and ways to integrate the research findings into clinical practice. The information presented in this article and in the referenced articles provides information on: outcome tools’ discriminatory ability and responsiveness; readily available comparison data on 7 commonly used outcome tools that can be used at the point of care; prediction equations for the Parent report Pediatric Outcomes Data Collection Instrument (PODCI) by age and Gross Motor Function Classification System (GMFCS) level; and minimum clinically important difference thresholds by GMFCS level. This information can help clinicians select the best outcome tools to discriminate among severity levels and avoid ceiling effects. The scores provided allow direct comparisons between a specific patient and a matched cohort, assisting clinicians in the creation of comprehensive and individualized evaluation and management plans. Use of appropriate outcome tools to assess children with cerebral palsy can lead to best practices and reduced costs in the clinical setting.
An increasing demand for evidence-based decision-making has challenged the medical community to clearly show that treatment improves a person’s functional abilities within their environment. This has resulted in increased use of outcome tools in the clinical setting to supplement technical measures, such as physical examination. Outcome tools are used to measure functional performance, as a baseline descriptive assessment, to select treatment goals, and to evaluate treatment. Outcome tools can help maintain patient and family motivation and provide scientific evidence on outcomes that are meaningful to patients and clinicians.
Outcome tools objectively quantify some aspect (physical, mental, or emotional) of a person. They are typically developed with a specific purpose for a target population and are used in clinical care to assess or establish treatment efficacy. They should be valid, reliable, and responsive to change. Validity is the degree to which a scale measures what it is intended to measure. Reliability is the extent to which a measure yields consistent values over time when no change has occurred and yields consistent values between examiners. Responsiveness establishes a tool’s ability to detect change when a change has occurred.
Since clinicians may lack an understanding of what to measure, lack information regarding appropriate tools, have difficulty assessing psychometric properties of tools, have doubts about clinical utility of tools, or lack the time or money for training and implementation, outcome tools are not always used by clinicians consistently, as intended, or at all. Clinicians tend to use the easiest, most familiar, or available measure rather than what is “best” theoretically.
With a goal of providing clinicians and clinician-researchers with more information on outcome tools used to assess ambulatory children with cerebral palsy (CP), this article provides an overview of several outcome tools commonly used in this population, provides research findings from a recent large multicenter study, and provides ways to integrate the research findings into clinical practice and clinical outcomes research. The information presented is the result of work conducted by the Functional Assessment Research Group (FARG), a group comprising individuals from 7 pediatric orthopedic facilities (Shriners Hospitals for Children [SHC]-Lexington, KY; SHC-Northern California; SHC-Springfield; SHC-Houston; SHC-Salt Lake City; Washington University, St. Louis; and the University of Virginia). The FARG mission is to conduct research contributing to the evidence base for treatment of the neuromusculoskeletal system in ambulatory children with CP.
In 2001, the World Health Organization (WHO) released the International Classification of Functioning, Disability and Health (ICF) framework. The ICF provides a comprehensive framework to approach outcome measurement of children with disabilities. The ICF dimensions include Body Functions and Structures, Activities and Participation, Environmental factors, and Personal factors. Health-related quality of life is not formally included in the ICF framework, but was assessed by FARG, and is defined as “what people ‘feel’ about their health condition or its consequences.”
Several outcome tools assessing the ICF dimensions of Body Functions and Structures and of Activities and Participation that are frequently used in pediatric orthopedics were selected for this study. These included: the Gillette Functional Assessment Questionnaire (FAQ), Gross Motor Function Measure (GMFM), Pediatric Quality of Life Inventory (PedsQL), Pediatric Outcomes Data Collection Instrument (PODCI), Pediatric Functional Independence Measure (WeeFIM), temporal-spatial gait parameters (velocity, stride length, and cadence), and energy cost during walking (O 2 cost). With the exception of FAQ Questions Two and Three and PODCI subscales of Satisfaction and Expectations, each tool has been tested for content validity and reliability. These tools encompass technical measures (temporal-spatial gait parameters, O 2 cost), clinician-rated tools (GMFM D (standing) & E (walking, running, jumping), GMFM-66), parent-reported measures (WeeFIM, FAQ, PODCI, PedsQL), and child-reported measures (PODCI, PedsQL).
The Gross Motor Function Classification System (GMFCS) is a standardized classification system that categorizes the function of individuals with CP into 5 levels, with level I being the most functional and level V the least. The GMFCS is valid and reliable with excellent inter-rater reliability. The FARG study focused on ambulatory children of GMFCS levels I to III: the child at level I has some motor limitations but does not use devices, at level II has limitations walking indoors and outdoors but does not use devices, and at level III has limitations walking indoors and outdoors and uses assistive devices.
Goals of this article
The FARG study methodology, statistical analyses, and results are fully described in the series of articles published in Developmental Medicine and Child Neurology ( DMCN ) and Journal of Pediatric Orthopedics . This article will summarize the FARG study findings and succinctly answer common questions related to the clinical applications of outcome tools in children with CP. Questions related to both the cross-sectional and longitudinal study data are answered, followed by questions related to how to apply the data clinically. In addition to summarizing the study findings, examples of ways to integrate the findings into clinical practice are presented.
Cross-sectional and Longitudinal Questions
For ambulatory children with CP:
- •
Do outcome scores reflect differences based on GMFCS levels?
- •
What outcome tools best discriminate physical function?
- •
Do parent and child reports differ on outcome tools?
- •
Do children with hemiplegia function differently from those with diplegia?
- •
How do scores on outcome tools change over 1 year?
- •
Are outcome tools responsive to change?
- •
Does orthopedic surgery change function as measured by changes in scores on the studied outcome tools?
- •
What is a minimum clinically important change in outcome scores?
Clinical Practice Questions
- •
How do I integrate this information into my practice?
- •
Can I tell how the child is doing compared with others with similar severity level?
- •
Can I predict the normal range of scores for children with CP based on age and GMFCS level?
- •
How do I know if a changed score is clinically meaningful?
- •
Can I use outcome tools to focus on the treatment plan?
In answering these questions, comparison data are presented that will assist clinicians to understand how a child is functioning, develop individualized treatment plans, and know when a clinically meaningful change has occurred. Use of these standardized assessments can also improve communication among parents, patients, and care providers.
Methods
This 6-year prospective multicenter study was entitled “A Cross-sectional and Longitudinal Assessment of Outcome Instruments in Patients with Ambulatory Cerebral Palsy.” The study was conducted at 7 pediatric orthopedic facilities across the United States that treat children from several surrounding states. Institutional Review Board approval was obtained at each site and all participants signed consent, assent as appropriate, and privacy and confidentiality forms.
Participants
Participants were enrolled in the study which included cross-sectional and longitudinal endpoints. Inclusion criteria were: diagnosis of CP, GMFCS levels I to III, ages 4 to 18 years, and ability to complete a gait evaluation. Exclusion criteria were: previous selective dorsal rhizotomy, lower extremity orthopedic surgery within the past year, botulinum toxin-A injections within the past 6 months, or a currently operating baclofen pump.
Five hundred and sixty-two individuals completed the baseline assessments and 387 completed the follow-up evaluation (68.7%). Of those who did not complete the follow-up evaluation: 95 were unable to be contacted (16.9%), 29 declined (5.2%), 9 did not come for their follow-up appointment (1.6%), 4 were no longer ambulatory (0.7%), 2 had surgery less than 1 year from baseline to study completion (0.4%), and 36 did not participate for other reasons (6.4%). Six (1.6%) were excluded from analysis because of incomplete data, resulting in a final longitudinal sample of 381. There were no differences at baseline for age, height, weight, type of involvement, gender, GMFCS level, birth history, and ethnicity between those who completed the follow-up and those who did not.
Of the 562 participants who completed the cross-sectional component, there were 339 (60%) males, 223 (40%) females; 240 (43%) GMFCS level I, 196 (35%) Level II, and 126 (22%) Level III; 400 (71%) diplegic and 162 (29%) hemiplegic; and 83% were Caucasian. Mean age at baseline was 11.0 (standard deviation [SD] 4.3) years (range, 4.1–18.3).
For the 381 participants who completed the follow-up assessments, there were 230 (60%) males, 151 (40%) females; 174 (46%) GMFCS level I, 132 (34%) level II, and 75 (20%) level III; 265 (69%) diplegic and 116 (31%) hemiplegic; and predominately Caucasian (83%). Mean age at follow-up was 12.4 (SD 3.4) years (range, 5.2–20.5). Mean time between assessments was 1.4 (SD 0.4) years. There was no significant difference in mean age, height, weight, sex, or ethnicity distributions among GMFCS levels at either baseline or follow-up.
Between assessments, individuals received treatments based on their physician’s recommendations. During the study period, 87 participants (23%) had orthopedic surgery. The remaining participants had treatments that included physical therapy, bracing, and observation.
Tools and Study Methodology
GMFM Dimensions D and E , Parent and Child PedsQL, Parent and Child PODCI, FAQ, WeeFIM, O 2 cost, temporal-spatial gait parameters, and GMFCS level were collected at baseline and follow-up done at least 1 year later. Before study initiation, consistency among coordinators was verified; local coordinators were trained in GMFCS classification, tool administration, and data collection procedures. Data were collected in a study-specific database via direct computer entry and reviewed by the project manager for completeness and accuracy. Statistical analyses performed to address each study hypothesis are described briefly in the following results sections and they were also fully described in the previously published articles.
Methods
This 6-year prospective multicenter study was entitled “A Cross-sectional and Longitudinal Assessment of Outcome Instruments in Patients with Ambulatory Cerebral Palsy.” The study was conducted at 7 pediatric orthopedic facilities across the United States that treat children from several surrounding states. Institutional Review Board approval was obtained at each site and all participants signed consent, assent as appropriate, and privacy and confidentiality forms.
Participants
Participants were enrolled in the study which included cross-sectional and longitudinal endpoints. Inclusion criteria were: diagnosis of CP, GMFCS levels I to III, ages 4 to 18 years, and ability to complete a gait evaluation. Exclusion criteria were: previous selective dorsal rhizotomy, lower extremity orthopedic surgery within the past year, botulinum toxin-A injections within the past 6 months, or a currently operating baclofen pump.
Five hundred and sixty-two individuals completed the baseline assessments and 387 completed the follow-up evaluation (68.7%). Of those who did not complete the follow-up evaluation: 95 were unable to be contacted (16.9%), 29 declined (5.2%), 9 did not come for their follow-up appointment (1.6%), 4 were no longer ambulatory (0.7%), 2 had surgery less than 1 year from baseline to study completion (0.4%), and 36 did not participate for other reasons (6.4%). Six (1.6%) were excluded from analysis because of incomplete data, resulting in a final longitudinal sample of 381. There were no differences at baseline for age, height, weight, type of involvement, gender, GMFCS level, birth history, and ethnicity between those who completed the follow-up and those who did not.
Of the 562 participants who completed the cross-sectional component, there were 339 (60%) males, 223 (40%) females; 240 (43%) GMFCS level I, 196 (35%) Level II, and 126 (22%) Level III; 400 (71%) diplegic and 162 (29%) hemiplegic; and 83% were Caucasian. Mean age at baseline was 11.0 (standard deviation [SD] 4.3) years (range, 4.1–18.3).
For the 381 participants who completed the follow-up assessments, there were 230 (60%) males, 151 (40%) females; 174 (46%) GMFCS level I, 132 (34%) level II, and 75 (20%) level III; 265 (69%) diplegic and 116 (31%) hemiplegic; and predominately Caucasian (83%). Mean age at follow-up was 12.4 (SD 3.4) years (range, 5.2–20.5). Mean time between assessments was 1.4 (SD 0.4) years. There was no significant difference in mean age, height, weight, sex, or ethnicity distributions among GMFCS levels at either baseline or follow-up.
Between assessments, individuals received treatments based on their physician’s recommendations. During the study period, 87 participants (23%) had orthopedic surgery. The remaining participants had treatments that included physical therapy, bracing, and observation.
Tools and Study Methodology
GMFM Dimensions D and E , Parent and Child PedsQL, Parent and Child PODCI, FAQ, WeeFIM, O 2 cost, temporal-spatial gait parameters, and GMFCS level were collected at baseline and follow-up done at least 1 year later. Before study initiation, consistency among coordinators was verified; local coordinators were trained in GMFCS classification, tool administration, and data collection procedures. Data were collected in a study-specific database via direct computer entry and reviewed by the project manager for completeness and accuracy. Statistical analyses performed to address each study hypothesis are described briefly in the following results sections and they were also fully described in the previously published articles.
Results and clinical applications
Do Outcome Scores Reflect Differences Based on GMFCS Level?
To gain a better understanding of the outcome tools and provide clinical comparison data on the function of individual children to their peers with CP, a descriptive analysis was completed. The results showed that children in GMFCS levels I, II, and III functioned differently as measured by the study outcome tools. The mean scores for each outcome tool were clearly separated by GMFCS level. The completed results for all studied tools which included the mean, SD, range, interquartile range, and 95% confidence intervals were previously reported.
Analysis of variance (ANOVA) was used to assess the relationships between outcome scores and GMFCS levels. It was determined that there were direct relationships between outcome scores and GMFCS level for Parent and Child PODCI Global Function, Transfers and Basic Mobility, Sports and Physical Function, Parent PODCI Upper Extremity and Physical Function, WeeFIM domains of Self Care and Mobility, FAQ Question One (walking scale), GMFM Dimensions D and E, GMFM-66 score, O 2 cost, and temporal-spatial gait parameters. For all of the functional assessments, as the child’s severity level increased (from GMFCS I to II to III), the mean functional outcome scores decreased; an example is shown in Fig. 1 . However, for the quality of life measures of Parent and Child PODCI Comfort/Pain, Parent and Child PODCI Happiness, Child PODCI Expectations, and Parent and Child PedsQL Emotional Functioning, children of different GMFCS levels reported a similar quality of life.
Although there was clear separation of the means, other descriptive measures of central tendency overlapped. The large sample size in this study may have contributed to finding statistically significant differences among GMFCS levels that may not have clinical significance. The ANOVA examines if the means of the groups are different and provides little information about the variation within the groups. Therefore, the degree of separation and overlap of interquartile ranges (middle half of population, 25%–75%) among GMFCS levels were reviewed. Overlap between levels shows the heterogeneous nature of CP. There was minimal overlap between the lower end of GMFCS level I scores and the upper end of level III scores. However, there was substantial overlap between the lower end of level I scores and the upper end of level II scores. These findings are consistent with the difficulties reported in classifying patients between levels I and II. Minimal overlap was seen between the lower end of level II and the upper end of level III.
The FAQ data are nonparametric and required analyses using odds ratios. The analysis revealed that FAQ Question One showed a difference among all GMFCS levels indicating that parents’ report of their child’s walking ability is consistent with the clinician’s gross motor rating using the GMFCS. Parents reported on FAQ Question Two what they felt limited their child’s walking ability and on FAQ Question Three what skills their child is able to do under typical conditions. Only 27% of parents reported that pain limits their child’s walking ability and 39% reported safety as a limiting factor. As severity level increased, a greater percentage reported safety as a problem. Regardless of GMFCS level, the majority of parents felt that balance, endurance, and weakness were limiting factors for walking. For FAQ Question Three, parents indicated that nearly all children were able to walk up and down stairs using a railing (87%), kick a ball (92%), and step over an object (90%). Only 11% of the children were able to jump rope and 15% to ice or roller skate. The parent’s perspective is important to help focus goals and clinicians should be sensitive to parents’ concerns when developing treatment plans.
Since the GMFCS was designed as a broad classification system based on function, one would expect children in different GMFCS levels to score differently on outcome tools. The significant differences in mean outcome scores found on the tools used in this study indicate that children classified into different GMFCS levels function differently. These findings support the use of the study outcome tools to assess function and support the GMFCS as an appropriate method of classifying severity, despite some overlap between levels. Examples of the differences among GMFCS levels for the studied outcome tools are reported in Table 1 .
| Outcome Tool | GMFCS I Mean (SD) | GMFCS II Mean (SD) | GMFCS III Mean (SD) |
|---|---|---|---|
| GMFM D- standing | 94 (6) | 84 (8) | 52 (24) |
| GMFM E- walking | 93 (8) | 75 (16) | 32 (20) |
| Energy (O 2 ) cost (mL O 2 /kg/m) | 0.28 (0.1) | 0.38 (0.2) | 0.57 (0.3) |
| Parent PODCI global function | 81 (11) | 72 (12) | 62 (13) |
| Parent PODCI sports | 68 (17) | 51 (17) | 35 (17) |
What Outcome Tools Best Discriminate Physical Function in Ambulatory Children with CP?
Knowing which tools best discriminate among GMFCS levels can assist in the selection of tools for clinical or research use, since it is not practical and often not feasible to administer multiple tools during a single visit. The best outcome tool, however, is dependent upon the clinical/research question and clinical/research endpoints of interest. The recommendations presented here focus on the tools assessed in this study that were the most discriminatory among GMFCS levels.
Using effect size indices (ESIs) for parametric variables and odds ratios for nonparametric data, the magnitude of differences in scores on the study tools across GMFCS levels were quantified. Binary logistic regression models determined discrimination, and receiver operating characteristic (ROC) curves addressed sensitivity and specificity of the measures.
Based on the study tools’ ability to discriminate (large ESIs) among GMFCS levels, the best tools to measure physical function in ambulatory children with CP are the GMFM Dimension E or GMFM-66, the Parent report PODCI Sports and Physical Function, Global Function, and the FAQ Question One. Tools that performed well and that are good secondary measures are the WeeFIM Self Care and Mobility domains and temporal-spatial gait parameters. The least discriminatory tools were the quality of life and cognition measures; however, these are important in comprehensive assessments of treatment effects.
ROC curves were constructed to determine sensitivity and specificity for the most discriminant measures identified by regression models. Based on this analysis, the single most discriminatory measure of physical function, as classified by the GMFCS, is the GMFM-66 score. This is the Rasch-analyzed score of the GMFM using the Gross Motor Activity Estimator program. Rasch analysis equalizes the difference between points on the scale, creating an interval scale, such that the interval of change from 15 to 25 is equivalent to the change from 65 to 75. The GMFM-66 is the best clinical choice for discriminating among GMFCS levels I through III.
A tool’s discriminatory ability may be limited by ceiling effects. Ceiling effects suggest that either the participants do not have significant limitations in these areas, or that there is a need for more challenging “upper end” items or questions. Parent and Child reports of PODCI Comfort/Pain and WeeFIM Cognition have large ceiling effects for all three GMFCS levels, suggesting that pain and limitations in cognition are not major issues for these children or that the questions are too easy to identify if such issues are present.
As an example of a tool with limited discriminatory ability, Fig. 2 shows a box plot of the Parent report of PODCI Comfort/Pain scores. Note that there are no upper fences for any GMFCS level, showing ceiling effects. The boxes represent the interquartile ranges for the data, presenting the range of scores obtained by the middle 50% of the participants for each GMFCS level. The size of the boxes shows variability within GMFCS level and degree of overlap between GMFCS levels. For the example shown, ceiling effects and the overlap of interquartile ranges limit the discriminatory ability of this subscale.
Do Parent and Child Reports Differ on Outcome Tools?
The PODCI and PedsQL have parent and child report forms. Given the time constraints often present in the clinical setting, it would helpful to know if completion by both parent and child are necessary. Therefore, differences in parent and child responses on the same tools were studied. A direct relationship (as severity increased scores decreased) between outcome scores and GMFCS levels was seen for both the parent and child reports, despite the differences noted between scores. For 2 of the tools studied, the PODCI and PedsQL, there are both a parent and child report version. Results showed that there were differences between the scores reported by the parent and the child. Children scored themselves higher than their parents for almost all PODCI and PedsQL subscales. As the child’s GMFCS level increased, the differences between parent and child scores increased. Physical subscales of PODCI Sports and Physical Function and PedsQL Physical Functioning showed the greatest differences. Example data are presented in Table 2 .
| Outcome Tool | Parent Mean (SD) | Child Mean (SD) |
|---|---|---|
| PODCI sports | 57 (21) | 70 (19) |
| PODCI upper extremity | 84 (14) | 93 (10) |
| PODCI transfers | 86 (13) | 93 (9) |
| PedsQL physical functioning | 57 (19) | 69 (19) |
| PedsQL social functioning | 57 (19) | 65 (22) |
This is not the case for able-bodied children, who tend to report the same scores as their parents. The authors speculate that these findings are likely related to perspectives of disability. The child’s perception is one of ability, since the impairment was not acquired after a period of normal development. The child emphasizes what he can do. Parents have the expectation that their child should be able to do everything able-bodied children can. Therefore, the parent’s perspective is more likely one of disability and emphasizes what the child cannot do. For these reasons, it is important to obtain both parent and child perspectives on goals and outcomes.
Do Children with Hemiplegia Differ from Children with Diplegia?
Children with hemiplegia compared with those with diplegia were assessed, since grouping of children with CP with notably different clinical presentations, and presumably different profiles of brain injury, may lead to inaccurate predictions of developmental rate and prognosis.
Based on the scores of the studied outcome tools, children with hemiplegia function differently from those with diplegia. Differences in outcome scores between these groups within GMFCS levels were tested and children with hemiplegia consistently scored lower on tools related to upper extremity and school function, and higher on nearly all tools that assessed gait or lower extremity function.
Ceiling effects were shown to be a function of the diagnostic subtype, GMFCS level, and outcome tool. They were mostly higher in GMFCS level I. The ceiling effects reflected the different strengths in each of the groups and were not consistently higher in either diagnostic subtype. For example, the children with hemiplegia had a higher percentage of ceiling effects in the PODCI Transfers and Basic Mobility subscale and WeeFIM Mobility, which shows their reliably higher scores on lower extremity functional assessments. However, the children with diplegia had a higher percentage of ceiling effects on WeeFIM Self Care, which shows their somewhat better upper extremity function. It is important to note that these ceiling effects may have made differences between diagnostic subtypes less evident.
Objective evidence of the distinct differences between children with diplegia and hemiplegia within the same GMFCS classification level exists in various aspects of motor functioning, activity, participation, and quality of life. Using GMFCS level alone as a predictor of motor prognosis may overestimate the lower extremity skills of children with diplegia, and underestimate those of children with hemiplegia.
How Do Scores Change on Outcome Tools for Ambulatory Children with CP Over 1 Year?
Since children with chronic conditions such as CP are assessed repeatedly over time, it is important to know the expected amount of change as a result of natural history. Expected changes can be used by clinicians to assess whether observed changes with time are atypical, thus influencing treatment planning. Change data on the 381 individuals who completed both the baseline and follow-up assessments were analyzed. For the studied tools, small mean changes with large SDs were seen over a 1-year period. The small mean changes were a result of the outcome tool change scores being normally distributed, with some large positive and negative changes. The large variability in change scores is likely because of the heterogeneity of CP. Complete results were published by Oeffinger and colleagues in 2008; sample results are presented in Table 3 . In addition, the average change scores and SDs are presented by GMFCS level and surgical or nonsurgical group by Oeffinger and colleagues (2008). Since changes that occur over a year are often small, one should be particularly cognizant of those individuals that present with large changes in function in a short period of time.
Related posts:
New Clinical and Research Trends in Lower Extremity Management for Ambulatory Children with Cerebral Palsy
The Adult with Cerebral Palsy: A Provider-Consumer Perspective
Bone Density in Cerebral Palsy
Access to Employment and Economic Independence in Cerebral Palsy
Understanding Function and Other Outcomes in Cerebral Palsy
Establishing Access to Technology: An Evaluation and Intervention Model to Increase the Participation of Children with Cerebral Palsy
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
