Gross motor capability and performance of mobility in children with cerebral palsy: a comparison across home, school, and outdoors/community settings.

Cerebral palsy (CP), the most common physical in children, (1) represents the most frequent diagnosis of children who receive physical therapy. (2) The severity of limitations in gross motor function among children with CP is highly variable, such that some children with CP walk independently with or without assistive devices, whereas others use battery-powered wheelchairs or are transported by an adult. In research and practice, mobility in children with CP has generally been examined in a clinical setting using standardized measures in which very specific directions are followed. (3-6) Most standardized tests are administered in a controlled setting without environmental distracters (eg, noise, other people, physical obstacles). By minimizing environmental factors, such tests measure a child's capability but may not reflect a child's performance in everyday settings.

Recent research has examined differences in mobility of children with CP across environmental settings. (7-10) Although these studies provided evidence of the influence of environmental settings on mobility in children with CP, the relationship between capability and performance has not been examined. The distinction between capability and performance is made on the basis of environmental context. (11) Capability can be defined as the child's abilities in a defined situation apart from real life. (11) Capability reflects what a child can do when, among other things, the environment is controlled to eliminate contextual factors that are usually present in everyday settings. (12,13) In contrast, performance can be defined as the child's execution of activities in everyday settings, (11) such as the home, school, and community. Performance reflects what a child does do in everyday settings. (12,13) Previous research has documented differences in the capability and performance of activities of daily living in children with physical disabilities, indicating that information on capability may not necessarily be useful to extrapolate to performance. (11)

The concept of person-environment interaction provides a framework for understanding the relationship between capability and performance in children with CP (Fig. 1). A child with CP has a particular capability in each developmental domain (eg, gross motor, fine motor, cognition, vision). The construct of "person" includes both capability and personal factors (eg, age, personality). The interaction of the person with the environment leads to the performance of an activity. (14) For children with CP, the contextual features (physical, temporal, and social) of their home, school, and community are likely to have an important impact on the performance of mobility. For example, contextual features of the school setting may include physical features, such as variable surfaces (eg, stairs, carpeting), temporal features, such as keeping up with peers, and social features, such as expectations for age-appropriate mobility.

[FIGURE 1 OMITTED]

The purpose of our study was to examine the capability and performance of children with CP. Capability was defined as the highest of 3 mobility items (crawling, walking with support, or walking alone) completed on the Gross Motor Function Measure (GMFM). (15) Because the GMFM was administered in a controlled situation, we believe it measures capability. Children were grouped according to the most difficult item that they passed, representing their highest capability. Performance was defined as the usual mobility methods used in the home, at school, and in the outdoors or community. Even though children may use more than one mobility method in each setting, "usual" mobility was defined in this study as the one method most often used in a setting, as reported on a parent questionnaire. The research question was: Among children with CP who have similar gross motor capability, is there a difference in performance of mobility methods across home, school, and outdoors or community settings? Knowing the differences between capability and performance can have implications for examination and interventions used to improve functional mobility in children with CP.

Method

Subjects

The subjects were 307 children with CP (168 male, 139 female), aged 6 to 12 years ([bar.X] = 8.7, SD = 1.7), who participated in a prospective longitudinal study of the development of gross motor function. (16) We previously reported the effect of environmental setting on mobility methods (7) and the changes that occur in mobility methods over time (10) of these 307 children. A parent or guardian provided informed consent. Subjects were randomly selected from 19 centers throughout Ontario, Canada, and were stratified by age and Gross Motor Function Classification System (GMFCS) level. (17) Children were included in the study if they had a diagnosis of CP made by a physician. The definition of CP proposed by Bax (18) was used. The age range of 6 to 12 years was selected because evidence indicates that gross motor capability of children with CP begins to plateau at an average age of 2.7 to 4.8 years depending on gross motor classification level. (16) Children were excluded from the sampling frame if they had received selective dorsal rhizotomy surgery, intrathecal baclofen, or butulinum toxin injections in the lower limbs prior to study recruitment. These interventions potentially alter gross motor function. Of the 370 children in the original dataset, the data of 63 children were excluded from the data analysis for one or more of the following reasons:

* One month or longer interval between the measurements of capability (GMFM) and performance (parent questionnaire) (n=45).

* A person other than a parent completed the parent questionnaire (n=2).

* A parent indicated more than one mobility method for a setting (n=10).

* Children whose GMFM scores for the items analyzed did not fit the pattern of difficulty predicted by Rasch item response analysis (19) (n=6). These children reflected inconsistencies in the usual pattern predicted by Rasch analysis, because they received credit for difficult items (such as walking alone) but did not pass lower items (such as crawling). These children did not pass these items either because the items were not tested or the child did not cooperate with testing.

Children were grouped according to their gross motor capability (highest item passed on the GMFM). The 4 mutually exclusive groups were: children who were unable to crawl (n=99, 32% of sample), children who were capable of crawling (n=40, 13% of sample), children who were capable of walking with support (n=34, 11% of sample), and children who were capable of walking alone (n=134, 44% of sample).

Table 1 describes the GMFCS levels, distribution of CP, and type of CP for each capability group. The majority of children who were unable to crawl were at GMFCS levels IV or V and had spastic quadriplegia. The majority of children who were capable of crawling were at GMFCS levels III or IV and had spastic quadriplegia or diplegia. Children who were capable of walking with support were mostly at GMFCS level III and had a variety of distributions of CP, including spastic diplegia, triplegia, and quadriplegia. Children who were capable of walking alone were predominantly at GMFCS level I and had spastic diplegia or hemiplegia.

Measures

Parent questionnaire. A parent questionnaire, developed for this study, included information pertaining to the child's usual mobility methods in the home, school, and outdoors or community settings. Several authors support using parent reports to measure the performance of children in everyday settings. Long (20) stated that parent reports are appropriate when measuring the typical performance of children, in order to consider performance across various settings. In support of this idea, Wilson et al (21) stated that parent questionnaires provide a qualitative, accurate assessment of children's skills in a natural environment. Parent reports of children's current skills have consistently been shown to be a sensitive, reliable, and valid source of information. (21) The questionnaire in this study utilized a recognition format (eg, "Does your child use walking alone in the home?") that has greater reliability than an identification format (eg, "What mobility method does your child use at home?"). (22)

In the questionnaire, parents were told to choose the one mobility method that best described the child's "usual way of getting around" in each of the following settings: (1) home, (2) school, and (3) outdoors or in the community. For each setting, parents had to choose one of the following mobility methods: (1) carried by an adult, (2) pushed by an adult in a stroller, wheelchair, or other similar piece of equipment, (3) rolls, creeps, or crawls on the floor, (4) takes steps holding on to walls or furniture, (5) takes steps holding on to an adult's hands, (6) walks using a walking aid (a piece of equipment), (7) walks alone without any assistance, (8) propels self in regular wheelchair, (9) operates a battery-powered wheelchair, or (10) not applicable.

Gross Motor Function Measure. The GMFM (15) is a standardized, criterion-referenced test designed to measure change in the gross motor function of children with CP. Evidence of the reliability and validity of GMFM scores has been reported. (15,23-26) The GMFM was administered following standardized procedures, including encouraging the child's best possible effort for each item attempted. (22) It was administered in a setting without environmental interferences. For example, the GMFM manual. (15) states that the floor should be a smooth, firm surface. The score assigned represents the child's best effort over a maximum of 3 trials and ranges from 0 (does not initiate item) to 3 (completes item). For our study, scores were converted into a dichotomy of pass/not pass, with a passing score defined as a score of 3. Only the GMFM items administered directly by the therapist, and not the items generated by the parent report, were used to represent capability. The entire GMFM was administered without mobility aids or orthoses. If the child typically used mobility aids or orthoses, standing and walking items were administered a second time with the typical mobility aids or orthoses. For consistency among children and test items, the scores obtained without the use of mobility aids or orthoses were used to represent capability.

The GMFM has recently undergone Rasch item response analysis, which enables the items to be arranged in a hierarchical order of relative difficulty. (15,19) An item's difficulty, corresponds to the ability required to pass that item. In our study, capability was represented by the highest of 3 GMFM items attained, based on Rasch item response analysis. (15,19) The 3 GMFM items were "crawling," "walks with support," and "walks without support" (Tab. 2). These items were chosen because they were the mobility items that best corresponded to the mobility methods listed in the parent questionnaire and they represent the largest differences in difficulty derived from Rasch item response analysis. These items also represent self-initiated movements that are used by children in everyday settings.

Procedure

The parent questionnaire was completed by the children's mother (n=255, 83%), father (n=37, 12%), adoptive mother (n=12, 4%), stepfather (n=1, 0.3%), or both parents (n=2, 0.7%). Although parents were instructed to choose one method when reporting their children's usual mobility, 10 parents (3% of original sample) reported more than one method, and therefore data for these children were excluded from the data analysis.

The GMFM was administered and a therapist questionnaire was completed by 89 physical therapists, 3 occupational therapists, and 1 kinesiologist. Prior to administration of the GMFM, all therapists were trained to administer and score the GMFM and were tested to ensure that they reached a high level of agreement (weighted kappa >.80) against a criterion test videotape. (27) Therapists had an average of 10.3 years of experience (SD=7.3), ranging from less than 1 year to 32 years of experience. The therapist who performed the testing did not necessarily provide services to the child.

The majority of the parent questionnaires (n=164, 53% of sample) were completed on the same day of GMFM administration. For the sample, the mean time interval between administration of the parent questionnaire and administration of the GMFM was 5 days (SD=8, range=0-31).

Data Analysis

Descriptive statistics were used to describe all mobility methods for each group of children. For descriptive analyses, the mobility methods "takes steps holding on to walls or furniture," "takes steps holding on to an adult's hands," and "walks using a walking aid" were combined into one category: "walks with support."

For each group, mobility methods were then converted from nominal to dichotomous data in order to analyze the differences in mobility methods across settings. The dichotomy reflected whether children performed their highest capability item in each setting, with 0 indicating "child does not perform the capability method in that setting" and 1 indicating "child does perform the capability method in that setting." For example, children who were capable of walking alone (as indicated by a score of 3 on GMFM item 70), received a score of 0 if they did not walk alone in the outdoors or community and a score of 1 if they did walk alone in the outdoors or community (as their usual method of mobility reported on the parent questionnaire). This method of scoring enabled a direct comparison of scores across settings for each group. Frequencies and percentages of these dichotomous scores were calculated. The modified Wald procedure (28) was used to calculate 95% confidence intervals (CIs) for children who did not perform their capability method in each setting.

A Cochran Q analysis was conducted separately for each of the 3 capability groups. The Cochran Q is a nonparametric test that is appropriate for related dichotomous data (29) and can be used to indicate whether children with similar capability demonstrate differences in performance across settings. When the Cochran Q was significant (alpha level of <.05), additional Cochran Q analyses were conducted to determine where differences in performance among settings occurred. For these post hoc analyses, we used the Bonferroni correction, whereby alpha level was divided by the number of comparisons, in order to adjust for multiple comparisons. (30) The determination of significance for post hoc comparisons, therefore, was based on an alpha level of <.02 (Bonferroni corrected P value for 3 comparisons).

Results

Children who were unable to crawl (n=99) (Fig. 2) were reported to perform the following mobility methods across settings: At home, 33% of children were pushed by an adult, and 28% were carried by an adult. At school and in the outdoors or community, most children were pushed by an adult (school: 65%; outdoors or community: 75%). Ten children (10%) used a battery-powered wheelchair at home, and 18 children (18%) used a battery-powered wheelchair at school and in the outdoors or community. Even though they did not pass the crawling item on the GMFM, 18 children (18%) used rolling, creeping, or crawling at home.

[FIGURE 2 OMITTED]

Among children who were capable of crawling (n=40), 33% did not crawl at home, 98% did not crawl at school, and 100% did not crawl in the outdoors or community (Tab. 3). Children who were capable of crawling were reported to perform the following mobility methods across settings (Fig. 3): at home, 68% of the children used rolling, creeping, or crawling, and 23% walked with support; at school, 30% of the children used a regular wheelchair, 28% walked with support, 20% were pushed by an adult, and 13% used a battery-powered wheelchair; and in the outdoors or community, 55% of the children were pushed by an adult, 15% used a regular wheelchair, and 15% used a battery-powered wheelchair. The overall Cochran Q analysis was significant ([Q.sub.(2)] = 50, P<.0001). Post hoc analyses indicated that the children performed crawling more at home than at school ([Q.sub.(1)] = 24, P<.0001) and more at home than in the outdoors or community ([Q.sub.(1)] = 27, P<.0001). There was no difference between crawling at school and crawling in the outdoors or community.

[FIGURE 3 OMITTED]

Among the children who were capable of walking with support (n=34), 56% did not walk with support at home, 32% did not walk with support at school, and 59% did not walk with support in the outdoors or community (Tab. 3). Children who were capable of walking with support were reported to perform the following mobility methods across settings (Fig. 4): at home, 44% of the children walked with support, 38% used rolling, creeping, or crawling, and 15% walked alone; at school, 68% of the children walked with support and 18% used a battery-powered wheelchair; and in the outdoors or community, 41% of the children walked with support, 24% were pushed by an adult, 18% used a battery-powered wheelchair, and 12% used a regular wheelchair. The overall Cochran Q analysis was significant ([Q.sub.(2)] = 8, P<.021). Post hoc analyses indicated that the children performed walking with support more at school than in the outdoors or community ([Q.sub.(1)] = 9, P<.003). There was no difference between walking with support at home and walking with support at school. Similarly, there was no difference between walking with support at home and walking without support in the outdoors or community.

[FIGURE 4 OMITTED]

Among the children who were capable of walking alone (n=134), 5% did not walk alone at home, 10% did not walk alone at school, and 19% did not walk alone in the outdoors or community (Tab. 3). In the outdoors or community, 13% of the children used walking with support (Fig. 5). The overall Cochran Q analysis was significant ([Q.sub.(2)] = 26, P<.0001). Post hoc analyses indicated that the children performed walking alone more at home than at school ([Q.sub.(1)] = 7, P<.008) and more at home than in the outdoors or community ([Q.sub.(1)] = 18, P<.0001). They also performed walking alone more at school than in the outdoors or community ([Q.sub.(1)] = 9, P<.002).

[FIGURE 5 OMITTED]

Discussion

Our results demonstrate differences between gross motor capability and parent reports of performance of mobility methods used by children with CP. The results also indicate that, among children with CP who have similar gross motor capability, there are differences in performance across settings. This discrepancy between capability' and performance may be attributed to differences in the contextual features of settings. (11, 13) The measurement of capability, usually involves standardized testing devoid of contextual features, which can be present in children's daily lives. In contrast, performance occurs within the context of everyday settings such as home, school, and outdoors or community. In terms of mobility for children with CP, our results indicate that capability for mobility (what a child can do) often is not the same as performance of mobility during daily activities and routines (what a child does do).

Our results are based on parents' reports of their child's mobility in everyday settings from a questionnaire that was developed for this study. The reliability and validity of the data obtained with the parent questionnaire used in this study were not examined. Reliability and validity of parent reports have been demonstrated elsewhere, (21) especially when using the recognition format used in this study. (22) We believe parent reports are a feasible and practical way to collect information about a child's performance across all settings, (20) especially because direct observation of a child's usual mobility is not always realistic. Parent reports might differ from measuring mobility via direct observation, especially in the school setting where parents may be least familiar with their child's performance. Direct observation at one point in time, however, may not reflect usual mobility methods as accurately as parent reports, which, in theory, are based on routine, ongoing involvement with the child.

For children who were capable of crawling, a variety of mobility methods in addition to crawling were performed, especially at school and in the outdoors or community. The variability may reflect the impracticality of crawling in the school and outdoors or community settings. In the school and the outdoors or community, various contextual features (eg, surfaces, distances, time constraints, social expectations) may contribute to the preference for using a wheelchair or being pushed by an adult.

For children who were capable of walking with support, several contextual features may explain differences between capability and performance across everyday settings. For each setting, between 33% and 59% of the children did not perform walking with support. At home, 38% of the children performed rolling, creeping, or crawling. In the outdoors or community, 24% of the children were pushed by an adult, 18% used a battery-powered wheelchair, and 12% used a regular wheelchair. Differences in the mobility method usually used may reflect the difficulty in using mobility aids (eg, walkers) at home (with small spaces and shorter distances) and in the outdoors or community (with varied surfaces and longer distances). In addition, when faced with time constraints, children may choose mobility methods other than walking with support. At home, floor mobility may be faster than walking with support. In the outdoors or community, using a wheelchair may be a faster or more efficient mobility method than walking with support. Social expectations also may influence performance of usual mobility. For instance, social expectations of age-appropriate mobility methods may differ between home and school settings. At home, because of less peer interaction, an older child may feel comfortable using mobility methods such as rolling or crawling. In the school setting, however, walking with support is a more age-appropriate mobility method.

For children who were capable of walking alone, between 81% and 95% of children performed walking alone across settings. The results indicate that most of the children walked alone at home, a lesser number of children walked alone at school, and the least number of children walked alone in the outdoors or community. Some children who were capable of walking alone either walked with support (eg, walking aid, adult hand) or were pushed by an adult (in a stroller or wheelchair), especially in the outdoors or community. This finding may reflect the greater contextual demands in the outdoors or community (eg, varied surfaces, greater distances, time constraints, safety issues).

We attempted in this study to choose items from the GMFM (capability measure) that closely corresponded to mobility methods described in the parent questionnaire (performance measure). Some of the GMFM items, however, may not have been directly equivalent to the mobility methods described in the parent questionnaire. Some children, therefore, may have performed a mobility method in everyday settings, even though they did not pass the corresponding GMFM item. For example, some children did not pass GMFM item 44 ("crawling from a 4-point position"), but performed rolling, creeping, and crawling in everyday settings. In addition, a few children did not pass GMFM item 68 ("walks with one-hand support"), but performed walking with support in some settings. These children generally used a walking aid, such as a walker, that supports both upper extremities. The task of walking with one-hand support (capability measure) may not have been directly comparable to walking with 2-hand support (performance measure).

Testing children without their orthoses may not represent the child's capability for gross motor function. We defined capability according to scores obtained on the GMFM, without using mobility aids or orthoses. We did this in order to maintain consistency among children and among different GMFM items. For some children, however, capability without the use of orthoses is different from their capability with orthoses. For instance, a few children were not able to pass GMFM item 70 ("walking alone") without orthoses, but were able to pass this item when tested with orthoses. This finding should be considered when measuring capability in clinical practice because testing children without their usual orthoses may not accurately represent their capability of gross motor function. Testing the child's "natural" capability before examining the impact of an intervention (such as an orthosis), however, may be a useful common starting point for all examinations.

Measures of capability, such as the GMFM, are often recommended for evaluation of interventions following, for example, dorsal rhizotomy surgery, (31) intrathecal baclofen, (32) or botulinum toxin injections. (33,34) Standardized criteria (ie, instructions, testing environment, and scoring) theoretically permit direct comparison of scores among children. Furthermore, clinical observations during the administration of standardized measures can provide valuable information on how the child moves, which can be used to identify potential areas (eg, range of motion, balance) to address with interventions. Although measures of capability can provide information on the child in a controlled clinical setting, we contend there are situations where measures of performance are necessary.

Our results suggest that physical therapists should examine performance in the settings that are important to the child's daily life. The variation in performance across settings is indicative of the extent to which performance is context dependent. Examining performance in everyday settings enables the therapist not only to observe the child's ability, but also to assess environmental features that either facilitate or hinder mobility. This information, we believe, is essential for making decisions on whether interventions should focus on the child's motor ability, modification of the task, or adaptation of the environment. In addition, because there are differences in the settings where children function, the measurement of performance appears to be necessary, when comparing a child with his or her own performance (eg, comparing the effects of intervention over time or across settings).

Knowledge of the difference between capability and performance has implications for decision making in physical therapy. By comparing a child's capability and performance, appropriate goals can be set to maximize the child's performance in everyday settings. For instance, if a child is capable of walking alone, but performs walking with support at school, an intervention outcome may include walking alone at school. The factors contributing to the discrepancy between capability and performance, we believe, should be identified and addressed in physical therapy intervention. Certain aspects of the person-environment interaction (Fig. 1) may clarify which features leg, personal factors, contextual features) may be influencing performance. Interventions may include adapting the child's environment, working toward improvement in the child's functional ability, or modifying the task. Physical therapists often use adaptive equipment to facilitate the child's performance in everyday settings. In our study, children with relatively low capability used mobility aids to maximize performance. Some children in this study who were not capable of crawling used battery-powered wheelchairs, and therefore utilized an independent mobility method in the home, at school, and in the outdoors or community. Other children who were not capable of crawling were pushed by an adult and therefore were reliant on others for mobility. By addressing the factors that contribute to a child's performance, physical therapists can focus on the most efficient and functional methods of mobility in everyday settings.

Conclusion

Our results provide evidence of differences between capability and performance of mobility for a sample of children with CP. The performance of mobility methods varied across the home, school, and outdoors or community settings. Although capability is generally measured in a clinical setting, performance of mobility in daily life involves different environmental settings, with each setting having unique contextual features. Physical therapists, in our view, may use standardized tests to examine capability, to reflect what children can do in a controlled clinical setting, but these tests may have limited generalizability to everyday performance. Performance, we argue, may be measured via systematic observation, parent questionnaires, and child self-reports, and, we believe, should include an examination of the contextual features of the environment. Examination of capability and performance in children with CP, we believe, provides therapists with information that can be useful in evaluating whether interventions to improve mobility should address neuromuscular and musculoskeletal impairments, modification of mobility method, or accommodations in the environment.

Further research is necessary to understand the contextual features that affect mobility and the relationship between capability and performance. Contextual features, we contend, should be examined specific to each setting to explain the differences in performance that occur across the home, school, and outdoors and community settings. Understanding the environmental factors (physical, social, and attitudinal) and personal factors contributing to the performance of mobility may enable physical therapists to more effectively identify preferred interventions to improve functional mobility in children with CP.

Table 1.
Gross Motor Function Classification System (GMFCS) Level,
Distribution of Cerebral Palsy (CP), and Type of CP for
Children in Each Capability Group

                       Children Unable   Children Capable
                       to Crawl          of Crawling
                       (n=99)            (n=40)

                       N    %            N    %

GMFCS
  I                     0                 0
  II                    1    1            1    2
  III                   3    3           21   53
  IV                   41   41           17   43
  V                    54   55            1    2

Distribution
  Quadriplegia         81   82           20   50
  Triplegia            11   11            7   17
  Diplegia              6    6           13   33
  Hemiplegia            1    1            0

Type
  Spastic              69   70           30   75
  Dystonic/athetotic    9    9            4   10
  Hypotonic             7    7            3    8
  Ataxic                0                 1    2
  Mixed                14   14            2    5
  Not applicable        0                 0

                       Children Capable
                       of Walking         Children Capable
                       With Support       of Walking Alone
                       (n=34)             (n=134)

                       N    %             N     %

GMFCS
  I                     0                  95   71
  II                    4   12             33   25
  III                  26   76              6    4
  IV                    4   12              0
  V                     0                   0

Distribution
  Quadriplegia          7   21             15   11
  Triplegia             8   23              9    7
  Diplegia             17   50             67   50
  Hemiplegia            2    6             43   32

Type
  Spastic              27   79            111   83
  Dystonic/athetotic    1    3              4    3
  Hypotonic             1    3              4    3
  Ataxic                2    6              6    4
  Mixed                 3    9              8    6
  Not applicable        0                   1    1

Table 2.
Operational Definitions for Gross Motor Capability

                     GMFM (a)   Item Difficulty (b)

Capability           Item       [bar.X]   SE

Crawling             44         42.44     0.47
Walks with support   68         49.15     0.47
Walks alone          70         57.39     0.53

Capability           Description of Passing Score (Score of 3)

Crawling             Crawling. From a 4-point position, child
                       crawls (moves on hands and knees) or hitches
                       (moves forward using arms and/or legs while
                       maintaining some variation of sitting, may
                       include bunny hopping or bottom hitching)
                       forward 1.8 m (6 ft).
Walks with support   Walks forward with one hand held. From a
                       standing position with one hand held, child
                       walks forward 10 steps.
Walks alone          Walks forward, arms free. From a standing
                       position, child walks forward 10 steps,
                       stops, turns 180[degrees], and returns.

(a) GMFM = Gross Motor Function Measure.

(b) The mean (with a maximum score of 100) and standard error
(SE) are based on Rasch item response analysis. (19)

Table 3.
Performance of Mobility Methods in the Home, at School, and in
the Outdoors or Community for Children in Each Capability Group

              Children Capable of Crawling (n=40)

              Perform Rolling,   Do Not Perform
              Creeping, or       Rolling, Creeping
              Crawling           or Crawling

Home          27 (67%)           13 (33%)
                                 CI: (a) 20%-48%

School         1 (2%)            39 (98%)
                                 CI: 86%-100%

Outdoors or    0                 40 (100%)
  community                      CI: 91%-100%

              Children Capable of Walking With
              Support (n=34)

                                Do Not Perform
              Perform Walking   Walking With
              With Support      Support

Home          15 (44%)          19 (56%)
                                CI: 39%-71%

School        23 (68%)          11 (32%)
                                CI: 19%-49%

Outdoors or   14 (41%)          20 (59%)
  community                     CI: 42%-74%

              Children Capable of
              Walking Alone (n=134)

              Perform
              Walking      Do Not Perform
              Alone        Walking Alone

Home          127 (95%)    7 (5%)
                           CI: 2%-11%

School        120 (90%)    14 (10%)
                           CI: 6%-17%

Outdoors or   109 (81%)    25 (19%)
  community                CI: 13%-26%

(a) CI = 95% confidence, interval.

References

(1) Hutton JL, Cooke T, Pharoah PO. Life expectancy in children with cerebral palsy. Br Med J. 1994;13:430-435.

(2) Hayes MS, McEwen IR, Lovett D, et al. Next step: survey of pediatric physical therapists' educational needs and perceptions of motor control, motor development and motor learning as they relate to services for children with developmental disabilities. Pediatric Physical Therapy. 1999;11:164-182.

(3) Palmer GB, Shapiro BK, Wachtel RC, et al. The effects of physical therapy on cerebral palsy: a controlled trial in infants with cerebral palsy. N Engl J Med. 1988;318:803-808.

(4) Calderon-Gonzalez R, Calderon-Sepulveda R, Rincon-Reyes M, et al. Botulinum toxin A in management of cerebral palsy. Pediatr Neurol. 1994;10:284-288.

(5) Steinbok P, Reiner AM, Beauchamp R, et al. A randomized clinical trial to compare selective posterior rhizotomy plus physiotherapy with physiotherapy alone in children with spastic diplegic cerebral palsy. Dev Med Child Neurol. 1997;39:178-184.

(6) McLaughlin JF, Bjornson KF, Astley SJ, et al. Selective dorsal rhizotomy: efficacy and safety, in an investigator-masked randomized clinical trial. Dev Med Child Neurol. 1998;40:220-232.

(7) Palisano RJ, Tieman BL, Walter SD, et al. Effect of environmental setting on mobility methods of children with cerebral palsy. Dev Med Child Neurol. 2003;45:113-120.

(8) Berry ET, McLaurin SE, Sparling JW. Parent/caregiver perspectives on the use of power wheelchairs. Pediatric Physical Therapy. 1996;8: 146-150.

(9) Haley SM, Coster WJ, Binda-Sundberg K. Measuring physical disablement: the contextual challenge. Phys Ther. 1994;74:443-451.

(10) Tieman BL, Palisano RJ, Gracely EJ, et al. Changes in mobility of children with cerebral palsy over time and across environmental setting. Physical and Occupational Therapy in Pediatrics. In press.

(11) Young NL, Williams JI, Yoshida KK, et al. The context of measuring disability: does it matter whether capacity or performance is measured? J Clin Epidemiol. 1996;49:1097-1101.

(12) Brown M, Gordon WA, Diller L. Rehabilitation indicators. In: Halpern AS, Fuhrer MJ, eds. Functional Assessment in Rehabilitation. Baltimore, Md: Paul H Brookes Publishing Co Inc; 1984:187-203.

(13) International Classification of Functioning, Disability, and Health (ICF). Geneva, Switzerland: World Health Organization; 2001.

(14) Law M, Cooper B, Strong S, et al. The person-environment-occupation model: a transactive approach to occupational performance, Can J Occup Ther. 1996;63:9-23.

(15) Russell DJ, Rosenbaum PL, Avery LM, Lane M. The Gross Motor Function Measure (GMFM-66 and GMFM-88) Users' Manual. London, United Kingdom: MacKeith Press; 2002.

(16) Rosenbaum PL, Walter SD, Hanna SE, et al. Prognosis for gross motor function in cerebral palsy: creation of motor development curves. JAMA. 2002;288;1357-1363.

(17) Palisano RJ, Rosenbaum PL, Walter SD, et al. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dew Med Child Neurol. 1997;39:214-223.

(18) Bax MCO. Terminology and classification of cerebral palsy. Dev Med Child Neurol. 1964;6:295-297.

(19) Russell DJ, Avery LM, Rosenbaum PL, et al. Improved scaling of the Gross Motor Function Measure for children with cerebral palsy: evidence of reliability and validity. Phys Ther. 2000;80:873-885.

(20) Long TM. The use of parent report measures to assess infant development. Pediatric Physical Therapy. 1992;4:74-77.

(21) Wilson BN, Kaplan BJ, Crawford SG, et al. Reliability and validity of a parent questionnaire on childhood motor skills. Am J Occup Ther. 2000;54:484-493.

(22) Glascoe FP, Dworkin PH. The role of parents in the detection of developmental and behavioral problems. Pediatrics. 1995;95:829-836.

(23) Russell DJ, Rosenbaum PL, Cadman DT, et al. The Gross Motor Function Measure: a means to evaluate the effects of physical therapy. Dev Med Child Neurol. 1989;31:341-352.

(24) Trahan J, Malouin F. Changes in the Gross Motor Function Measure in children with different types of cerebral palsy: an eight-month follow-up study. Pediatric Physical Therapy. 1999;11:12-17.

(25) Bjornson KF, Graubert CS, Buford VL, McLaughlin JF. Validity of the Gross Motor Function Measure. Pediatric Physical Therapy. 1998;10: 43-47.

(26) Bjornson KF, Graubert CS, McLaughlin JF, et al. Test-retested reliability, of the Gross Motor Function Measure in children with cerebral palsy. Physical & Occupational Therapy in Pediatrics. 1998;18(2): 51-61.

(27) Russell DJ, Rosenbaum PL, Raina PS, et al. Training users in the use of the Gross Motor Function Measure: methodological and practical issues. Phys Ther. 1994;74:630-636.

(28) Agresti A, Coull BA. Approximation is better than "exact" for interval estimation of binomial proportions. Am Stat. 1998;52:119-126.

(29) Huck SW. Reading Statistics and Research. 3rd ed. New York, NY: Longman; 2000.

(30) Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. 2nd ed. East Norwalk, Conn: Appleton & Lange; 2000.

(31) Steinbok P. Outcomes after selective dorsal rhizotomy for spastic cerebral palsy. Childs Nerv Syst. 2001;17:1-18.

(32) Almeida GL, Campbell SK, Girolami G, et al. Multidimensional assessment of motor function in a child with cerebral palsy following intrathecal administration of baclofen. Phys Ther. 1997;77:751-764.

(33) Mall V, Heinen F, Kirschner J, et al. Evaluation of botulinum toxin: a therapy in children with adductor spasm by gross motor function measure. J Child Neurol. 2002;15:214-217.

(34) Kirschner J, Linder M, Berweck S, et al. Treatment of pes equines in cerebral palsy with botulinum toxin A and functional benefit according to gross motor function measure. Eur J Neurol. In press.

BL Tieman, PT, PhD, is Assistant Professor, Department of Physical Therapy, Georgia State University, MSC 8L0388, 33 Gilmer St SE, Unit 8, Atlanta, GA 30303-3088 (USA) ([e-mail]). Address all correspondence to Dr Tieman.

RJ Palisano, PT, ScD, is Professor, Programs in Rehabilitation Sciences, Drexel University, Philadelphia, Pa, and Co-investigator, CanChild Centre for Childhood Disability Research, McMaster University, Hamilton, Ontario, Canada.

EJ Gracely, PhD, is Associate Professor, Department of Family, Community, and Preventive Medicine, Drexel University.

PL Rosenbaum, MD, FRCP(C), is Professor of Paediatrics, McMaster University, Co-Director, CanChild Centre for Childhood Disability Research, and Canada Research Chair in Childhood Disability, Ontario, Canada.

Dr Tieman provided concept/idea/research design. Dr Tieman, Dr Palisano, and Dr Rosenbaum provided writing. All authors provided data analysis. Dr Palisano and Dr Rosenbaum provided project management and fund procurement. The authors thank Lisa A Chiarello, PT, PhD, PCS, and Margaret E O'Neil, PT, PhD, MPH, for their dissertation committee work.

The ethics review boards of Hamilton Health Sciences Corp, the Bloorview MacMillan Centre (Toronto, Ontario, Canada), and the Thames Valley Children's Centre (London, Ontario, Canada) approved the study.

This work was supported by grant MCJ429391 from the Department of Health and Human Services, Maternal and Child Health Bureau, awarded to Drexel University; grant MT-13476 from the Canadian Institutes of Health Research (formerly the Medical Research Council of Canada); and grant RO1-HD-34947 from the National Center for Medical Rehabilitation Research of the National Institute of Child Health and Human Development.

This research was presented as platform presentations at the American Academy of Cerebral Palsy and Developmental Medicine (AACPDM), New Orleans, La, September 13, 2002, and the American Physical Therapy Association Combined Sections Meeting, Tampa, Fla, February 13, 2003.

This article was received July 22, 2003, and was accepted October 29, 2003.