- Top of page
Aim Parental report is often relied on to measure performance of activities in children with cerebral palsy (CP). This study examined whether the Functional Mobility Scale (FMS) accurately reflects performance of mobility in children with CP.
Method Eighteen children with spastic CP (11 males, seven females; mean age 12y 8mo, SD 2y 8mo, range 8–17y) were recruited from a special development school. Children were in Gross Motor Function Classification System (GMFCS) levels II (n=5), III (n=4), or IV (n=9), and had quadriplegia (n=9), diplegia (n=7), or hemiplegia (n=2). The children’s mobility was observed directly around and outside the home and at school and their mobility methods were recorded. The parent’s FMS rating was obtained on the telephone by a physiotherapist. Agreement between direct observation and the FMS rating was examined using quadratic weighted kappa (κ) statistics.
Results Agreement between direct observation and the FMS was as follows: FMS 5m κ=0.71, 45%; FMS 50m κ=0.76, 94%; FMS 500m κ=0.74, 95%. Differences in the range and number of mobility methods were observed by GMFCS level across environmental settings.
Interpretation Substantial agreement was found between FMS ratings and direct observation, particularly over longer distances, providing evidence of the validity of the FMS as a measure of performance in children with CP.
Children with cerebral palsy (CP) often experience activity limitations, defined in the World Health Organization’s International Classification of Functioning, Disability and Health (ICF) as ‘difficulties in executing tasks or actions’.1 To determine how these limitations affect the lives of children with CP, reliable and valid measurement tools are required to measure change over time and after interventions. A number of outcome measures are available to measure activity limitations in children with CP.2 They are focused on different aspects of activity and have various methods of administration. For example, the Gross Motor Function Measure (GMFM)3 is used to assess gross motor function through direct observation by the therapist in the clinic. Some tools, such as the Functional Assessment Questionnaire (FAQ),4 Activities Scale for Kids (ASK),5 and Functional Mobility Scale (FMS)6, use child or parent reports to gain insight into the child’s activities away from the clinical environment. The present study is focused on the validity of the FMS as a parent-report tool for children with CP.
When selecting outcome measures it is important to consider the impact of the environmental setting and whether actual performance is being measured. Environmental factors influence health conditions such as CP, and these factors could have an impact on the performance of activities.1 For example, the mobility of children with CP has been shown to vary across the environmental settings of home, school, and community.7,8 The term performance refers to what an individual does in his or her current environment and can be understood as ‘involvement in life situations’.1 It can be thought of as what a person does do. This is in contrast to capability, which can be thought of as what a person can do, the difference being based on capacity, context, and choice.9 In considering these different concepts, the primary advantage of measures that reflect performance is that they consider the environment while providing a direct assessment of actual community function, and they measure limitations of direct relevance to children and their families.9 Once identified, these limitations can then, potentially, be addressed.
It is not always practical or feasible for clinicians and researchers to observe and assess children in their own environments; hence self- or parent-report measures are relied on to assess performance. When using these measures the clinician or researcher assumes that the information gathered accurately reflects the performance of the child; however, this is often not known. The FMS was developed as a performance measure, giving a rating of the assistance required by children with CP for mobility in the environmental settings of the home (i.e., mobility over 5m), school (50m), and community (500m).6 It is administered by a clinician in a semi-structured interview using child or parent reports. It has been shown to have good reliability, and it is able to detect change in mobility after single-event multilevel surgery.6,10
The aim of this study was to examine the validity of the FMS to determine whether parent reports of mobility accurately reflect performance in children with CP by analysing the agreement between the FMS and direct observation. The hypothesis was that the mobility status of children with CP as observed directly would agree substantially with the FMS. The study also aimed to describe the different mobility methods used by children with CP.
- Top of page
Participants were recruited through the Royal Children’s Hospital, Melbourne, Australia, and a special development school. Included were children and young people aged 4 to 18 years with a diagnosis of CP, classified in Gross Motor Function Classification System (GMFCS) levels I to IV.11 All children satisfying the inclusion criteria who attended the school were invited to participate. The school used for recruitment provides an educational environment for children with physical disabilities or health impairments, and the students have a mixture of complex physical, cognitive, and behavioural problems. For this reason few children in GMFCS level I attend the school, and there is a preponderance of children in GMFCS level IV, who benefit from access to special educational facilities in combination with specialized therapy. Children were excluded if their parents were not able to understand the study information or instructions because of language, cognitive, or other difficulties.
The study was approved by the ethics committees of the Royal Children’s Hospital Melbourne and the University of Melbourne, the school principal, and the regional department of education. A list of eligible children who attended the school and who were clients of the hospital was compiled by the school. Parents of all eligible children were then sent letters explaining the study and inviting them to participate. Returned signed consent forms designated the children to be included.
There were three stages of data collection: obtaining an FMS rating with a parent by telephone, direct observation of the child at school, and direct observation of the child at home and outside the home.
An experienced physiotherapist not involved in the study with no previous knowledge of the mobility status of the children performed the FMS rating with the parents by telephone. The rating was performed and recorded according to usual administration of the scale;6 that is, the physiotherapist asked a few simple questions of the parent about the child’s usual method of mobility around the home (FMS 5m), at school (FMS 50m), and in the community (FMS 500m). A rating of 1 to 6 was assigned for each setting according to the assistance required, ranging from independent on all surfaces (rated 6) to use of a wheelchair (rated 1). A rating of C was assigned if the child crawled or N if the distance was not covered. When administering the FMS the clinician records what the child does do, rather than what they know or believe they can do, to reflect performance. This rater had no further involvement in the study.
Direct observation of the children at school was performed by the first author (AH) who attended the school for one school day, observed all children, and recorded performance without disturbing the usual routines. The children were observed at different times moving around the various environments of the school, including from the school bus to the classroom, around the classroom, between classes, and around the playground. The methods of mobility that the children used in each of these settings were recorded. These included powered or manual wheelchair, a range of walkers, crutches, sticks, holding hands or walls, independent walking without assistive devices on level surfaces, independent walking without assistive devices on all surfaces, being carried, and crawling.
Direct observation of the children at home was carried out by one of two physiotherapists from the school (JH and MH). These observers were used because the parents were familiar with them and were more likely to accept them in the home. A second reason was that the school physiotherapists perform home visits from time to time, so the children were more likely to perform as they would normally rather than if an unfamiliar clinician were to observe them at home. Similarly to the school visit, the home observers recorded the methods of mobility that the children chose to use. They observed the children moving within one room, between rooms, and outside the home for longer distances where they went for a ‘walk’ around the block or down the street.
Each child was rated once on the FMS by telephone and observed once at school and once at home. For all three stages of data collection the various observers and raters were unaware of previous information collected.
An observed FMS score was generated by a person blinded to the information gathered after all school and home observations had been recorded. To choose one mobility method for each environmental setting, the ‘within one room’ category for the home visit was used for the FMS 5m, the ‘between classes’ category for the school visit for the FMS 50m, and the ‘outside the home’ category for the FMS 500m. The observed FMS score was compared with the telephone FMS score for data analysis. All methods of mobility used were recorded to provide descriptive details of the number and variety of methods used in each environment and the number of methods used by each child.
Descriptive statistics for participant characteristics were calculated as means and standard deviations for continuous variables and as frequency and percentages for categorical variables. Descriptive statistics were used to illustrate the number and variety of mobility methods used by each child and by the group as a whole.
Because of the categorical nature of the data, as well as percentage agreements, quadratic-weighted κ values and 95% confidence intervals were calculated to analyse agreement between the telephone FMS and the observed FMS.12 Guidelines for the range of κ values were used to describe the strength of agreement, whereby values of 0.00 to 0.20 indicate slight agreement, 0.21 to 0.40 fair agreement, 0.41 to 0.60 moderate agreement, 0.61 to 0.80 substantial agreement, and 0.81 to 1.00 almost perfect agreement.13 Data were stored and organized using the EpiData for Windows program (http://www.epidata.dk/). All analyses were preformed using Stata statistical software, release 9.0 (StataCorp, College Station, TX, USA).
- Top of page
Twenty-seven children were eligible for the study, and parents of 18 of those gave written consent for their child to participate; there were 11 males and seven females. The age range was 8 to 17 years (mean 12y 8mo, SD 2y 8mo). Five children were in GMFCS level II, four were in level III, and nine were in level IV. Nine children had spastic quadriplegia, seven had spastic diplegia, and two had spastic hemiplegia. There was no difference in demographics between children whose parents provided consent and children whose parents did not consent.
The mean time between FMS ratings performed by telephone and the home observation was 10.2 days (SD 5.44) with a range of 0 to 17 days. The mean time between FMS ratings and the school observation was 20.9 days (SD 6.36) with a range of 14 to 40 days.
Agreement rates between the parent-reported and observed FMS ratings are presented in Table I. The weighted κ values show substantial agreement for all three environmental settings. Percentage agreement was highest for the FMS 50m (school) and 500m (community) ratings. The actual ratings are shown in Table II.
Table I. Agreement between parent-reported and observed Functional Mobility Scale (FMS) scores by environmental setting
| ||FMS 5m||FMS 50m||FMS 500m|
|Weighted kappa (CI)||0.71 (0.41, 1.0)||0.76 (0.31, 1.00*)||0.74 (0.33, 1.00a)|
Table II. Mobility ratings for each of the environmental settings, on the parent-reported (telephone) and observed Functional Mobility Scale (FMS) assessments
| ||Observed FMS rating|
|FMS 5m (home)|
|Telephone FMS rating||1||3||1||0||1||1||0||0||6|
|FMS 50m (school)|
|Telephone FMS rating||1||6||2||0||0||0||0|| ||8|
|FMS 500m (community)|
|Telephone FMS rating||1||13||0||1||1||0||0|| ||15|
Direct observation showed differences in the range and number of mobility methods used within the home, school, and community. Within the home the most frequent mobility method was independent walking, followed by crawling and wheelchair use. At school the most frequent method was wheelchair use, followed by walking. Outside the home, wheelchair use was the most frequent method.
Overall, more than one mobility method was used by seven children at home and by 13 children at school. Outside the home for longer distances all children were observed to use only one method. There were some subtle differences in the number of mobility methods in each setting according to the children’s GMFCS levels (Table III). Children in GMFCS level II used mainly one method at home and two methods at school. Children in GMFCS level III used mainly one method at home and three methods at school. Children in level IV used between one and two methods at home and two methods at school.
Table III. Differences in the number of mobility methods used in each setting according to Gross Motor Function Classification System (GMFCS) level
| 1 method used||4||2||5|
| 2 methods used||1||3||0|
| 3 methods used||0||0||0|
| 1 method used||3||0||4|
| 2 methods used||0||1||0|
| 3 methods used||1||3||0|
| 1 method used||5||3||9|
| 2 methods used||4||6||0|
| 3 methods used||0||0||0|
| 1 method used||12||5||18|
| 2 methods used||5||10||0|
| 3 methods used||1||3||0|
- Top of page
This study found substantial agreement between parent reports of mobility obtained from the FMS and direct observation of mobility in children with CP. The large confidence intervals around the κ values, presumably due to the small sample size, preclude any conclusions to be drawn about whether agreement between the three distances was significantly different. Despite this, the trend was for better agreement for the longer distances.
Although the agreement was substantial,14 no κ value was above 0.76. Possible reasons for disagreement include the variety of mobility methods observed, issues surrounding parental reporting, and the concept being measured. The small numbers in the study may also have contributed to lower agreement. Many children were observed to use more than one mobility method at home and at school. This variety may have created difficulties for clinicians administering the tool and for the parent reporting. It is potentially challenging to choose one method for each setting when in reality children often use more than one. For example, one male child aged 12 years 3 months, in GMFCS level III, walked independently around his home and at school in the classroom. He used crutches from class to class at school and a wheelchair in the playground and from the bus to the classroom. Consequently he used three methods at school. Considering the variety of methods used by many of the children in this study, it is possible that observation on a different day may have provided different results. However, the agreement results in this study are robust in the light of this potential source of variation.
The children in this sample displayed a range of mobility methods in different environmental settings. There was variety within the group by GMFCS levels and variety within individual children. These findings support earlier observations that children use a range of methods across settings, particularly those in GMFCS levels II to IV.7,8 Similarly to the findings by Tieman et al.8, this study found that the most common methods were independent walking or crawling at home, walking with a walking aid or wheelchair use at school, and wheelchair use in the community. In the current study all forms of mobility observed in each setting were recorded, rather than just one ‘preferred’ method as reported by Tieman et al.8 and Palisano et al.7 The current study also differentiated between various types of walking aids rather than grouping them together as a single category.
A number of factors may influence a child’s choice of mobility method in different settings. The home environment is more familiar, with shorter distances to travel. Children are likely to need less assistance there or may choose crawling if that allows them to be independent. Safety and distances, as well as liability issues, within the school environment may potentially influence mobility methods used there. It is possible that more supportive devices or wheelchairs are used to ensure safety, even if the child could use different mobility methods. Convenience and time-saving factors might also influence mobility choices. Many parents reported that they use the wheelchair in the community to save time, for behavior management, or if the child tires easily. Some families reported that they leave the walker at school for convenience and use the wheelchair for longer distances.
This study highlights some important issues around the use of parent- or self-report measures. As with other parent- or self-report scales such as the FAQ,4 ASK,5 CHQ15 and the Pediatric Outcomes Data Collection Instrument,16 it is assumed that what is reported reflects what is truly happening. The reliability, but not the validity, of this method of reporting has been studied: for example, high intraclass correlation coefficients were found in GMFCS levels between clinician ratings and family-completed questionnaires in a study involving 97 children aged 6 to 11 years.17 The reliability of child reports was examined in 28 children during the development of the ASK18 with high retest reliability. These studies did not address the agreement between parent or self reports and direct observation, i.e. the validity of the information collected with the measures. Without this evidence it is uncertain whether these measures are reporting what they are intended to report.
The FMS is intended to measure performance rather than capability. It is possible in this study that parents reported capability, particularly if they were keen to emphasize their child’s best ability. Studies have shown differences between capability and performance in children.9,19 Children with similar capabilities have been shown to demonstrate differences in performance,19 and the same children have been shown to score consistently higher for capability than performance of the same activities.9 The FMS is administered by a clinician asking a few questions of the child or parent. The way the questions are worded is integral to whether or not performance is reported.
Performance reflects function relevant to children and families;9 thus parental or self reporting remains the most feasible and practical method of measuring it. Despite observations in daily life potentially being more valid, it is impractical for clinicians and researchers to observe every child within their environmental settings. Capacity may also be important for clinicians to measure, particularly after interventions. However, interventions, measurements, and outcomes should always consider what is important for the child and family in their own reality.
The limitations of this study include the small sample size, the variation in time between ratings, and the limited recruitment process, although these issues do not invalidate the findings. A larger sample size might have resulted in higher agreement, with smaller confidence intervals. The time between the telephone FMS rating and observations at home and school ranged between 0 and 40 days. This could have allowed change in mobility status of some of the children, thus providing a source of variation, although given the age of the children this is unlikely. An interval of 2 to 14 days is suggested as most appropriate.20 Only one school was used to recruit participants, and this school had a relatively high proportion of children classified in GMFCS level IV. The results are, therefore, not generalizable to the whole population of children with CP. Future research should examine this issue using larger numbers representing a more even spread of GMFCS levels from a range of schools.