Randomization and Interventions
After the first author identified a matched pair, she notified the second author who numbered the children and then generated 1 of the numbers using a random number generator. The child whose number appeared was assigned to the experimental group and the match was assigned to the control group.
A research assistant, who was blinded to the children's group assignment, gathered pretest data in the families' homes. Children in both groups continued to receive early intervention services, as outlined by their individualized family services plans. Project staff contacted families in the control group by phone each month for the duration of the study to answer any questions and to verify continued participation in the study. For children in the experimental group, the first author delivered a power wheelchair to the children's homes and provided initial fitting and training. All children showed a beginning understanding of how to make the chair “go” after the initial fitting and training. Individually customized Invacare Power Tiger wheelchairs (Invacare Corporation, One Invacare Way, Elyria, OH) were used without charge to the families, with funding from a grant from the United States Department of Education. Most of the children (10) used a hand to control the joystick on the chair. Other control mechanisms included proximity switches (3) and a head array (1). We provided ramps, when necessary to enable the child to get in and out of the home. None of the families had wheelchair-accessible vans.
The power mobility training was based on our own experiences,8 reports by Butler,10,16 and recommendations by Kangas.18 Training methods were consistent with the conclusions of a critical review of training young children to use power mobility that was published after the completion of our study.17 We provided written guidelines (Appendix 1) to the children's parents and their early intervention therapists. Parents were primarily responsible for providing practice opportunities and instruction, with therapists and research staff helping to solve any problems. We asked parents to (1) provide the child with daily opportunities to sit in the device with the motor turned on during play; (2) encourage the child to experiment with movement in a relatively large space and not be concerned if the child moved in circles; and (3) avoid telling the child what to do, but rather to let the child experiment unless frustrated or unsafe. We stressed the importance of parental supervision, as one would supervise any young child.
To measure the children's developmental and functional abilities at the beginning of the study and 12 months later, we used (1) the Battelle Developmental Inventory (BDI),19 (2) Pediatric Evaluation of Disability Inventory (PEDI),20 and (3) The Early Coping Inventory (ECI).21 We collected all data in the children's homes. The BDI is designed to evaluate development in 5 domains compared with children developing typically: cognitive, adaptive (self-help), motor, communication, and personal/social development. Scoring consists of rating each item as 0, 1, or 2, based on direct observation of children or through parent interview. Following a review of studies that evaluated or used the BDI, Berls and McEwen22 concluded that the BDI has good psychometric properties and value for measuring developmental change in longitudinal studies with young children.
The PEDI is designed to measure children's functional abilities in the mobility, self-help, and social domains. It is one of the few tests designed for children with disabilities. The PEDI is a judgment-based assessment, meaning that children's skills and behaviors need not be directly tested, but can be reported by someone who knows the child well, such as a parent. In addition to measuring functional skills, the PEDI measures the amount of caregiver assistance required for tasks and the modifications a child requires. It can, therefore, show change through a decrease in caregiver assistance even if the child cannot perform a task independently. Inter–interviewer and inter–respondent reliability estimates for the PEDI are more than 0.90.20 Studies have supported its use as an evaluative measure that is responsive to change.23–25
The ECI is a criterion-referenced measure for children with and without disabilities aged 4 to 36 months. It has 3 scales: sensorimotor organization (behaviors used to regulate psychophysiologic functions and to integrate sensory and motor processes), reactive behaviors (responses to demands of the social and physical environments), and self-initiated behaviors (self-directed actions used to meet personal needs and to interact with objects and people). The scales can be used independently of each another. Because we did not expect changes in sensorimotor organization, we only used the reactive behavior scale and the self-initiated behavior scale. Scoring consists of rating each available item on a scale from 1 to 5 by someone who knows the child well, such as a parent. The first author also measured the children's competence in maneuvering the power wheelchairs by achievement of the skills that Butler et al26 used in their study (Table 3).
Reliability of Measurements
Two research assistants, a pediatric physical therapist with 10 years of experience and a pediatric occupational therapist with 4 years of experience, interviewed the families to score the PEDI and ECI and completed the BDI using a combination of observation and interview procedures and the general adaptations and supplementary materials for children with disabilities, as described in the testing manuals. The first author, a pediatric physical therapist with 13 years of experience, was the second rater for reliability estimates for 10% of the study sample. We calculated percent agreement and weighted κ27 for all domains of the BDI, ECI, and PEDI. Overall percent agreement for the BDI was 96.6% and κ was 0.97. Percent agreement for individual domains ranged from 94.7% to 98.1% and κ scores ranged from 0.95 to 0.99. For the ECI, overall percent agreement was 91.8% and weighted κ for individual domains ranged from 0.95 to 0.97. For the PEDI, overall percent agreement was 98.0%, weighted κ for the functional skills domains were 1.0, and weighted κ for the caregiver assistance domains ranged from 0.85 to 0.99.
Comparison of the participants' standard scores for the BDI and PEDI from baseline to 12 months indicated that their development was either not changing or they became further behind over time compared with their same age typical peers. Because age equivalent scores on the BDI and scaled scores on the PEDI are not adjusted for age and an increase in score represented increased performance, we used these scores for analyses because they were more likely to show individual change over time. We used mean scores on the ECI. Because the small sample size increased risk of a type II error, the a priori α-level was .10. The risk of a type II error was greater than the risk of a type I error, and could lead to disregarding potentially useful findings with a lower α-level.28
We used baseline scores to examine differences in the 2 groups at baseline and difference (change) scores (posttest minus pretest scores) for each child to compare the 2 groups. We used exact nonparametric methods because some of our data did not meet parametric assumptions and contained outliers, and because of our small sample size. Because the matched pairs design produced correlated data, we used the Wilcoxon signed rank test to compare the change scores between the experimental and control groups. Because 3 children in the experimental group discontinued use of the power wheelchairs before the end of the study, we did both an on-protocol analysis and an intention to treat analysis (ITT). An on-protocol analysis includes only the data of participants who complete the study according to the protocol.27 An ITT analysis includes all available data regardless of whether participants complete the study as planned.27 Our on-protocol analysis included data of the 11 matched pairs for whom we had baseline and 12-month data and the children in the experimental group used the wheelchairs for the year-long period. The ITT analysis included the 3 additional pairs for whom we had baseline data and 12-month data, but the children in the experimental group stopped using the wheelchairs before the end of the year.
We also calculated differences in the groups' medians (effect size) and 90% confidence intervals for all comparisons using exact Hodges-Lehman estimates. We performed analyses using either SAS version 9.1.329 or, for exact nonparametric procedures, Cytel Studio 9.30
Table 4 summarizes baseline data for experimental and control groups in both the intention to treat and the on-protocol samples. The groups did not differ on any of the measures at baseline. Table 5 shows the results of the on-protocol analysis and Table 6 shows the results of the ITT analysis.
When comparing the 2 groups' median change scores in the on-protocol sample (Table 5), the experimental groups' BDI receptive communication score increased more than the control group's median change scores, as did their PEDI mobility functional skills, mobility caregiver assistance, and self-care caregiver assistance median change scores. Other median change scores on the BDI and PEDI, and all scores on the ECI did not differ. The ITT analysis (Table 6) upheld the results of our on-protocol analysis, with an additional significant difference between the median change in BDI total scores, with the experimental groups' scores improving more than the control group's scores.
Of the 7 skills in Table 3, which we used to measure competence in maneuvering a wheelchair, all the children learned to drive in circles and start and stop on command. Four children mastered all 7 skills in 12 to 42 weeks. The other 7 children mastered between 2 and 6 skills. All but 2 children could move forward 10 feet in wide areas, requiring 2 to 34 weeks to learn. Only the 4 children who mastered all 7 skills could move forward 10 feet in a narrow area. No adverse events, such as accidents or injuries, occurred as part of this study.
The on-protocol analysis showed that the median change scores from baseline to 12 months increased significantly more in the group that used power wheelchairs than in the control group for BDI receptive communication, and PEDI mobility functional skills, mobility caregiver assistance, and self-care caregiver assistance. The results of the intention to treat analysis were similar, with 1 addition: BDI total scores of children who used power wheelchairs also increased significantly more than scores of children in the control group. Because the differences all favored the group of children who used power wheelchairs, we believe the study provides support for the use of power mobility to enhance the development of young children with severe motor limitations. The study also showed that some children as young as 14 months can begin to use power mobility, but the time to become proficient in its use varied among children, ranging from 12 weeks to 42 weeks.
The study had several unanticipated findings. First, the length of time children required to become proficient in maneuvering the power wheelchairs was longer than we expected. Previous reports suggested that young children could learn within a few weeks,8,10 but the children in our study required at least 12 weeks, and 7 of the 11 children had not mastered all of the skills by the end of 1 year. Because our inclusion criteria were quite broad, the possibility exists that the children had greater sensorimotor and cognitive impairments than children in the previous reports, which may have contributed to the time required to acquire the wheelchair skills. The possibility also exists that children might learn to use a robotic device, which Lynch et al,11 described, more quickly than they can learn to use a power wheelchair.
More intense training in a structured and controlled environment also might facilitate more rapid wheelchair skill acquisition. The children's families were primarily responsible for providing training opportunities by incorporating practice into daily and routine activities, which also was true for other studies identified in a review of evidence for teaching young children to use power mobility.17,31 Other investigators have provided structured training for infants to use a powered device11 and structured and controlled wheelchair practice experiences with older children12,13 with some success. Infants who are developing typically while learning to walk practice for more than 6 hours throughout a day and by the end of a day may have taken 9000 steps and traveled the length of 29 football fields.32 Therefore, the next study our research team designed incorporated initial intensive practice in an effort to allow children to acquire the skills necessary for independent mobility more rapidly than the children in this study. Although we requested that parents keep a power mobility log to track how long, where, and when children used the power wheelchairs, the data we received from the logs were inconsistent. On the basis of the limited data we have, the average amount of time children practiced was 5.2 hours per week. Variability in practice time may have influenced individual as well as group outcomes.
Although, based on previous research with typical infants,2 we hoped to find differences in the expressive communication skills of children in the experimental and control groups. We found differences in receptive communication on the BDI, but not expressive communication. All the children had severe speech impairments, and expressive communication was the one domain in which neither group of children showed much improvement over 1 year. Most children communicated primarily by facial expressions, gestures, and unintelligible vocalizations, and none had augmentative or alternative communication methods to assist them. Children in the experimental group may have experienced an enriched vocabulary because the power wheelchairs allowed them exposure to objects in their environments that parents would name. Parental interaction also might have increased and the child's ability to move in response to such interaction could have been a result of improved understanding of language.
The functional performance of self-care skills for children in the experimental group did not change when compared with children in the control group, but they did require less assistance from their caregivers. This was similar to findings by Bottos et al13 who reported differences in self-care. Examination of the items in the PEDI self-care functional skills domain suggested that power mobility was unlikely to have had a direct effect on self-care skills, but may have given children a greater sense of independence and control, leading to more active participation in their self-care.
Our results do not support the commonly expressed concern that motor development will plateau or decline if children use power mobility. This study had limited power, and although we did not identify statistically significant changes in the motor skills between the 2 groups, an actual difference may exist. Another common concern is safety, but no accidents or injuries occurred. Parents of children in the experimental group reported having to provide levels of supervision that were more consistent with the child's age than they had to provide when their children were immobile.
Limitations and Future Research
The small sample size limited the power of the study; therefore, we are unsure whether we would identify additional differences between groups with a larger sample. The sample size also did not allow for multiple testing correction, so a possibility exists that some of the significant results are spurious, and the .10 a priori α-level increased the risk of a type I error.
Future research with a larger number of children could allow identification of variables other than power wheelchair use that might contribute to outcomes, such as age, diagnosis, cognition, and type of wheelchair access and control (eg, joystick vs switches). Because our data indicated that some children had limited practice opportunities, future studies also should consider the influence of increased practice, including access to accessible transportation and more varied learning environments (eg, shopping malls, child care settings) on outcomes or proficiency of power wheelchair use. Opportunities to practice in large areas, rather than only the relatively small area inside a home, might also facilitate faster learning. In general, the children in our study who became more proficient had larger change scores than those who did not become proficient.
An additional limitation of the study was use of the BDI to measure the children's cognitive abilities, particularly as they grew older. Many of the items for measuring cognition required fine motor or speech abilities that children participating in the study did not have. Future studies incorporating cognitive and spatial cognitive measures that do not require fine motor or oral responses would more likely reflect true cognitive abilities and their potential for change.
Use of the ECI for children older than 36 months was another limitation. All children were younger than 30 months at the beginning of the study, but because we followed them for 12 months, some were older than 36 months at the end of the study. Although children who were older than 36 months were still functioning within the developmental age range appropriate for the test, it has not been validated for use with children this age group. We could not find a tool that allowed us to measure children's coping skills that crossed the age ranges in our study and allowed for comparison using the same measure over time.
We followed children for 1 year, which is longer than other studies, and measured their development and functional skills over that year. We might have found different results or detected greater differences between groups, however, if we had collected data after children gained independence in maneuvering the power chairs. It seems logical that only after children become proficient in maneuvering the power chairs would we see accelerated changes in development. Future longitudinal studies that follow children after they become proficient would provide insight about long-term effects of power mobility on development.
Several factors affect the generalizability of the results of this study. We did not draw a random sample from a pool of eligible children. We recruited a relatively small sample of children who met the inclusion criteria, with referrals primarily from therapists working in the early intervention program in 1 state. As a result, the participants most likely were not representative of all children who met the inclusion criteria, such as children who lived in different states, participated in different types of early intervention programs, or whose parents chose not to participate. Another factor that limits generalizability is that the children achieved different levels of proficiency with power wheelchair use and, because of the sample size, we were not able to examine the child and family characteristics that might be related to proficiency. The study also did not address the generalizability of wheelchair use in the home to use in other settings.
Despite the limitations of the study, support was demonstrated for the use of power wheelchairs with children as young as 14 months of age to enhance their mobility, receptive communication, and self-care skills measured over a period of 12 months. We believe clinicians and families should consider power mobility as an intervention for children as young as 14 months of age who have not yet achieved a means of independent mobility and are unlikely to achieve it. This study shows that although children may require extended trials and training to become proficient, power mobility can promote independence and development of young children with severe motor impairments. We recommend that clinicians discuss power mobility with families, not as a last resort, but as one of the options available to support early independent mobility for young children whose independent mobility is severely limited.
The authors thank the children and their families who participated, the therapists who referred children to the study, and Invacare and Adaptive Switch Labs.
1. Campos JJ, Berthenthal BI. Locomotion and psychological development in infancy. In:Jaffe KM, ed. Childhood Powered Mobility: Developmental, Technical and Clinical Perspectives. Arlington, VA: Rehabilitation Engineering and Assistive Technology Society of North America; 1987:11–42.
2. Gustafson GE. Effects of the ability to locomote on infants' social and exploratory behaviors: an experimental study. Dev Psychol. 1984;20:397–405.
3. Benson JB, Uzgiris IC. Effect of self-initiated locomotion on infant search activity. Dev Psychol. 1985;21:923–931.
4. Kermoian R, Campos JJ. Locomotor experience: a facilitator of spatial cognitive development. Child
5. Yan JH, Thomas JR, Downing JH. Locomotion improves children's spatial search: a meta-analytic review. Percept Mot Skills. 1998;87:67–82.
6. Chugani HT, Behen ME, Muzik O, Juhasz C, Nagy F, Chugani DC. Local brain functional activity following early deprivation: a study of postinstitutionalized Romanian orphans. Neuroimage. 2001;14:1290–1301.
7. Black JE. How a child
builds its brain: some lessons from animal studies of neural plasticity. Prev Med. 1998;27:168–171.
8. Jones MA, McEwen IR, Hansen L. Use of power mobility for a young child
with spinal muscular atrophy. Phys Ther. 2003;83:253–262.
9. Tefft D, Guerette P, Furumasu J. Cognitive predictors of young children's readiness for powered mobility. Dev Med Child
10. Butler C. Effects of powered mobility on self-initiated behaviors of very young children with locomotor disability. Dev Med Child
11. Lynch A, Ryu JC, Agrawal S, Galloway JC. Power mobility training for a 7-month-old infant with spina bifida. Pediatr Phys Ther. 2009;21:362–368
12. Deitz J, Swinth Y, White O. Powered mobility and preschoolers with complex developmental delays. Am J Occup Ther. 2002;56:86–96.
13. Bottos M, Bolcati C, Sciuto L, Ruggeri C, Feliciangeli A. Powered wheelchairs
and independence in young children with tetraplegia. Dev Med Child
14. Redcay E, Haist F, Courchesne E. Functional neuroimaging of speech perception during a pivotal period in language acquisition. Dev Sci. 2008;11:237–252.
15. Kochanek TT. Early identification of developmentally disabled and at-risk preschool children. Except Child
16. Ragonesi CB, Chen X, Agrawal S, Galloway JC. Power mobility and socialization in preschool: a case study of a child
with cerebral palsy. Pediatr Phys Ther. 2010;22:322–329.
17. Livingstone R. A critical review of powered mobility assessment and training for children. Disabil Rehabil Assist Technol. 2010;5:392–400.
18. Kangas KM. Clinical assessment and training strategies for the child
's mastery of independent powered mobility. In:Furumasu J, ed. Pediatric Powered Mobility: Developmental Perspectives, Technical Issue and, Clinical Perspectives. Arlington, VA: Rehabilitation Engineering and Assistive Technology Society of North America; 1997:33–47.
19. Newborg J, Stock JR, Wnek L, Guidubaldi J, Svinicki J. Battelle Developmental Inventory. Chicago, IL: Riverside Publishing; 1984.
20. Haley SM, Coster WJ, Ludlow LH, Haltiwanger JT, Andrellos PJ. Pediatric Evaluation of Disability Inventory. Boston, MA: Department of Rehabilitation Medicine, New England Medical Center; 1992.
21. Zeitlin S, Williamson GG, Szczepanski M, Early Coping Inventory. Bensenville, IL: Scholastic Testing Service; 1988.
22. Berls AT, McEwen IR. Battelle developmental inventory. Phys Ther. 1999;79:776–783.
23. Nordmark E, Jarnlo GB, Hagglund G. Comparison of the Gross Motor Function Measure and Pediatric Evaluation of Disability Inventory in assessing motor function in children undergoing selective dorsal rhizotomy. Dev Med Child
24. Wright FV, Boschen K, Jutai J. Exploring the comparative responsiveness of a core set of outcome measures in a school-based conductive education programme. Child
Care Health Dev. 2005;31:291–302.
25. Vos-Vromans DC, Ketelaar M, Gorter JW. Responsiveness of evaluative measures for children with cerebral palsy: the gross motor function measure and the pediatric evaluation of disability inventory. Disabil Rehabil. 2005;27:1245–1252.
26. Butler C, Okamoto GA, McKay TM. Motorized wheelchair driving by disabled children. Arch Phys Med Rehabil. 1984;65:95–97.
27. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice
. 3rd ed. Upper Saddle River, NJ: Prentice-Hall; 2009.
28. Upshur RE. The ethics of alpha: reflections on statistics, evidence and values in medicine. Theor Med Bioeth. 2001;22:565–576.
29. SAS Institute, Inc. Procedures and User's Guide. Version 9.1. Cary, NC: SAS Institute Inc; 2004.
30. Cytel Studio 9
[statistical software]. Cambridge, MA: Cytel; 2010.
31. Campbell PH, Milbourne S, Dugan LM, Wilcox MJ. A review of evidence on practices for teaching young children to use assistive technology devices. Topics Early Child
Spec Educ. 2006;26(1):3–13.
32. Adolph K, Vereijken B, Shrout PE. What changes in infant walking and why. Child
33. Wright-Ott C. The transitional powered mobility aid: a new concept and tool for early mobility. InFurumasu J, ed. Pediatric Powered Mobility: Developmental Perspectives, Technical Issues, Clinical Approaches. Arlington, VA: RESNA; 1997:58–69.
APPENDIX 1 Instructions for Teaching Children to Use Power Mobilitya
Young children in power wheelchairs MUST BE supervised at all times. Adults should be close to monitor all activities and to ensure safety.
- Encourage the child to explore the joystick/switches/head array first, then the movement, and then the environment. Let them learn by doing, giving them time to learn and react.
- Provide positive feedback, for example, “You found the _______(object the child ran into) rather than “Oops you crashed.”
- Help the child by only these words, “come closer,” “go,” “lift your hand off,” or “let's go for a walk.”
- Give the child time to figure out a situation before intervening. If the child looks distressed, then intervene immediately.
- Expect the child to learn how to operate and maneuver the device within a day or week. This is a gradual learning experience. The goal is not to move accurately at first, rather the goal is to give the child a tool to begin moving, exploring, discovering, and problem solving at his or her own speed. The process should be enjoyable and rewarding.
- Say anything that sounds negative, for example, “You crashed into the wall again or you're going the wrong way.”
- Describe how to move by using directionality commands, for example, “Turn this way, press the colored button and go forward, push the joystick and come here, turn left, go in reverse.” The child may not understand these directions in the beginning. In addition, such directions sound like commands and many children will resist this type of interaction.
aModified from Wright-Ott.33
Keywords:© 2012 Lippincott Williams & Wilkins, Inc.
activities of daily living; child; child/preschool; child development; cognition; communication; disabled child; disability evaluation; female; humans; learning; male; motor skills; physical therapy/methods; practice; social development; spatial behavior; treatment outcome; wheelchairs