Test retest reliability of balance tests in children with cerebral palsy

Transcription

1 Test retest reliability of balance tests in children with cerebral palsy Hua-Fang Liao* RPT PhM, Associate Professor, School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University; Pai-Jun Mao RPT MsPT, Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital; Ai-Wen Hwang RPT MsPT, School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan, Republic of China. *Correspondence to first author at School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, 7 Chun-Shan S. Road, Taipei, Taiwan, Republic of China. hfliao@ha.mc.ntu.edu.tw To investigate intrasession and intersession reliability of balance tests in children with or without disabilities, 50 children without disabilities (ND) and 36 children with cerebral palsy (CP) aged from 5 to 12 years were tested. Intrasession reliability of postural stability of the Smart Balance Master System and one-leg standing test were assessed in both groups and intersession reliability of the Smart Balance Master System and balance subtest of the Bruininks Oseretsky Test of Motor Proficiency (BOTMP) were assessed in ND children. Intersession reliability of the postural stability test in ND children, obtained using the Smart Balance Master System, was of moderate to good reliability in centre target (CT), sway vision (SV), eyes open and sway surface (EOSS), and sway vision and sway surface (SVSS; ICC 0.72 to 0.84). In children with CP, intrasession reliability was high in CT (ICC 1). One-leg standing tests in both groups also had moderate to good intersession reliability (ICC 0.56 to 0.99). Agreement of failure score of lateral rhythmic shifting (LRS) at 1 second and 2 seconds pace was 85% and 93% respectively in ND children. Within the balance subtest of BOTMP, only two items had 100% agreement. Results suggest that postural stability tests in four conditions (CT, SV, EOSS, and SVSS), LRS, one-leg standing, and walking on a line are reliable and can be used to monitor balance control in ND children. Postural stability in CT condition and one-leg standing test are reliable in children with CP. Further study is needed to establish more reliable balance tests for children with CP. Ability to control posture while standing, including stability in different sensory conditions, organized muscle responses to external perturbation, and feed-forward control, begins to develop in early childhood. Some of these postural control abilities are still developing after the age of 13 years (Riach 1987, Haas et al. 1989, Woollacott and Sveistrup 1992, Shimizu et al. 1994). Children with developmental disabilities may have poor control of either compensatory postural reaction, or anticipatory postural control, which may contribute to their delayed or deviated motor development. Balance deficits have been well demonstrated in children with cerebral palsy (CP; Bobath 1964, Nashner et al. 1983, Liao et al. 1997b), Down syndrome (Shumway-Cook and Woollacott 1985, Haley 1987), learning disability (Horak et al. 1988), and hearing impairments (Crowe and Horak 1988). Physical therapy to improve motor function in children with disabilities usually emphasizes balance training (Liao 1990, Heriza 1991, Woollacott 1994). However, a review of present techniques for assessing balance function, revealed that few standardized, reliable, and comprehensive measures exist (Atwater et al. 1990, Westcott et al. 1997). Common clinical tests used by physical therapists and occupational therapists include the duration of maintaining a static posture (e.g. one-leg standing, tandem standing), postural stability under different sensory conditions (e.g. eyes closed, standing on a sponge), degrees of tilting on a tiltboard, and the ability to maintain balance dynamically (e.g. walking on a balance beam; Atwater et al. 1990, Crowe et al. 1990, Westcott et al. 1997). Previous studies of the reliability of different balance tests have found that: (1) the one-leg standing test with eyes open or closed has good interrater reliability (Spearman s r=0.87 to 0.99) and fair to good test retest reliability (Spearman s r=0.59 to 1.0; Atwater et al. 1990). (2) The tiltboard tip test has good interrater reliability (Spearman s r=0.98), but poor test retest reliability in children with or without disability (Atwater et al. 1990, Broadstone et al. 1993). (3) The Pediatric Clinical Test of Sensory Interaction for Balance (P-CTSIB; Crowe et al. 1990) has good interrater reliability in the tandem position (Spearman s r=0.73 to 0.92) and fair to good reliability in the feet-together position (Spearman s r=0.69 to 0.90). It has also been reported that because of the difficulty of performing the P-CTSIB in the tandem position, only the feet-together position is suitable for testing children under 4 years old (Richardson et al. 1992). Test retest reliabilities in P- CTSIB duration and sway scores are low to moderate. However, when these scores were added according to the sensory system, test retest reliabilities improved (Spearman s r=0.45 to 0.69 in the feet-together position, r=0.44 to 0.83 in the tandem position). The tandem P-CTSIB shows a great learning effect between test and retest and is, therefore, not suitable for balance evaluation in children aged from 4 to 9 years old (Westcott et al. 1994). The Bruininks Oseretsky Test of Motor Proficiency (BOTMP; Bruininks 1978) has been used as an evaluative or discriminative assessment tool (Connolly et al. 1993, Wilson et al. 1995). However, clinical use of the balance subtest has been discouraged because of its low reliability (Wilson et al. 1995). Postural stability tests based on the extent of the excursion from centre of pressure have also been used for many years, such as the Smart Balance Master System, (version 3.4; NeuroCom International, Clackamas, OR, USA). However, test retest reliability data are still lacking (Liao et al. 1997b, Westcott et al. 1997). 180 Developmental Medicine & Child Neurology 2001, 43:

2 Most previous studies of balance tests have involved children without disabilities (ND) and have used the Spearman s correlation test as the reliability analysis test. Whether the results obtained from these studies can be applied to children with developmental disabilities needs further investigation. Results of the linear coefficient test may show a linear association but not necessarily demonstrate agreement between test and retest (Ottenbacher and Tomchek 1992). Intraclass correlation tests (ICC) have been cited as the most suitable type of reliability test (Shrout and Fleiss 1979, Ottenbach and Tomchek 1992). Children s performance on the balance task appears to fluctuate from one session to another. Therefore, in this study, the test retest reliability was evaluated using data from tests performed on the same day (intrasession), and from tests performed on different days within one week (intersession) in order to determine the extent of performance stability. The purposes of this study were: (1) to establish the intrasession reliability of postural stability of ND children and children with CP using the Smart Balance Master System under different sensory conditions, (2) to establish the intersession reliability of the postural stability and dynamic rhythmic shifting test of the Smart Balance Master System in ND children, (3) to establish the reliability of the one-leg standing test in ND children and children with CP, and (4) to evaluate the intersession reliability of the balance subtest of the BOTMP test in ND children. Method PROCEDURE The study was conducted in four parts (see Table I). Informed consent was obtained from parents of all participants after providing a detailed explanation of each part of the procedure. Part 1: Intrasession reliability of postural stability using Smart Balance Master System under different sensory conditions Sixteen ND children, 10 males and six females, with ages Table I: Test procedure and participants Testing Purposes Participants procedures nr/(y:mo) Part 1 Intrasession reliability of 16 ND children/ postural stability using Smart (mean 8:7, SD 2:3) Balance Master System 13 children with CP/ (mean 8:9, SD 2:2) Part 2 Intersession reliability of 14 ND children/ postural stability and dynamic (mean 10:2, SD 2:2) balance test using Smart Balance Master System Part 3 Reliability of clinical 14 ND children/ balance test one-leg standing (mean 10:2, SD 2:2) 9 children with CP/ (mean 9:3, SD 0:7) Part 4 Intersession reliability of 20 ND children/ balance subtest of BOTMP (mean 10:6, SD 2:3) ND, non-disabled children; CP, children with cerebral palsy. ranging from 6 years 9 months to 12 years 2 months (mean 8 years 7 months, SD 2 years 3 months) were recruited from two elementary schools near the College of Medicine, National Taiwan University, Republic of China. Children in the ND category were excluded if they had any neuromuscular, musculoskeletal, or cardiopulmonary disease. Thirteen children with spastic diplegic cerebral palsy (SDCP) were also recruited from the Department of Rehabilitation Medicine, National Taiwan University Hospital. The inclusion criteria were: (1) ability to stand for more than 1 minute, (2) ability to follow instructions, and (3) no visual or hearing impairment. The SDCP group comprised six males and seven females with ages ranging from 5 to 12 years (mean 8 years 9 months, SD 2 years 2 months). Postural stability was tested using the Smart Balance Master System under seven sensory conditions, including centre target (visual feedback) with fixed support (CT), eyes open with fixed support (EO), eyes closed with fixed support (EC), swayed vision reference and fixed support (SV), eyes open with swayed support (EOSS), eyes closed with swayed support (ECSS), and swayed vision reference and swayed support (SVSS). Each set of tests included seven sensory conditions and the sequence of these sensory conditions was randomly arranged. Every set of tests was performed three times during the same day with a 5-minute rest between every two sets. Children were barefoot for all the tests. The Smart Balance Master System includes a force plate that measures the vertical component of the centre of foot pressure (COP). In the postural stability test, percentage of limit of stability (LOS) was used to represent the sway of COP. For details on the test see our previous report (Liao et al. 1997b). Table II: Intrasession reliability of postural stability of children using Smart Balance Master System in different sensory conditions Sensory ICC 95% CI F value p condition ND children (n=16) CT < EO <0.01 EC < SV < EOSS ns ECSS ns SVSS <0.01 Children with CP (n=13) CT < EO ns EC <0.05 SV <0.05 EOSS <0.05 ECSS <0.05 SVSS ns CT, centre target; EO, eyes open; EC, eyes closed; SV, sway vision; EOSS, eyes open, sway surface; ECSS, eyes closed, sway surface; SVSS, sway vision, sway surface. ND, non-disabled children; CP, children with cerebral palsy. Test retest Reliability of Balance Tests in CP Hua-Fang Liao et al. 181

3 Part 2: Intersession reliability of postural stability and dynamic balance test Fourteen ND children were recruited from two nearby elementary schools for participation in intersession reliability testing. The ND group comprised seven males and seven females with ages ranging from 6 years 9 months to 13 years 2 months (mean 10 years 2 months, SD 2 years 2 months). Postural stability tests were performed as described in Part 1. The dynamic balance test was the lateral rhythmic weight-shift test of the Smart Balance Master System. It evaluates the ability of the child to move the centre of a pressure cursor by shifting his/her bodyweight rhythmically between two lateral endlines at a 50% limit of stability for three cycles, and to repeat this sequence within 1, 2, and 3 seconds from line to line. A failure score was recorded if the child was unable to perform three cycles continuously or unable to follow the designated speeds (the average transition time of 0.5 seconds more or less than the set time), or if the child fell. The failure score ranged from 0 to 3 for each pace (i.e 1, 2, or 3 seconds) because the test was administered a total of three times for each pace during each session. A score of 0 was assigned to a child who performed rhythmic shifting 3 times without failure. All children received testing and retesting within one week, and each test was performed three times in a randomized sequence. The test retest reliability was analyzed by ICC. The data from the static balance test were collected on the same day and were either analyzed using the average value or by the value of the best performance. Part 3: Reliability of clinical balance test one-leg standing The 14 ND children were the same as in Part 2. All were rightleg dominant. Nine children with CP were included: three males and six females, aged from 8 years to 9 years 11 months (mean 9 years 3 months, SD 7 months). Three of the children had spastic hemiplegia and six spastic diplegia. The less involved legs of the children with CP were right in six and left in three. Selection criteria were the same as in Part 1. These children were recruited from the hospitals or developmental centres in Taipei. The inclusion criteria were: (1) ability to perform one-leg stand for more than 1 second, and (2) ability to follow instructions. The dominant leg of each child was decided at the beginning of the test by kicking a tennis ball to the examiner. The ND children underwent a one-leg standing test using their dominant leg under two different sensory conditions: eyes open and standing on a medium density sponge (one-leg- EOSS), and eyes closed and standing on a medium density sponge (one-leg-ecss). Each condition was tested three times. In the EO condition, the child was asked to place both hands on hips and to look at a target on the wall at eye level at 1 metre distance. The child was asked to stand with the dominant leg and keep the other leg in a position of 90 degrees of flexion. A stopwatch was used to record the time the child was able to maintain the testing position. The examiner asked the child to stop if the child was able to maintain the position for more than 30 seconds. If the child did not meet the criteria of the BOTMP, the trial was discontinued and the time recorded. Criteria for trial discontinuance included dropping the raised leg so that the foot touched the floor, dropping the leg below a 45 degree angle after warning, hooking the raised leg behind the supporting leg, and shifting the supporting foot out of place. In a pilot study, most of the children with CP could not maintain EC or EOSS positions. In this study therefore, only the test of oneleg standing with eyes open on a firm base (one-leg-eo) was Table III: Intersession reliability of postural stability of ND children using Smart Balance Master System in different sensory conditions (n=14) Sensory condition Body sway Body sway ICC 95% CI F value p 1st day 2nd day (% of LOS) (% of LOS) Best value of three repetitive tests, mean (SD) CT 0.09 (0.06) 0.11 (0.07) <0.005 EO 0.12 (0.10) 0.13 (0.12) ns EC 0.17 (0.12) 0.17 (0.12) ns SV 0.17 (0.14) 0.18(0.17) < EOSS 0.47 (0.36) 0.40 (0.49) <0.001 ECSS 1.59 (0.91) 1.08 (1.26) <0.05 SVSS 0.17 (0.14) 0.18 (0.17) <0.001 Mean value of three repetitive tests, mean (SD) CT 0.15 (0.12) 0.19 (0.23) ns EO 0.19 (0.20) 0.25 (0.26) ns EC 2.67 (8.89) 0.25 (0.16) ns SV 0.27 (0.23) 0.25 (0.21) <0.01 EOSS 5.54 (11.96) 0.73 (1.07) ns ECSS 9.24 (13.92) 3.94 (8.68) <0.05 SVSS (20.07) 3.63 (9.74) ns LOS, limit of stability; CT, centre target; EO, eyes open; EC, eyes closed; SV, sway vision; EOSS, eyes open, sway surface; ECSS, eyes closed, sway surface; SVSS, sway vision, sway surface. 182 Developmental Medicine & Child Neurology 2001, 43:

4 used for children with CP. All the children were retested within one week. Test retest reliability was analyzed by ICC. The data of static balance test collected on the same day were analyzed using the average value or value of the best performance. Part 4: Intersession reliability of balance subtest of BOTMP in ND children Twenty ND children, six males and 14 females, were included in this part of the study. Their ages ranged from 6 years 9 months to 13 years 2 months (mean 10 years 6 months, SD 2 years 3 months). Most children with CP do not have good enough motor ability to carry out the BOTMP evaluation. Therefore we did not test the reliability of the balance subtest of the BOTMP for the children with CP. The dominant leg was right in 18 children and left in two children. Selection criteria were the same as in Part 1. Eight items of the balance subtest of the BOTMP test were used, including standing on the preferred leg on the floor; standing on the preferred leg on a balance beam; standing on the preferred leg on a balance beam with EC; walking forward on the balance beam; walking forward heel-to-toe on a walking line; walking forward heel-to-toe on a balance beam; and stepping over a response speed stick on the balance beam. The test and retest were performed within one week, and at the same time of day. Point scores of the balance subtest and eight individual test items were analyzed to investigate their reliability. STATISTICAL ANALYSIS The Statistical Package for Social Sciences (SPSS for Windows version 6) was used for data analysis. Intraclass correlation coefficients (ICC 2,1) were used in the intrasession or intersession reliability study of postural stability tests and one-leg standing test (Shrout and Fleiss 1979). The percentage of agreement was used to assess the reliability of rhythmic weight-shifting tests and balance tests of the BOTMP. Statistical significance was set at p<0.05. Table IV: Test retest percentage of agreement for failure score at lateral rhythmic weight shift in ND children Speed a Failure score (n=14) Agreement % At 1s pace test retest At 2s pace test retest At 3s pace test retest a To move centre of pressure cursor rhythmically between two lateral endlines at 50% limit of stability for three cycles, and to repeat sequence at 1, 2, or 3 seconds pace. Table V: Intersession reliability of one-leg standing test using mean duration of three repetitive tests in ND children and children with CP ND children (n=14) Children with CP (n=9) ICC 95% CI F value ICC 95% CI F value One-leg-EO d c One-leg-EOSS d a One-leg-ECSS d a One-leg-EOSS e a One-leg-ECSS e b EO, eyes open; EOSS, eyes open, sway surface; ECSS, eyes closed, sway surface. a p<0.05; b p<0.01; p< ; d Analyzed using mean duration of three repetitive tests; e Analyzed using longest duration three repetitive tests. ND, non-disabled children; CP, children with cerebral palsy. Table VI: Intersession reliability of individual tests of balance subtest for Bruininks Oseretsky Test of Motor Proficiency in ND children (n=20) Mean (SD) difference Range of difference Agreement % of point score of point score One-leg standing on floor on balance beam 0.4 (0.97) on balance beam, eyes closed 1.7 (1.89) Walking forward on walking line on balance beam 0.6 (1.07) heel-to-toe on walking line 0.6 (0.7) heel-to-toe on balance beam 0.8 (0.92) Stepping over stick on balance beam 0.5 (0.53) Test retest Reliability of Balance Tests in CP Hua-Fang Liao et al. 183

5 Results PART 1 The results for ICC of the intrasession reliability of postural stability in seven sensory conditions are shown in Table II. In general, ICC values of less than 0.5 can be referred to as indicating poor reliability, between 0.5 and 0.75, moderate reliability, and above 0.75, good reliability (Portney and Watkins 1993). For the ND group, with the exception of EOSS and ECSS, the test retest scores were significantly correlated with the other five sensory conditions, with a range from poor to high reliability. The correlation coefficients were: CT, 0.79; EC, 0.62; SV, 0.59; SVSS, 0.45; and EO, In children with CP, with the exception of EO and SVSS, the test retest scores were significantly correlated on the other five sensory conditions, with weak to high reliability (ICC 1 in CT, 0.42 in EC, 0.36 in EOSS, 0.34 in ECSS, and 0.33 in SV). Therefore, only postural stability on CT conditions showed good intrasession reliability in both ND children and children with CP. PART 2 Test retest reliability, with a one-week interval between testing and retesting, was examined for using a sample of 14 ND children. The reliability coefficient and magnitude of differences between the best scores or the mean score for postural stability under seven sensory conditions are shown in Table III. When the best performance score of the three tests and the lowest percentage of the LOS were used in the correlation analysis, the reliability of postural stability was statistically significant with the exception of EO and EC in ND children. Postural stability under the three conditions revealed a high reliability, (ICC 0.84 in SV, 0.80 in EOSS, and 0.75 in SVSS). The postural stability test in CT was also of moderate reliability (ICC 0.72). Comparison of the best scores in the test and retest revealed that the average difference was small and was not significantly different on paired t-test. No learning effect was found in the postural stability tested by the Smart Balance Master System, because the body sway of the retest was not always smaller than that in the test score. Analysis of the means of the percentage of LOS of three tests showed that the correlation coefficient was only significant in SV (ICC 0.68 ) and ECSS (ICC 0.42). For the dynamic balance test, the agreement of failure scores in the test and retest at different time paces are shown in Table IV. The reliability of the failure score of the lateral rhythmic weight shift was computed using percentage of agreement (Bartko 1976). The percentage of agreement was high at 1 and 2 seconds pace, but was low at 3 seconds pace. The children in this study also showed a learning effect in lateral rhythmic weight-shift ability at 3 seconds pace. PART 3 The ICC of the intrasession reliability of one-leg-eoss was 0.53 (F=4.2, p<0.01), while that of the one-leg-ecss was not significant (F=0.97, p>0.05) in ND children. In children with CP, the intrasession reliability was good (ICC 0.98, F=346, p<0.001). The reliability of the mean duration of three repetitive tests of one-leg standing on the dominant leg is shown in Table V. They were all moderately reliable in two different sensory conditions, with ICC ranging from 0.56 to 0.72 in ND children. In children with CP, the one-leg-eo was highly reliable (ICC 1). When the best performance of three repetitive tests was used for intersession reliability tests, the result in the oneleg-eoss was low (ICC 0.47) and was moderate in the one-leg- ECSS (ICC 0.74). PART 4 The ICC between test and retest of the balance subtest of the BOTMP was non-significant (ICC 0.4, F=2.3, p>0.05). Further analysis of each balance item found that the agreement between two tests were 100% in item 1: standing on the preferred leg on the floor, and item 4: walking forward on a walking line. The percentages of agreement were under 80% for the other six items, and the differences of point scores between test and retest ranged from 0 to 5. Therefore, only two items in the balance subtest of BOTMP were found to be reliable for test retest in this study (Table VI). Discussion Human performance on various balance tests are task specific (Drowatzky and Zuccato 1967). For example, a person may perform one type of balance task well but perform another type at only an average level (Haywood 1993). The degree of postural sway and reliability of the measurement itself are influenced by the age of the child and the equipment employed (Baker et al. 1998). According to Maki s classification, balance control can be compensatory or volitional, fixed-support or change-in-support (Maki and McIlroy 1997). Increased understanding of balance control could lead to the development of new diagnostic and therapeutic approaches for detecting and treating specific causes of imbalance. There are relatively few measurements of postural stability in children which have acceptable reliability and validity documentation (Westcott et al. 1997). This impedes clinicians in definitive determination of the factors contributing to postural problems, and further complicates the process of designing the most effective treatment. Results of this study indicate that some balance tests that were of moderate to good reliability are valid clinical tests, while some with poor reliability may not be clinically useful. There are several methods that may be used to assess the sensory organization of postural stability. These tests include the Romberg test, tandem walking test, tiltboard tip test, postrotatory nystagmus test, Pediatric Clinical Test of Sensory Interaction for Balance, and platform posturography of sensory organization. Among these tests, platform posturography of sensory organization provides a means for evaluating deficits in central sensory organization (Forssberg and Nashner 1982), and is being used with increasing frequency in clinics (Westcott et al. 1997). However, only one reliability study of such posturography measurement has been reported. Intersession reliability of static standing as determined by using the sway area as a variable was 0.94 in young healthy adults (Hoffman and Koceja 1997). However, these results may not necessarily be applicable to children. The degree of postural sway and reliability of the measurement itself are influenced by the age of the child and the equipment employed (Baker et al. 1998). In this study, postural stability tests using the Smart Balance Master System showed moderate to high intersession reliability under four sensory conditions (CT, SV, EOSS, SVSS) in ND children when the best performance of three repetitive tests was used in the analysis. When the mean of 184 Developmental Medicine & Child Neurology 2001, 43:

6 three repetitive tests was used for the intersession reliability test, the reliability was relatively poor. This may be due to the low intrasession reliability of the postural stability test or the assignment of 100 to the failure stability. These results suggest that the best performance of three repetitive tests in CT, SV, EOSS, and SVSS can be used to monitor the progression of development of postural stability in children aged from 6 to 12 years old. Previously, we found that self-induced weight-shift ability was closely correlated with the ambulatory function in children with CP (Liao et al. 1997b). The reliability of rhythmic weight-shifting ability has never been investigated. In this study, the lateral rhythmic weigh-shifting ability at 1 and 2 seconds pace were high, but were low at 3 seconds pace. The children in this study also showed a learning effect in lateral rhythmic weight-shift ability at 3 seconds pace. The walking cycle time is about 0.8 to 1.1 seconds (Sutherland 1984). The faster rhythmic weight-shifting rate is similar to the walking cycle, which is an automatic movement pattern, and explains why more consistent performances can be observed. Slower rhythmic shifting requires more voluntary control and, therefore, shows better performance after practice. The one-leg standing test is a common assessment used by physical therapists and occupational therapists. Various standardized balance assessment tools used in children from 4 to 12 years of age include the Peabody Developmental Motor Scales (Folio and Fewell 1983) and the BOTMP. Visual cues and the body position can affect balance during one-leg standing (Clark and Watkins 1984, Schulmann et al. 1987). Therefore, the standing position and a fixed visual cue in the EO condition is standardized in these assessment tools. A previous study found that the one-leg-eo and one-leg-ec had fair to good test retest reliability (Spearman s r=0.59 to 1.0), and that the reliability was lower in the EC condition (Atwater et al. 1990). Results of this study also show that reliability in the EC condition was lower than that in the EO condition when analysis was done using the mean of three repetitive tests. No previous studies have reported results for reliability of the one-leg standing test on sponges in either normally developing children or those with CP. Our results indicate that all of the tests in the four sensory conditions had moderate intersession reliability in ND children when the mean of three repetitive tests was used. The one-leg-eo was highly reliable in children with CP. The balance subtest of the BOTMP had a low reliability in this study, which is similar to the results of a previous study done by Wilson and colleagues (1995). These data strongly contraindicate its clinical use. Using the balance subtest of the BOTMP, Siegel found that the balance deficit of the child with a hearing impairment was not age related (Siegel et al. 1991). Although this result may have been simply due to the poor test retest reliability of the balance subtest of the BOTMP, the authors also found that two items in the balance subtest of BOTMP were highly reliable, with 100% agreement. These two items were standing on the preferred leg on the floor and walking forward on a walking line. Along with the moderate to high reliability of the one-leg standing test found in this study, these two items can be used individually as clinical assessment tools in children. In summary, this study has established the test retest reliability of four different balance tests in ND children. Those are postural stability and dynamic rhythmic shifting ability using the Smart Balance Master System, one-leg standing test, and the balance subtest of the BOTMP test. Also, test retest of postural stability, and one-leg standing were evaluated in children with CP. Moderate to high reliability of some of the assessments suggests that stable results will be obtained with these assessments in children in clinical practice. However, test retest reliability does not provide information about the effectiveness of using such assessments in documenting changes in clinical practice. The ICC pattern differed in the intrasession reliability under different sensory conditions by using the Smart Balance Master System. Determination of whether differences between children with CP and ND children represent different developmental patterns, i.e. whether the low ICC values represent the most inefficient balance strategies, will require further investigation. Accepted for publication 27th April Acknowledgments This study was supported by grants from the National Science Council of the Executive Yuan, Taiwan, Republic of China. We thank the parents and children who participated this study. References Atwater SW, Crowe TK, Deitz JC, Richardson PK. (1990) Interrater and test retest reliability of two pediatric balance tests. Physical Therapy 70: Baker CP, Newstead AH, Mossberg KA, Nicodemus CL. (1998) Reliability of static standing balance in non disabled children: comparison of two methods of measurement. Pediatric Rehabilitation 2: Bartko JJ, Carpenter WT. (1976) On the methods and theory of reliability. Journal of Nervous and Mental Disease 163: Bobath K, Bobath B. (1964) The facilitation of normal postural reactions and movements in the treatment of cerebral palsy. Physiotherapy 50: Broadstone BJ, Westcott SL, Deitz JC. (1993) Test retest reliability of two tiltboard tests in children. Physical Therapy 73: Bruininks RH. (1978) Bruininks Oseretsky Test of Motor Proficiency Examiner s Manual. Circle Pines, MN: American Guidance Service. p 34 7, Clark JE, Watkins DL. (1984) Static balance in young children. Child Development 55: Connolly BH, Morgan SB, Russel FF, Fulliton WL. (1993) A longitudinal study of children with Down syndrome who experienced early intervention programming. Physical Therapy 73: Crowe TK, Horak FB. (1988) Motor proficiency associated with vestibular deficits in children with hearing impairments. Physical Therapy 68: Deitz JC, Richardson PK, Atwater SW. (1990) Interrater reliability of the pediatric clinic test of sensory interaction for balance. Physical and Occupational Therapy in Pediatrics 10: Drowatzky JN, Zuccato FC. (1967) Interrelationships between selected measures of static and dynamic balance. Research Quarterly 38: Folio MR, Fewell R. (1983) Peabody Development Motor Scales Manual. Allen, TX: DLM Teaching Resource. Forssberg H, Nashner LM. (1982) Ontogenetic development of postural control in man: adaptation to altered support and visual condition during stance. Journal of Neuroscience 2: Haas G, Diener HC, Rapp H, Dichgams J. (1989) Development of feedback and feed forward control of upright stance. Developmental Medicine & Child Neurology 31: Hayley SM. (1987) Sequence of development of postural reactions by infants with Down syndrome. Developmental Medicine & Child Neurology 29: Test retest Reliability of Balance Tests in CP Hua-Fang Liao et al. 185

7 Haywood KM. (1993) Life Span Motor Development, 2nd edn. Champaign, IL, Human Kinetics Publishers. p 224. Heriza CH. (1991) Motor development: traditional and contemporary theories. In: Lister MJ, editor. Contemporary Management of Motor Control Problems. Alexandria, VA: Foundation for Physical Therapy. p Hoffman MA, Koceja DM. (1997) Dynamic balance testing with electrically evoked perturbation: a test of reliability. Archives of Physical Medicine and Rehabilitation 78: Horak FB, Shumway-Cook A, Crowe TK, Black FO. (1988) Vestibular function and motor proficiency of children with impaired hearing, or with learning disability and motor impairments. Developmental Medicine & Child Neurology 30: Liao HF. (1990) The development of gross motor patterns in spastic diplegics and the treatment approaches based on Bobath concepts. Journal of the Physical Therapy Association of the Republic of China 15: Suen MP, Soong WT, Lien IL. (1987) Developmental profile of different types of cerebral-palsied children. Journal of Formosan Medical Association 86: Huang HH, Lee SC, Jeng SF, Chow WP. (1997a) Rehabilitation services for children with cerebral palsy in two medical institutions in Taipei City. Formosan Journal of Medicine 1: Jeng SF, Lai JS, Cheng CK, Hu MH. (1997b) The relation between standing balance and walking function in children with spastic diplegic cerebral palsy. Developmental Medicine & Child Neurology 39: Maki BE, McIlroy WE. (1997) The role of limb movements in maintaining upright stance: the change-in-support strategy. Physical Therapy 77: Nashner LM, Shumway-Cook A, Marin O. (1983) Stance posture control in select groups of children with cerebral palsy: deficits in sensory organization and muscular coordination. Experimental Brain Research 49: Ottenbacher KJ, Tomchek SD. (1992) Reliability analysis in therapeutic research: practice and procedure. American Journal of Occupational Therapy 47: Portney LG, Watkins MP. (1993) Foundations of Clinical Research: applications to practice. East Norwalk, CT: Appleton & Lange. p Riach CL, Hayes KC. (1987) Maturation of the postural sway in young children. Developmental Medicine & Child Neurology 29: Richardson PK, Atwater SW, Crowe TK, Deitz JC. (1992) Performance of preschoolers of the pediatric clinical test of sensory interaction for balance. American Journal of Occupational Therapy 46: Schulmann DL, Godfrey B, Fisher AG. (1987) Effect of eye movements on dynamic equilibrium. Physical Therapy 67: Shimizu K, Asai M, Takata S, Watanabe Y. (1994) The development of equilibrium function in childhood. In: Taguchi K, Igarashi M, Mori S, editors. Vestibular and Neural Front. Amsterdam: Elsevier Science. p Shrout PE, Fleiss JL. (1979) Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin 86: Shumway-Cook A, Woollacott MH. (1985) Dynamics of postural control in the child with Down syndrome. Physical Therapy 65: Siegel JC, Marchetti M, Tecklin JS. (1991) Age-related balance changes in hearing-impaired children. Physical Therapy 71: Sutherland DH. (1984) Gait Disorders in Childhood and Adolescence. London: Williams & Wilkins. p Wilson BN, Polatajko HJ, Kaplan BJ, Faris P. (1995) Use of the Bruininks Oseretsky test of motor proficiency in occupational therapy. American Journal of Occupational Therapy 49: Westcott SL, Crowe TK, Deitz JC, Richardson P. (1994) Test retest reliability of the pediatric clinical test of sensory interaction for balance (P-CTSIB). Physical & Occupational Therapy in Pediatrics 14: Lowes LP, Richardson PK. (1997) Evaluation of postural stability in children: current theories and assessment tools. Physical Therapy 77: Woollacott MH. (1994) Changes in balance control across the lifespan: can training improve balance efficiency? In: Taguchi K, Igarashi M, Mori S, editors. Vestibular and Neural Front. Amsterdam: Elsevier Sciences. p Sveistrup H. (1992) Changes in the sequencing and timing of muscle response coordination associated with developmental transitions in balance abilities. Human Movement Science 11: Developmental Medicine & Child Neurology 2001, 43: