Research in Developmental Disabilities 37 (2015) 102–111
Contents lists available at ScienceDirect
Research in Developmental Disabilities
Predictors of participation change in various areas for preschool children with cerebral palsy: A longitudinal study Katie P. Wu a, Yu-fen Chuang b, Chia-ling Chen a,c,*, I-shu Liu c, Hsiang-tseng Liu c, Hsieh-ching Chen d a
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital at Linkou, 5 Fu-Hsing St., Kwei-Shan, Tao-Yuan 333, Taiwan Department of Physical Therapy, Chang Gung University, 259 Wen-Hwa 1st Rd, Kwei-Shan, Tao-Yuan 333, Taiwan c Graduate Institute of Early Intervention, Chang Gung University, 259 Wen-Hwa 1st Rd, Kwei-Shan, Tao-Yuan 333, Taiwan d Department of Industrial Engineering and Management, National Taipei University of Technology, 1, Sec. 3, Zhongxiao E. Rd., Taipei 106, Taiwan b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 24 July 2014 Received in revised form 9 November 2014 Accepted 9 November 2014 Available online
This study identifies potential predictors of participation changes in various areas for preschool children with cerebral palsy (CP). Eighty children with CP (2–6 years) were enrolled. Seven potential predictors were identified: age; sex; socioeconomic status, CP subtype; cognitive function, Function Independence Measure for Children (WeeFIM), and motor composite variable from 5 motor factors (gross motor function classification system (GMFCS) level; bimanual fine motor function level; selective motor control score; Modified Ashworth Scale score; and Spinal Alignment and Range of Motion Measure). Outcome was assessed at baseline and at 6-month follow-up using the Assessment of Preschool Children’s Participation (APCP) including diversity and intensity scores in the areas of play (PA), skill development (SD), active physical recreation, social activities (SA), and total areas. Dependent variables were change scores of APCP scores at baseline and 6month follow-up. Regression analyses shows age and sex together predicted for APCPtotal, APCP-SD diversity and APCP-total intensity changes (r2 = 0.13–0.25, p < 0.001); cognitive function and WeeFIM were negative predictors for APCP-SA and APCP-PA diversity changes, respectively. CP subtype, motor composite variable, and socioeconomic status predicted for APCP changes in some areas. Findings suggest that young boys with poor cognitive function and daily activity predicted most on participation changes. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Cerebral palsy Predictor Participation Preschool children Longitudinal study
1. Introduction Manifestations of cerebral palsy (CP) include spasticity, loss of selective motor control (SMC), muscle weakness, and limited range of motion (ROM), which further limit performance at activities of daily living (ADL) and participation in various activities (Calley et al., 2012; Engel-Yeger, Jarus, Anaby, & Law, 2009). Participation of CP children in skill-based, communitybased, and active physical activities is low (Majnemer et al., 2008). For instance, preschool children with CP with poor motor function participated less in activities than those with good motor function (Law, King, Petrenchik, Kertoy, & Anaby, 2012).
* Corresponding author at: Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital at Linkou, 5 Fu-hsing St., Kwei-shan, Taoyuan 333, Taiwan. Tel.: +886 3 3281200x3846; fax: +886 3 3281320. E-mail addresses: [email protected] (K.P. Wu), [email protected] (Y.-f. Chuang), [email protected] (C.-l. Chen), [email protected] (I.-s. Liu), [email protected] (H.-t. Liu), [email protected] (H.-c. Chen). http://dx.doi.org/10.1016/j.ridd.2014.11.005 0891-4222/ß 2014 Elsevier Ltd. All rights reserved.
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
103
Potential predictors, such as age, sex, environmental factors, body function, and activity limitation, for children and youth with CP, have been proposed to be linked to participation in leisure activities (Shikako-Thomas, Majnemer, Law, & Lach, 2008). High physical ability, young, and female are associated with higher intensity of participation in 205 children with CP aged 13–21 years (Palisano, Orlin et al., 2011). Gross Motor Function Classification System (GMFCS) levels are related to participation for children and youth with CP (Brunton & Bartlett, 2010; Palisano et al., 2009; Palisano, Orlin et al., 2011). The GMFCS level and bimanual fine motor function (BFMF) level are associated with participation for children with CP aged 5–8 years (Beckung & Hagberg, 2002). Physical independence and performance in the mobility domains have predicted well by movement and manual ability of children with CP aged 6–12 years (Morris, Kurinczuk, Fitzpatrick, & Rosenbaum, 2006). The CP subtype is related to participation level for children and adolescents with CP (Fauconnier et al., 2009; Kerr, Parkes, Stevenson, Cosgrove, & McDowell, 2008; Schenker, Coster, & Parush, 2005). Bartlett and Palisano proposed a model on determinants for motor changes in children with CP. This model targets on the relationships among child characteristics (e.g. temperament, personality, and cognitive impairments), family ecology, and health care services (Bartlett & Palisano, 2000). Temperament patterns varied among different CP subtypes (Chen et al., 2011) and may further influenced their participation. The cognitive impairments and acitivities of daily living (ADL) limitations were associated with participation in children and adolescents with CP (Fauconnier et al., 2009; Law et al., 2012; Majnemer et al., 2008). The environmental factors, such as socioeconomic status and family factors, were also associated with participation in children or youth with CP (Chan, Lau, Fong, Poon, & Lam, 2005; Colver et al., 2012; Law et al., 2012; Mihaylov, Jarvis, Colver, & Beresford, 2004; Shikako-Thomas, Shevell, Schmitz et al., 2013). Since most studies examined participation intensity for school children and adolescents with CP, knowledge on preschool participants are therefore limited. A 3-year follow-up study indicated that factors associated with change in participation intensity were depended on activity type, sex and age for 402 children/youth with physical disabilities (King et al., 2009). Wright et al. reported poor-tofair relationships between measures of body function and structure, activity, and participation for CP children who were injected with botulinum toxin type-A (Wright, Rosenbaum, Goldsmith, Law, & Fehlings, 2008). Activity and participation gains following injection are likely influenced by environmental factors, GMFCS level, or age for ambulatory children with spastic CP (Wright et al., 2008). Another 2-year study followed on children with CP aged 9–15, indicated that muscle strength, involved limb distribution, SMC, ROM, and spasticity measured by the modified Ashworth scale (MAS) were linked to gross motor function measure (GMFM) score corrected by GMFCS level (Voorman, Dallmeijer, Knol, Lankhorst, & Becher, 2007). A 6-month follow-up study showed that GMFCS level and age are robust negative predictors for change in most developmental domains, such as cognition, language skills, motor function, social function, and self-help in preschool children with CP (Chen, Hsu, et al. 2013). The Spinal Alignment and Range of Motion Measure (SAROMM) (Bartlett & Purdie, 2005) was a negative predictor for cognitive change (Chen, Hsu, et al. 2013). Relationships between potential predictors and participation change are complex; however, few studies have investigated the relationship between potential predictors and participation change in various activities for preschool children with CP. Clinical demand is increasing for valid and responsive participation measures for preschool children to assess participation improvement and justify intervention. The Assessment of Preschool Children’s Participation (APCP) (Law et al., 2012) assesses the level of activity participation for preschool children aged 2–6. Subset data from this study used to determine the clinimetric properties of APCP have already been published (Chen, Chen, et al. 2013), showing that APCP score was markedly responsive to change at follow-up (Chen, Chen, et al. 2013). That is, the clinimetric properties of the APCP measure makes it an appropriate and valid tool to identify participation patterns in terms of diversity and intensity of various activities for preschool children with CP (C.L. Chen, Chen, et al. 2013). This study, attempts to identify potential predictors that can predict participation change in various areas for preschool children with CP. The APCP was selected as the participation measure in this study. This scale contains both diversity and intensity scales in the areas of play (PA), skill development (SD), active physical recreation (AP), and social activities (SA). Potential predictors tested in this study were age, sex, socioeconomic status, CP subtype, cognitive function, motor composite variable (GMFCS level, BFMF level, SMC score, spasticity score, and SAROMM), and ADL. We hypothesize that different predictor combinations can predict participation change (diversity and intensity) in different areas for children with CP. 2. Materials and methods 2.1. Participants Children with CP were recruited from rehabilitation clinics at three hospitals. A physiatrist and a physical therapist independently determined eligibility for inclusion for each participant. The inclusion/exclusion criteria were reported in our previous manuscript (Chen, Chen, et al. 2013). Inclusion criteria were diagnosis of CP and age 2 years to 5 years and 11 months. Children with a progressive neurological disorder, genetic or metabolic disorder, or severe concurrent illness or disease (e.g., active pneumonia or brain tumor) were excluded. Each participant was re-examined at 6 months after the initial assessment. The physiatrist confirmed the CP diagnosis, CP subtype and limb distribution based on history taking, physical examination, chart review, or brain imaging findings. Five participants were lost during follow-up for the 85 children that were initially recruited using convenience sampling (i.e., 1 due to active medical problems, 1 due to death, and contact with
104
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
3 was lost). Finally, 80 children with CP were enrolled (Table 1). During this 6-month study, 76% (n = 61) of participants received physical therapy, 72% (n = 58) of participants received occupational therapy, and 41% (n = 33) of participants received speech therapy. The Institutional Review Board for Human Studies at Chang Gung Memorial Hospital approved this study. Informed consent from each participant was obtained from his/her caregiver. 2.2. Procedure Test measures, the Comprehensive Developmental Inventory for Infants and Toddlers (CDIIT) (Liao, Wang, Yao, & Lee, 2005), GMFCS, BFMF, MAS, SMC, Spinal Alignment and Range of Motion Measure (SAROMM), Function Independence Measure for Children (WeeFIM) (Ottenbacher et al., 1996), and APCP, were administered at baseline and participation outcome measures were re-tested at 6-month follow-up by two trained raters (i.e., physical therapists). Rater training included careful review of written instructions and repeated practice. A senior certified physical therapist assessed rater competence. The test–retest and inter-rater reliabilities were 0.98 and 0.98, respectively, for CDIIT; 1.000 and 0.875, respectively, for the GMFCS levels; 1.000 and 0.875, respectively, for the BFMF levels; 0.855 and 0.865, respectively, for SMC; and 0.907 and 0.870, respectively, for the SAROMM.
Table 1 Demographic and clinical characteristics of participants (N = 80). Characteristics
Valuea Mean or N
Demographic Age (years) Sex Male Female Clinical GMFCS levels Level I Level II Level III Level IV Level V BFMF levels Level I Level II Level III Level IV Level V CP subtypes Unilateral Bilateral SAROMM MAS SMC 0–2 3–4 APCP Diversity (%) Total PA SD AP SA Intensity Total PA SD AP SA
3.9
SD or % 1.4
50 30
62.5% 37.5%
23 16 12 12 17
28.8% 20.0% 15.0% 15.0% 21.3%
16 19 16 12 17
20.0% 23.8% 20.0% 15.0% 21.3%
15 65 22.8 16.4
18.8% 81.3% 10.9 13.3
54 26
67.5% 32.5%
41.9 51.0 36.3 42.6 41.5
20.3 25.5 23.7 20.3 19.9
2.21 3.03 1.93 2.29 1.87
1.13 1.52 1.40 1.16 0.87
N, number of participants; SD, standard deviation; CP, cerebral palsy; GMFCS, gross motor function classification system; BFMF, bimanual fine motor function; SMC, selective motor control; MAS, Modified Ashworth Scale; SAROMM, Spinal Alignment and Range of Motion Measure; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities. a Values were expressed as mean SD for continuous variables and number (%) for categorical variables.
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
105
2.3. Outcome measures The APCP, a questionnaire administered by interview, was the outcome measure (Law et al., 2012); The APCP questionnaire was initially designed to determine the level of activity participation for children aged 2 years to 5 years and 11 months. To assess participation intensity, frequency with which a child participated in an activity was recorded over the last 4 months on a 7-point ordinal scale, where 1 represented the lowest frequency of participation in an activity (once in 4 months), while 7 represented the highest frequency of participation in an activity (at least once per day). Participation diversity is the sum of the total number of reported activities over the 4 months prior to study start. For group data, diversity is a percentage of activity type to all activities. Intensity is derived by dividing the sum of frequencies for all items by the number of items in each activity area. Diversity and intensity scores are generated for each item and in the four activity areas, PA, SD, AP, and SA. The diversity and intensity scores in total area are calculated as the average of the four area scores. The APCP had fair to excellent concurrent validity (r = 0.39–0.85) and predictive validity (r = 0.46–0.82) for children with CP (Chen, Chen, et al. 2013). The clinimetric properties, such as minimal detectable change (MDC) and minimal clinically important difference (MCID), of scales allow clinicians to determine whether a change in score is clinically meaningful (Chen, Chen, et al. 2013). This study is a secondary analysis of data related to APCP that were published earlier (Chen, Chen, et al. 2013). 2.4. Potential predictors According to empirical and theoretical considerations, seven potential predictors were selected: age; sex; Social Status Rating Scale (SSRS) (Rin, Schooler, & Caudill, 1973), CP subtype; cognitive function (CDIIT-COG), motor composite variable, and WeeFIM. A motor composite variable was derived from five motor factors, including GMFCS level (Palisano et al., 1997); BFMF level (Beckung & Hagberg, 2002); SMC (Smits et al., 2010) score; MAS score (Bohannon & Smith, 1987); and SAROMM score (Bartlett & Purdie, 2005), Each CP subtype was classified as unilateral or bilateral (Rosenbaum et al., 2007). A high level or score for the GMFCS, BFMF, MAS, and SAROMM indicate a poor performance. High scores on the SMC indicate a good control. A five point SSRS was used to classify the socioeconomic status of family, ranged from level I to level V according to the education and occupation of key person in their family (Rin et al., 1973). A high level indicates a low socioeconomic status. 2.4.1. GMFCS levels (levels I–V) The GMFCS–Expanded and Revised emphasizes the concepts from the World Health Organization’s International Classification of Functioning, Disability and Health (ICF) (Palisano, Rosenbaum, Bartlett, & Livingston, 2008) considers the impact of both environment and personal factors on mobility. The 5-level GMFCS grading system aim to determine the best level representing the child’s or youth’s present abilities and limitations in gross motor function and the need of wheeled mobility or other assistive technology. Emphasis is on the children’s usual performance in different settings rather than their capacity. A low level indicates good gross motor mobility. For example, level 1 indicates walk without limitations; level II indicates walk with some limitations; level III indicates walk using a hand-held mobility device; level IV indicates selfmobility with limitations that may use powered mobility; and level V indicates transport in a manual wheelchair (Palisano et al., 2008). 2.4.2. BFMF scale (levels I–V) The BFMF scale is a five-level scale for bimanual functional limitations (Beckung & Hagberg, 2002). Levels progress as restrictions on fine motor skills increase. Level I indicates the best bimanual function in which bilateral hands manipulate without restrictions or one hand may have restrictions in advanced fine motor skills. Levels I–III may indicate one hand manipulate without restriction while the other hand has increasing performance limitation or at least one hand has limitations in more advanced fine motor skills. Level V has the least bimanual function in which both hands have only the ability to hold or worse. 2.4.3. The SMC tests (score 0–4) The four-point SMC test (Boyd and Graham SMC test) was used to assess the selective dorsiflexion of the ankle: 0 is for no ankle movement; 1 is for limited dorsiflexion; 2 is for ankle dorsiflexion using the extensor hallucis longus, extensor digitorum longus, and some tibialis anterior activity; 3 is for the ability to dorsiflex using mainly tibialis anterior activity with hip and/or knee flexion; and 4 is for the ability to perform isolated dorsiflexion without hip and knee flexion. This SMC test has a moderate inter-rater reliability (Smits et al., 2010). A higher score indicates a better motor control. 2.4.4. MAS The six-point MAS (Bohannon & Smith, 1987) is administered to assess resistance during passive muscle stretching. The MAS grades spasticity as follows: 0, no increase in muscle tone; 1, slight increase in muscle tone with a catch and release or minimal resistance at the end of the ROM; 1+, slight increase in muscle tone in less than half of the ROM; 2, greater increase in muscle tone through most of the ROM with easy passive movement; 3, marked increase in muscle tone with difficult passive movement; and 4, rigid limb in flexion or extension. The MAS score is used to assess the extent of muscle tone in
106
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
upper extremities (i.e., elbow flexor, pronator, wrist flexors, finger flexor, and thumb adductor) and lower extremities (i.e., hip adductor, hamstrings, gastrocnemius and soleus) on bilateral sides. The MAS total score is the sum of for the bilateral upper and lower extremities. The intra-observer reliability of the MAS was low to average (Numanoglu & Gunel, 2012). A higher score indicates a greater spasticity. 2.4.5. SAROMM (score 0–4) The five-point SAROMM contains 26 items, including 4 items for spinal alignment and 11 items for ROM and muscle extensibility tested bilaterally (Bartlett & Purdie, 2005). A score of 0 represents the ability to align normally without passive limitations and a score of 4 indicates that that a subject has severe deviations in spinal alignment, limitations in joint ROM, and/or muscle extensibility. The SAROMM total score is the sum of spinal alignment and ROM scores (possible range, 0–104). The intraclass correlation coefficients for interrater and test–retest reliabilities for all SAROMM subscales and total scores for children and adolescents with CP are all >0.80 (Bartlett & Purdie, 2005). A higher score indicates a worse spinal alignment and ROM. 2.4.6. WeeFIM The WeeFIM (Ottenbacher et al., 1996) is an 18-item, 7-level ordinal scale instrument that is categorized into 2 main functional groups: ‘‘Dependent’’ (scores 1–5, helper requires) and ‘‘Independent’’ (scores 6–7, no helper requires). It has three functional subscales: self-care (WeeFIM-SC; eight items), mobility (five items), and cognition (five items). Level 1 indicates total assistance while Level 7 indicates no assistance for the child and the child can complete the task independently without requiring a device. The WeeFIM has excellent reliability (Ottenbacher et al., 1996). A higher score indicates a better ADL. 2.4.7. CDIIT The CDIIT (Liao et al., 2005) is widely used in Taiwan and includes five developmental subtests. The developmental age (DA) for the cognitive subtests (CDIIT-Cog) was selected for this study. The CDIIT has acceptable test–retest reliabilities (intra-class correlation coefficient (ICC > 0.76), inter-rater reliability (ICC > 0.76), and validity in children (Liao et al., 2005). A higher score indicates a better cognitive function. 2.5. Statistical analysis The change score in outcome measure (APCP), determined by subtracting pretest score from posttest score, is the dependent variable. A principal components analysis (PCA) was used to extract a motor composite variable from five motor factors (GMFCS level, BFMF level, reverse scores of SMC, MAS scores, and SAROMM scores). Predictors for APCP change were identified in two steps. First, predictors (age, sex, SSRS, CP subtype, motor composite variable, CDIIT-Cog, and WeeFIM) at baseline were examined for associations with change scores for APCP outcome measures (diversity and intensity scales) by using the Pearson correlation coefficient (r). The criterion for predictors inclusion in regression analysis was p 0.25 (Groff, Lundberg, & Zabriskie, 2009). Second, predictors in regression analysis were subjected to a forward stepwise procedure to generate a linear regression model for APCP change in each area. Adjusted r2, p values, and regression coefficients (b) were used to assess goodness-of-fit in the regression models. Regression diagnostics were also applied to examine multicollinearity (i.e., variance inflation factor (VIF) among predictors in the models. 3. Results Table 1 lists demographic and clinical characteristics of the 80 participants. Table 2 groups participants according to change status (improvement, no change, or deterioration) in each domain of the APCP at follow-up. Most participants improved both in diversity and intensity scores in all areas at 6 months. The 5 motor factors, including GMFCS level, BFMF level, reverse scores of SMC, MAS scores, and SAROMM scores, were entered as variables in PCA to extract the motor composite variable. The first component (GMFCS level) accounted for 80% of the variances. Table 3 lists Pearson coefficients for the correlation between the seven predictors and the change scores for outcome measures (between baseline scores and those at follow-up). Three predictors, age, SSRS, and CDIIT-Cog were entered into the APCP-AP intensity model; age, sex, and CP subtypes were entered into the APCP-SA intensity model. Four predictors, age, sex, and CDIIT-Cog, and WeeFIM were entered into the APCP-total, APCP-SD diversity, and APCP-PA intensity models; age, sex, SSRS, and CDIIT-Cog were entered into the APCP-AP diversity model; age, sex, and CDIIT-Cog, and motor composite variable were entered into the APCP-SD intensity model. All predictors except for SSRS and CP subtypes were entered into the APCP-PA diversity model. All predictors except for SSRS were entered into the APCP-SA diversity model. Values of all variables (VIF < 1.1) were within normal ranges (VIF 5), confirming multicollinearity assumption held. Table 4 presents the results of stepwise multiple regression analyses. Age and sex together were predictors in the APCPtotal, APCP-SD diversity and APCP-total intensity models (r2 = 0.13–0.25, p < 0.001). WeeFIM was a predictor in the APCP-PA diversity model (r2 = 0.08, p < 0.01). CDIIT-Cog and CP subtypes were predictors in the APCP-SA diversity model (r2 = 0.13, p < 0.01). Age, sex, and motor composite variable were strong predictors in the APCP-SD intensity models (r2 = 0.22, p < 0.001). Age and SSRS were predictors in APCP-AP intensity model (r2 = 0.07, p < 0.05). Age and CP subtype alone was a
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
107
Table 2 Number of participants according to status of change in the motor outcome measures. Outcome measure
Status of change
APCP
Improvement
No change
Deterioration
N
N
N
63 51 57 34 45
4 17 14 26 19
13 12 9 20 16
65 54 66 49 58
1 10 2 5 6
14 16 12 26 16
Diversity Total PA SD AP SA Intensity Total PA SD AP SA
N, number of participant; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities.
Table 3 Relationships between the potential predictors and the change scores of the APCP outcome measures. Predictors
Pearson’s r APCP Diversity Total
Demographic Age Sex Socioeconomic status Clinical CP subtypes Motor composite variable CDIIT-Cog WeeFIM
Intensity PA
SD
AP
SA
Total
PA
SD
AP
SA
0.43* 0.32* 0.06
0.28* 0.22* 0.03
0.38* 0.29* 0.03
0.21* 0.14* 0.15*
0.22* 0.15* 0.03
0.38* 0.28* 0.09
0.28* 0.21* 0.08
0.39* 0.27* 0.05
0.24* 0.13 0.16*
0.14* 0.17* 0.04
0.12 0.05 0.38* 0.25*
0.13 0.17* 0.26* 0.30*
0.09 0.13 0.27* 0.13*
0.11 0.01 0.17* 0.09
0.20* 0.14* 0.29* 0.15*
0.13 0.09 0.27* 0.13
0.12 0.10 0.25* 0.26*
0.10 0.18* 0.24* 0.08
0.13 0.10 0.16* 0.04
0.29* 0.06 0.11 0.03
CP, cerebral palsy; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities; CDIIT-Cog: Comprehensive Developmental Inventory for Infants and Toddlers-cognitive function; WeeFIM: Function Independence Measure for Children. * p-Values < 0.25.
predictor in APCP-PA and APCP-SA intensity models, respectively (r2 = 0.07, p < 0.05). However, no predictors were entered into the final APCP-AP diversity model. The 9 final regression equations are as follows:
Total diversity = (30.4) (3.1) age (6.5) sex PA diversity = (20.1) (0.2) WeeFIM SD diversity = (39.6) (3.9) age (8.6) sex SA diversity = (36.5) (0.3) CDIIT-Cog (11.1) CP subtypes Total intensity = (1.9) (0.2) age (0.4) sex PA intensity = (1.7) (0.2) age SD intensity = (2.3) (0.2) age (0.5) Sex (0.2) motor composite variable AP intensity = (1.6) (0.2) age (0.2) SSRS SA intensity = (1.8) (0.6) CP subtype
4. Discussion To the best of our knowledge, this study is the first to identify predictors for participation change in various areas for preschool children with CP. This study demonstrated that different factor combinations can indeed predict participation change in various areas. For example, age and sex combined were the strongest predictors of participation changes in total areas. Cognitive function, WeeFIM, CP subtype, motor composite variable, and socioeconomic status predicted for APCP
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
108
Table 4 Forward stepwise multiple regression analyses of the predictors for the changes in the APCP measures. Dependent variables
Independent variables
Coefficient (b)
95% CI lower
95% CI upper
Adjusted r2
Diversity Total
F
VIF
14.1 Constant Age Sex
30.4 3.1 6.5
22.2 4.6 10.9
38.7 1.6 2.2
Constant WeeFIM
20.1 0.2
12.7 0.3
27.4 0.0
0.171 0.248
PA
Contents lists available at ScienceDirect
Research in Developmental Disabilities
Predictors of participation change in various areas for preschool children with cerebral palsy: A longitudinal study Katie P. Wu a, Yu-fen Chuang b, Chia-ling Chen a,c,*, I-shu Liu c, Hsiang-tseng Liu c, Hsieh-ching Chen d a
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital at Linkou, 5 Fu-Hsing St., Kwei-Shan, Tao-Yuan 333, Taiwan Department of Physical Therapy, Chang Gung University, 259 Wen-Hwa 1st Rd, Kwei-Shan, Tao-Yuan 333, Taiwan c Graduate Institute of Early Intervention, Chang Gung University, 259 Wen-Hwa 1st Rd, Kwei-Shan, Tao-Yuan 333, Taiwan d Department of Industrial Engineering and Management, National Taipei University of Technology, 1, Sec. 3, Zhongxiao E. Rd., Taipei 106, Taiwan b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 24 July 2014 Received in revised form 9 November 2014 Accepted 9 November 2014 Available online
This study identifies potential predictors of participation changes in various areas for preschool children with cerebral palsy (CP). Eighty children with CP (2–6 years) were enrolled. Seven potential predictors were identified: age; sex; socioeconomic status, CP subtype; cognitive function, Function Independence Measure for Children (WeeFIM), and motor composite variable from 5 motor factors (gross motor function classification system (GMFCS) level; bimanual fine motor function level; selective motor control score; Modified Ashworth Scale score; and Spinal Alignment and Range of Motion Measure). Outcome was assessed at baseline and at 6-month follow-up using the Assessment of Preschool Children’s Participation (APCP) including diversity and intensity scores in the areas of play (PA), skill development (SD), active physical recreation, social activities (SA), and total areas. Dependent variables were change scores of APCP scores at baseline and 6month follow-up. Regression analyses shows age and sex together predicted for APCPtotal, APCP-SD diversity and APCP-total intensity changes (r2 = 0.13–0.25, p < 0.001); cognitive function and WeeFIM were negative predictors for APCP-SA and APCP-PA diversity changes, respectively. CP subtype, motor composite variable, and socioeconomic status predicted for APCP changes in some areas. Findings suggest that young boys with poor cognitive function and daily activity predicted most on participation changes. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Cerebral palsy Predictor Participation Preschool children Longitudinal study
1. Introduction Manifestations of cerebral palsy (CP) include spasticity, loss of selective motor control (SMC), muscle weakness, and limited range of motion (ROM), which further limit performance at activities of daily living (ADL) and participation in various activities (Calley et al., 2012; Engel-Yeger, Jarus, Anaby, & Law, 2009). Participation of CP children in skill-based, communitybased, and active physical activities is low (Majnemer et al., 2008). For instance, preschool children with CP with poor motor function participated less in activities than those with good motor function (Law, King, Petrenchik, Kertoy, & Anaby, 2012).
* Corresponding author at: Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital at Linkou, 5 Fu-hsing St., Kwei-shan, Taoyuan 333, Taiwan. Tel.: +886 3 3281200x3846; fax: +886 3 3281320. E-mail addresses: [email protected] (K.P. Wu), [email protected] (Y.-f. Chuang), [email protected] (C.-l. Chen), [email protected] (I.-s. Liu), [email protected] (H.-t. Liu), [email protected] (H.-c. Chen). http://dx.doi.org/10.1016/j.ridd.2014.11.005 0891-4222/ß 2014 Elsevier Ltd. All rights reserved.
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
103
Potential predictors, such as age, sex, environmental factors, body function, and activity limitation, for children and youth with CP, have been proposed to be linked to participation in leisure activities (Shikako-Thomas, Majnemer, Law, & Lach, 2008). High physical ability, young, and female are associated with higher intensity of participation in 205 children with CP aged 13–21 years (Palisano, Orlin et al., 2011). Gross Motor Function Classification System (GMFCS) levels are related to participation for children and youth with CP (Brunton & Bartlett, 2010; Palisano et al., 2009; Palisano, Orlin et al., 2011). The GMFCS level and bimanual fine motor function (BFMF) level are associated with participation for children with CP aged 5–8 years (Beckung & Hagberg, 2002). Physical independence and performance in the mobility domains have predicted well by movement and manual ability of children with CP aged 6–12 years (Morris, Kurinczuk, Fitzpatrick, & Rosenbaum, 2006). The CP subtype is related to participation level for children and adolescents with CP (Fauconnier et al., 2009; Kerr, Parkes, Stevenson, Cosgrove, & McDowell, 2008; Schenker, Coster, & Parush, 2005). Bartlett and Palisano proposed a model on determinants for motor changes in children with CP. This model targets on the relationships among child characteristics (e.g. temperament, personality, and cognitive impairments), family ecology, and health care services (Bartlett & Palisano, 2000). Temperament patterns varied among different CP subtypes (Chen et al., 2011) and may further influenced their participation. The cognitive impairments and acitivities of daily living (ADL) limitations were associated with participation in children and adolescents with CP (Fauconnier et al., 2009; Law et al., 2012; Majnemer et al., 2008). The environmental factors, such as socioeconomic status and family factors, were also associated with participation in children or youth with CP (Chan, Lau, Fong, Poon, & Lam, 2005; Colver et al., 2012; Law et al., 2012; Mihaylov, Jarvis, Colver, & Beresford, 2004; Shikako-Thomas, Shevell, Schmitz et al., 2013). Since most studies examined participation intensity for school children and adolescents with CP, knowledge on preschool participants are therefore limited. A 3-year follow-up study indicated that factors associated with change in participation intensity were depended on activity type, sex and age for 402 children/youth with physical disabilities (King et al., 2009). Wright et al. reported poor-tofair relationships between measures of body function and structure, activity, and participation for CP children who were injected with botulinum toxin type-A (Wright, Rosenbaum, Goldsmith, Law, & Fehlings, 2008). Activity and participation gains following injection are likely influenced by environmental factors, GMFCS level, or age for ambulatory children with spastic CP (Wright et al., 2008). Another 2-year study followed on children with CP aged 9–15, indicated that muscle strength, involved limb distribution, SMC, ROM, and spasticity measured by the modified Ashworth scale (MAS) were linked to gross motor function measure (GMFM) score corrected by GMFCS level (Voorman, Dallmeijer, Knol, Lankhorst, & Becher, 2007). A 6-month follow-up study showed that GMFCS level and age are robust negative predictors for change in most developmental domains, such as cognition, language skills, motor function, social function, and self-help in preschool children with CP (Chen, Hsu, et al. 2013). The Spinal Alignment and Range of Motion Measure (SAROMM) (Bartlett & Purdie, 2005) was a negative predictor for cognitive change (Chen, Hsu, et al. 2013). Relationships between potential predictors and participation change are complex; however, few studies have investigated the relationship between potential predictors and participation change in various activities for preschool children with CP. Clinical demand is increasing for valid and responsive participation measures for preschool children to assess participation improvement and justify intervention. The Assessment of Preschool Children’s Participation (APCP) (Law et al., 2012) assesses the level of activity participation for preschool children aged 2–6. Subset data from this study used to determine the clinimetric properties of APCP have already been published (Chen, Chen, et al. 2013), showing that APCP score was markedly responsive to change at follow-up (Chen, Chen, et al. 2013). That is, the clinimetric properties of the APCP measure makes it an appropriate and valid tool to identify participation patterns in terms of diversity and intensity of various activities for preschool children with CP (C.L. Chen, Chen, et al. 2013). This study, attempts to identify potential predictors that can predict participation change in various areas for preschool children with CP. The APCP was selected as the participation measure in this study. This scale contains both diversity and intensity scales in the areas of play (PA), skill development (SD), active physical recreation (AP), and social activities (SA). Potential predictors tested in this study were age, sex, socioeconomic status, CP subtype, cognitive function, motor composite variable (GMFCS level, BFMF level, SMC score, spasticity score, and SAROMM), and ADL. We hypothesize that different predictor combinations can predict participation change (diversity and intensity) in different areas for children with CP. 2. Materials and methods 2.1. Participants Children with CP were recruited from rehabilitation clinics at three hospitals. A physiatrist and a physical therapist independently determined eligibility for inclusion for each participant. The inclusion/exclusion criteria were reported in our previous manuscript (Chen, Chen, et al. 2013). Inclusion criteria were diagnosis of CP and age 2 years to 5 years and 11 months. Children with a progressive neurological disorder, genetic or metabolic disorder, or severe concurrent illness or disease (e.g., active pneumonia or brain tumor) were excluded. Each participant was re-examined at 6 months after the initial assessment. The physiatrist confirmed the CP diagnosis, CP subtype and limb distribution based on history taking, physical examination, chart review, or brain imaging findings. Five participants were lost during follow-up for the 85 children that were initially recruited using convenience sampling (i.e., 1 due to active medical problems, 1 due to death, and contact with
104
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
3 was lost). Finally, 80 children with CP were enrolled (Table 1). During this 6-month study, 76% (n = 61) of participants received physical therapy, 72% (n = 58) of participants received occupational therapy, and 41% (n = 33) of participants received speech therapy. The Institutional Review Board for Human Studies at Chang Gung Memorial Hospital approved this study. Informed consent from each participant was obtained from his/her caregiver. 2.2. Procedure Test measures, the Comprehensive Developmental Inventory for Infants and Toddlers (CDIIT) (Liao, Wang, Yao, & Lee, 2005), GMFCS, BFMF, MAS, SMC, Spinal Alignment and Range of Motion Measure (SAROMM), Function Independence Measure for Children (WeeFIM) (Ottenbacher et al., 1996), and APCP, were administered at baseline and participation outcome measures were re-tested at 6-month follow-up by two trained raters (i.e., physical therapists). Rater training included careful review of written instructions and repeated practice. A senior certified physical therapist assessed rater competence. The test–retest and inter-rater reliabilities were 0.98 and 0.98, respectively, for CDIIT; 1.000 and 0.875, respectively, for the GMFCS levels; 1.000 and 0.875, respectively, for the BFMF levels; 0.855 and 0.865, respectively, for SMC; and 0.907 and 0.870, respectively, for the SAROMM.
Table 1 Demographic and clinical characteristics of participants (N = 80). Characteristics
Valuea Mean or N
Demographic Age (years) Sex Male Female Clinical GMFCS levels Level I Level II Level III Level IV Level V BFMF levels Level I Level II Level III Level IV Level V CP subtypes Unilateral Bilateral SAROMM MAS SMC 0–2 3–4 APCP Diversity (%) Total PA SD AP SA Intensity Total PA SD AP SA
3.9
SD or % 1.4
50 30
62.5% 37.5%
23 16 12 12 17
28.8% 20.0% 15.0% 15.0% 21.3%
16 19 16 12 17
20.0% 23.8% 20.0% 15.0% 21.3%
15 65 22.8 16.4
18.8% 81.3% 10.9 13.3
54 26
67.5% 32.5%
41.9 51.0 36.3 42.6 41.5
20.3 25.5 23.7 20.3 19.9
2.21 3.03 1.93 2.29 1.87
1.13 1.52 1.40 1.16 0.87
N, number of participants; SD, standard deviation; CP, cerebral palsy; GMFCS, gross motor function classification system; BFMF, bimanual fine motor function; SMC, selective motor control; MAS, Modified Ashworth Scale; SAROMM, Spinal Alignment and Range of Motion Measure; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities. a Values were expressed as mean SD for continuous variables and number (%) for categorical variables.
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
105
2.3. Outcome measures The APCP, a questionnaire administered by interview, was the outcome measure (Law et al., 2012); The APCP questionnaire was initially designed to determine the level of activity participation for children aged 2 years to 5 years and 11 months. To assess participation intensity, frequency with which a child participated in an activity was recorded over the last 4 months on a 7-point ordinal scale, where 1 represented the lowest frequency of participation in an activity (once in 4 months), while 7 represented the highest frequency of participation in an activity (at least once per day). Participation diversity is the sum of the total number of reported activities over the 4 months prior to study start. For group data, diversity is a percentage of activity type to all activities. Intensity is derived by dividing the sum of frequencies for all items by the number of items in each activity area. Diversity and intensity scores are generated for each item and in the four activity areas, PA, SD, AP, and SA. The diversity and intensity scores in total area are calculated as the average of the four area scores. The APCP had fair to excellent concurrent validity (r = 0.39–0.85) and predictive validity (r = 0.46–0.82) for children with CP (Chen, Chen, et al. 2013). The clinimetric properties, such as minimal detectable change (MDC) and minimal clinically important difference (MCID), of scales allow clinicians to determine whether a change in score is clinically meaningful (Chen, Chen, et al. 2013). This study is a secondary analysis of data related to APCP that were published earlier (Chen, Chen, et al. 2013). 2.4. Potential predictors According to empirical and theoretical considerations, seven potential predictors were selected: age; sex; Social Status Rating Scale (SSRS) (Rin, Schooler, & Caudill, 1973), CP subtype; cognitive function (CDIIT-COG), motor composite variable, and WeeFIM. A motor composite variable was derived from five motor factors, including GMFCS level (Palisano et al., 1997); BFMF level (Beckung & Hagberg, 2002); SMC (Smits et al., 2010) score; MAS score (Bohannon & Smith, 1987); and SAROMM score (Bartlett & Purdie, 2005), Each CP subtype was classified as unilateral or bilateral (Rosenbaum et al., 2007). A high level or score for the GMFCS, BFMF, MAS, and SAROMM indicate a poor performance. High scores on the SMC indicate a good control. A five point SSRS was used to classify the socioeconomic status of family, ranged from level I to level V according to the education and occupation of key person in their family (Rin et al., 1973). A high level indicates a low socioeconomic status. 2.4.1. GMFCS levels (levels I–V) The GMFCS–Expanded and Revised emphasizes the concepts from the World Health Organization’s International Classification of Functioning, Disability and Health (ICF) (Palisano, Rosenbaum, Bartlett, & Livingston, 2008) considers the impact of both environment and personal factors on mobility. The 5-level GMFCS grading system aim to determine the best level representing the child’s or youth’s present abilities and limitations in gross motor function and the need of wheeled mobility or other assistive technology. Emphasis is on the children’s usual performance in different settings rather than their capacity. A low level indicates good gross motor mobility. For example, level 1 indicates walk without limitations; level II indicates walk with some limitations; level III indicates walk using a hand-held mobility device; level IV indicates selfmobility with limitations that may use powered mobility; and level V indicates transport in a manual wheelchair (Palisano et al., 2008). 2.4.2. BFMF scale (levels I–V) The BFMF scale is a five-level scale for bimanual functional limitations (Beckung & Hagberg, 2002). Levels progress as restrictions on fine motor skills increase. Level I indicates the best bimanual function in which bilateral hands manipulate without restrictions or one hand may have restrictions in advanced fine motor skills. Levels I–III may indicate one hand manipulate without restriction while the other hand has increasing performance limitation or at least one hand has limitations in more advanced fine motor skills. Level V has the least bimanual function in which both hands have only the ability to hold or worse. 2.4.3. The SMC tests (score 0–4) The four-point SMC test (Boyd and Graham SMC test) was used to assess the selective dorsiflexion of the ankle: 0 is for no ankle movement; 1 is for limited dorsiflexion; 2 is for ankle dorsiflexion using the extensor hallucis longus, extensor digitorum longus, and some tibialis anterior activity; 3 is for the ability to dorsiflex using mainly tibialis anterior activity with hip and/or knee flexion; and 4 is for the ability to perform isolated dorsiflexion without hip and knee flexion. This SMC test has a moderate inter-rater reliability (Smits et al., 2010). A higher score indicates a better motor control. 2.4.4. MAS The six-point MAS (Bohannon & Smith, 1987) is administered to assess resistance during passive muscle stretching. The MAS grades spasticity as follows: 0, no increase in muscle tone; 1, slight increase in muscle tone with a catch and release or minimal resistance at the end of the ROM; 1+, slight increase in muscle tone in less than half of the ROM; 2, greater increase in muscle tone through most of the ROM with easy passive movement; 3, marked increase in muscle tone with difficult passive movement; and 4, rigid limb in flexion or extension. The MAS score is used to assess the extent of muscle tone in
106
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
upper extremities (i.e., elbow flexor, pronator, wrist flexors, finger flexor, and thumb adductor) and lower extremities (i.e., hip adductor, hamstrings, gastrocnemius and soleus) on bilateral sides. The MAS total score is the sum of for the bilateral upper and lower extremities. The intra-observer reliability of the MAS was low to average (Numanoglu & Gunel, 2012). A higher score indicates a greater spasticity. 2.4.5. SAROMM (score 0–4) The five-point SAROMM contains 26 items, including 4 items for spinal alignment and 11 items for ROM and muscle extensibility tested bilaterally (Bartlett & Purdie, 2005). A score of 0 represents the ability to align normally without passive limitations and a score of 4 indicates that that a subject has severe deviations in spinal alignment, limitations in joint ROM, and/or muscle extensibility. The SAROMM total score is the sum of spinal alignment and ROM scores (possible range, 0–104). The intraclass correlation coefficients for interrater and test–retest reliabilities for all SAROMM subscales and total scores for children and adolescents with CP are all >0.80 (Bartlett & Purdie, 2005). A higher score indicates a worse spinal alignment and ROM. 2.4.6. WeeFIM The WeeFIM (Ottenbacher et al., 1996) is an 18-item, 7-level ordinal scale instrument that is categorized into 2 main functional groups: ‘‘Dependent’’ (scores 1–5, helper requires) and ‘‘Independent’’ (scores 6–7, no helper requires). It has three functional subscales: self-care (WeeFIM-SC; eight items), mobility (five items), and cognition (five items). Level 1 indicates total assistance while Level 7 indicates no assistance for the child and the child can complete the task independently without requiring a device. The WeeFIM has excellent reliability (Ottenbacher et al., 1996). A higher score indicates a better ADL. 2.4.7. CDIIT The CDIIT (Liao et al., 2005) is widely used in Taiwan and includes five developmental subtests. The developmental age (DA) for the cognitive subtests (CDIIT-Cog) was selected for this study. The CDIIT has acceptable test–retest reliabilities (intra-class correlation coefficient (ICC > 0.76), inter-rater reliability (ICC > 0.76), and validity in children (Liao et al., 2005). A higher score indicates a better cognitive function. 2.5. Statistical analysis The change score in outcome measure (APCP), determined by subtracting pretest score from posttest score, is the dependent variable. A principal components analysis (PCA) was used to extract a motor composite variable from five motor factors (GMFCS level, BFMF level, reverse scores of SMC, MAS scores, and SAROMM scores). Predictors for APCP change were identified in two steps. First, predictors (age, sex, SSRS, CP subtype, motor composite variable, CDIIT-Cog, and WeeFIM) at baseline were examined for associations with change scores for APCP outcome measures (diversity and intensity scales) by using the Pearson correlation coefficient (r). The criterion for predictors inclusion in regression analysis was p 0.25 (Groff, Lundberg, & Zabriskie, 2009). Second, predictors in regression analysis were subjected to a forward stepwise procedure to generate a linear regression model for APCP change in each area. Adjusted r2, p values, and regression coefficients (b) were used to assess goodness-of-fit in the regression models. Regression diagnostics were also applied to examine multicollinearity (i.e., variance inflation factor (VIF) among predictors in the models. 3. Results Table 1 lists demographic and clinical characteristics of the 80 participants. Table 2 groups participants according to change status (improvement, no change, or deterioration) in each domain of the APCP at follow-up. Most participants improved both in diversity and intensity scores in all areas at 6 months. The 5 motor factors, including GMFCS level, BFMF level, reverse scores of SMC, MAS scores, and SAROMM scores, were entered as variables in PCA to extract the motor composite variable. The first component (GMFCS level) accounted for 80% of the variances. Table 3 lists Pearson coefficients for the correlation between the seven predictors and the change scores for outcome measures (between baseline scores and those at follow-up). Three predictors, age, SSRS, and CDIIT-Cog were entered into the APCP-AP intensity model; age, sex, and CP subtypes were entered into the APCP-SA intensity model. Four predictors, age, sex, and CDIIT-Cog, and WeeFIM were entered into the APCP-total, APCP-SD diversity, and APCP-PA intensity models; age, sex, SSRS, and CDIIT-Cog were entered into the APCP-AP diversity model; age, sex, and CDIIT-Cog, and motor composite variable were entered into the APCP-SD intensity model. All predictors except for SSRS and CP subtypes were entered into the APCP-PA diversity model. All predictors except for SSRS were entered into the APCP-SA diversity model. Values of all variables (VIF < 1.1) were within normal ranges (VIF 5), confirming multicollinearity assumption held. Table 4 presents the results of stepwise multiple regression analyses. Age and sex together were predictors in the APCPtotal, APCP-SD diversity and APCP-total intensity models (r2 = 0.13–0.25, p < 0.001). WeeFIM was a predictor in the APCP-PA diversity model (r2 = 0.08, p < 0.01). CDIIT-Cog and CP subtypes were predictors in the APCP-SA diversity model (r2 = 0.13, p < 0.01). Age, sex, and motor composite variable were strong predictors in the APCP-SD intensity models (r2 = 0.22, p < 0.001). Age and SSRS were predictors in APCP-AP intensity model (r2 = 0.07, p < 0.05). Age and CP subtype alone was a
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
107
Table 2 Number of participants according to status of change in the motor outcome measures. Outcome measure
Status of change
APCP
Improvement
No change
Deterioration
N
N
N
63 51 57 34 45
4 17 14 26 19
13 12 9 20 16
65 54 66 49 58
1 10 2 5 6
14 16 12 26 16
Diversity Total PA SD AP SA Intensity Total PA SD AP SA
N, number of participant; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities.
Table 3 Relationships between the potential predictors and the change scores of the APCP outcome measures. Predictors
Pearson’s r APCP Diversity Total
Demographic Age Sex Socioeconomic status Clinical CP subtypes Motor composite variable CDIIT-Cog WeeFIM
Intensity PA
SD
AP
SA
Total
PA
SD
AP
SA
0.43* 0.32* 0.06
0.28* 0.22* 0.03
0.38* 0.29* 0.03
0.21* 0.14* 0.15*
0.22* 0.15* 0.03
0.38* 0.28* 0.09
0.28* 0.21* 0.08
0.39* 0.27* 0.05
0.24* 0.13 0.16*
0.14* 0.17* 0.04
0.12 0.05 0.38* 0.25*
0.13 0.17* 0.26* 0.30*
0.09 0.13 0.27* 0.13*
0.11 0.01 0.17* 0.09
0.20* 0.14* 0.29* 0.15*
0.13 0.09 0.27* 0.13
0.12 0.10 0.25* 0.26*
0.10 0.18* 0.24* 0.08
0.13 0.10 0.16* 0.04
0.29* 0.06 0.11 0.03
CP, cerebral palsy; APCP, Assessment of Preschool Children’s Participation; PA, play; SD, skill development; AP, active physical recreation; SA, social activities; CDIIT-Cog: Comprehensive Developmental Inventory for Infants and Toddlers-cognitive function; WeeFIM: Function Independence Measure for Children. * p-Values < 0.25.
predictor in APCP-PA and APCP-SA intensity models, respectively (r2 = 0.07, p < 0.05). However, no predictors were entered into the final APCP-AP diversity model. The 9 final regression equations are as follows:
Total diversity = (30.4) (3.1) age (6.5) sex PA diversity = (20.1) (0.2) WeeFIM SD diversity = (39.6) (3.9) age (8.6) sex SA diversity = (36.5) (0.3) CDIIT-Cog (11.1) CP subtypes Total intensity = (1.9) (0.2) age (0.4) sex PA intensity = (1.7) (0.2) age SD intensity = (2.3) (0.2) age (0.5) Sex (0.2) motor composite variable AP intensity = (1.6) (0.2) age (0.2) SSRS SA intensity = (1.8) (0.6) CP subtype
4. Discussion To the best of our knowledge, this study is the first to identify predictors for participation change in various areas for preschool children with CP. This study demonstrated that different factor combinations can indeed predict participation change in various areas. For example, age and sex combined were the strongest predictors of participation changes in total areas. Cognitive function, WeeFIM, CP subtype, motor composite variable, and socioeconomic status predicted for APCP
K.P. Wu et al. / Research in Developmental Disabilities 37 (2015) 102–111
108
Table 4 Forward stepwise multiple regression analyses of the predictors for the changes in the APCP measures. Dependent variables
Independent variables
Coefficient (b)
95% CI lower
95% CI upper
Adjusted r2
Diversity Total
F
VIF
14.1 Constant Age Sex
30.4 3.1 6.5
22.2 4.6 10.9
38.7 1.6 2.2
Constant WeeFIM
20.1 0.2
12.7 0.3
27.4 0.0
0.171 0.248
PA
