Investigation of Difference Between the Three Representatives of the Functional Status According to Upper Limb Functions and Participation in Children with Congenital Hemiplegic Cerebral Palsy

Background: Although there are many valid and reliable functional motor and communication functions classication tools to describe disabled children’s functional status in clinical or research environments; however, the actual difference among their levels based on actual performance remains to be determined. Aim: This study aimed to explore whether the three functional status representatives differentiate according to actual performance in everyday life in children with congenital hemiplegic cerebral palsy (CP). Method: High to moderate functional motor and communication performance levels as described by Manual Ability Classication System (MACS), Gross Motor Function Classication System, and Communication Function Classication System (CFCS) were investigated in the context of the scaled scores of upper limb functions and participation in different life situations on ABILHAND-Kids and Child and Adolescent Scale of Participation (CASP) questionnaires. The data were collected from 98 children with congenital hemiplegic CP from different rehabilitation centers (mean age 9,3 years [SD 3.0 years], [%95 CI:8.7-9.9]; 42 females, 56 males); 28,6% classied to level I, 29,6 % to level II, and 41,8 % to level III in MACS. Results: The study ndings demonstrated that scaled scores of the upper limb functions and participation in different life situations clearly increased with the greater MACS, GMFCS, and CFCS level, except for CFCS levels closest to each other. Conclusion: MACS and GMFCS was strongest predictor of a child’s actual performance in daily life.


Introduction
Congenital hemiplegic cerebral palsy (CP) constitutes one-third of all typologies of CP (1 in 1300 live birth) [1], predominantly characterized by one upper and lower extremity affected [2]. Assessment of the children with hemiplegic CP is essential in identifying the existing problems associated with upper limb functions [3] and participation in different life situations ('involvement in a life situation') [4]. Moreover, in either research or clinical settings, it is also bene cial to quick yet robust document the clinical characteristics of the disabled children [3]. Following that, classi cation systems were developed as practical tools to concisely provide knowledge to researchers or clinicians as a snapshot of functional  [13]. A previous study has reported that these three classi cation systems complement each other to depict the child's capacity associated with upper limb functions, mobility, and communication performance [14] onto the World Health Organization's (WHO) international classi cation of functioning, disability, and health [15]. A potential inter-relationship among GMFCS, MACS, and CFCS was determined to strong or moderate in 222 children with CP aged 2-7 years. Besides, in the same study, it was found out that the GMFCS had a high correlation with MACS, however, less with CFCS [16]. Similarly, a relatively more recent study by Compagnone [14] reported a strong correlation between these three functional classi cation systems.
One aspect of the validation of the classi cation systems is their correspondence to the external references by using outcome measures [9]. Therefore, it is crucial to explore what means each of the classi cation system levels in the context of real-life utilizing the most appropriate outcome measures. In other words, validating the classi cation systems within the functional performance in daily life and understanding each level's potential difference in terms of an external reference is essential for clinicians, health professionals and caregivers/families. Following that, some studies have investigated a potential inter-relationship between these three prevailing functional classi cation systems in perspective of upper limb function, demonstrating a meaningful correlation among classi cation systems in question and functional performance in daily living [6; 9; 17]. However, while their unique resultants, these studies included participants of different typologies of CP at all MACS and GMFCS levels, which strengthens better understanding of the differences or similarities between high to moderate levels (I-III) of classi cation systems. Furthermore, it would be crucial to differentiate between high to moderate levels of classi cation systems in the context of upper limb function and participation to enhance communication between families and health professionals in the clinic settings and as well as to construct homogenous groups in the research settings. Actual discrimination between classi cation systems' levels based on the actual performance in basic daily living activities and involvement in different life situations is required to practically delineate the functional status of children with hemiplegic CP because effectively measuring the functional status of this population is vital in monitoring their functional status in short periods [18].
Based upon the premises mentioned above, the purpose of this study was to further investigate (1) actual differences between high to moderate levels of the prevailing functional classi cation systems in respect of upper limb functions and participation in different life situations and (2) an inter-relationship between MACS, GMFCS, and CFCS. We hypothesized that there would be a potential difference between the high to moderate levels of the classi cation systems based on functional performance in real world on children with hemiplegic CP 2. Material And Methods

Participant
Participants in this cross-sectional study were the children with hemiplegic CP between the ages of 7-14 years and were at the MACS, CFCS, and GMFCS levels I-III. Human Research Ethics Committee at the XXXX University approved the study protocols numbered 79236777-050.01.04. A convenience sample size of the 98 children diagnosed with congenital hemiplegia (56 males, 46 females) born in XXXX city was set from different rehabilitation centers. Informed consent was obtained and signed by the parents/guardian of the children. Participants' functional motor and communication levels were evaluated by an experienced rehabilitation therapist who is expert in this eld. Children with lower levels (IV-V) of functional motor and communication skills or diagnosed with acquired brain injury were excluded from the study.

Classi cation Tools of Functional Motor and Communications Performance
The participants' functional status in the study was documented using the GMFCS (inter-rater reliability; interclass correlation coe cient [ICC]= 0,93)[19], MACS ( inter-rater reliability; ICC=0.89-0.98)[20], and CFCS ( inter-rater reliability ICC= 0.66)[12] based on both clinical observation and the reporting of child's physiotherapist or families. These three instruments are universally accepted to describe the functional performance as a snapshot of functional motor and communication skills [3; 14] on a ve-level scale from the score I to score V (lower level indicates the best capacity). Although some studies have attempted to explore the possible correlation between these instruments [12; 13; 17], the distinction among the classi cation systems' levels, especially between high levels based on the activity and participation in natural settings, remains to be determined.

Outcome Measures of Upper Limb Functions and Participation
The ABILHAND-Kids was developed as a Rasch-based measure to report perceived manual ability of children with CP aged 6 to 15 years by their caregiver/parents [21]. The ABILHAND-Kids questionnaire consists of 21 items that are the best representative of speci c daily activities requiring one or both hand use [22]. The potential advantage of this questionnaire is its usefulness in clinical and research settings, as it can be quickly completed [21]. The total score can be calculated based on converting the raw score into a logit measure on the website of http://rssandbox.iescagilly.be [23] . Finally, ABILHAND-Kids has excellent test-retest reliability (ICC=0.98) in evaluating upper limb functions of XXX children with CP[24] The Child and Adolescent Scale of Participation (CASP) was used to documenting the extent of child's involvement in different life situations, including participation in home, school, and community activities [25]. The CASP includes 20 ordinal-scaled items as to home participation (6 items), 2) community participation (4 items), 3) school participation (5 items), and 4) home and community living activities (5 items). Its assessment items are suitable for school-aged children (5 years or older) and coincide with many activities in different life situations. Psychometric properties of the CASP, such as test-retest reliability, were demonstrated to good in Turkish children with CP (ICC=0.95) [26]

Statistical Analysis
Statistical analyses were performed using the SPSS software version 24. The variables were investigated using visual (histogram, probability plots) and analytical methods (Kolmogorov-Smirnov) to determine whether or not they are normally distributed. While investigating the associations between non-normally distributed and/or ordinal variables, the correlation coe cients and their signi cance were calculated using Spearman Test. Inter-relationship among MACS, GMFCS and CFCS was interpreted according to Spearman's correlation coe cient strength: r<0.2 very weak relationship; 0.2-0.4 weak relationship; 0.4-0.6 moderate relationship; 0.6-0.8 strong relationship; r>0.8 very strong relationship [27]. Categorical data were presented with percentage or frequency as appropriate, while continuous data were given as mean ± standard deviation. One-way ANOVA and Tukey's Test were performed to explore possible differences in mean upper limb functions and participation in different life situations scaled scores in three representatives of functional status on MACS, GMFCS, and CFCS. A univariate linear regression model was conducted to investigate separately the extent to which variables affected upper limb functions and participation in different life situations. Then, multiple linear regression analysis was performed to explore how functional motor and communication skills together contribute scaled scores of upper limb functions and participation in different life situations. To do so, level III of each classi cation system was accepted as reference category (0). Standard error bars were used to demonstrate mean scores with %95 con dence interval (CI) on assessment tools by each classi cation system level. An overall p-value of less than 0.05 was considered to show a statistically signi cant result.

Results
All study participants' demographic characteristics and prognostic variables, and mean ages of participants classi ed into different GMFCS, MACS, and CFCS levels were shown in Table I (GMFCS II), and only % 6.1 required a hand-held device for mobility. Considering that the MACS is a classi cation system of disabled children's hand function in ve levels (a lower value describes higher manual ability), most of the children in this study had effective hand use (MACS I-II). For communication skills, the predominant levels were level I (% 56.1) and level II (% 29.6) respectively, while fewer participants had a moderate score (level III; %14.3). That is, most of the study participants could independently and effectively communicate with people in most environments. Finally, except for hearing impairment, vision, speech, and cognition impairment was reported in 9 (% 9.2), 11 (%11,2), and 12 (% 12,2) of study participants, respectively. The distribution of children at different MACS levels within GMFCS levels was demonstrated in Table II. As indicated, most children in MACS level I (89,3%) were quali ed with GMFCS I; on the other hand, a very few percent (10,7%; 3 children) presented GMFCS II. Of children in MACS II, 55,2% had the best mobility level on GMFCS (level I); whereas 44,8% relatively had a lower mobility level on GMFCS (level II). Finally, children with MACS III demonstrated greater variability in mobility level on GMFCS than those with MACS I and II, with a majority part representing GMFCS II (53,7%). Table III (Table IV). Furthermore, results of pairwise post-hoc tests suggested that participants in a higher functioning GMFCS level had more remarkable outcomes in both upper limb functions and participation in different life situations than those with a lower GMFCS level (Post-Hoc: a>b>c). That is, the higher GMFCS levels were found to be associated with greater mean scores on both the ABILHAND-kids and CASP-subtests. Children in GMFCS I presented greater variability for upper limb function (ABILHAND-Kids) than those in GMFCS II and III. In contrast, children in GMFCS III demonstrated greater variability in scaled scores on CASP-community participation subdomains than those in GMFCS I and II (Figures 2 and 5) Table V shows upper limb functions and participation in different life situation outcomes by communication skills on CFCS. Results of one-way ANOVA displayed statistical meaningful differences among three levels related to both manual ability and participation in different life situation outcomes (p=0,000). However, isolated comparisons (pairwise post-hoc tests with Tukey) revealed a statistically signi cant difference only between high and moderate levels (CFCS I-III), whereas it was not found out statistically meaningful differences between level I and II, and between level II and III (Post-Hoc: a>c, a=b, b=c). In other words, the mean upper limb functions and participation scaled scores obtained by children quali ed with one of three levels of CFCS differed signi cantly only between those quali ed with levels I and III. In contrast, a signi cant difference could not be found between children classi ed into levels closer to each other for upper limb functions and participation in different life situations. Furthermore, as demonstrated in Figures 3 and 6, children quali ed with CFCS III displayed greater variability in scaled scores on both upper limb functions and participation in different life situations than those characterized as CFCS I and II.
As demonstrated in Table VI, MACS levels were found to be moderately correlated with GMFCS levels (r=0,491, p=0,000), whereas, weakly correlated with the CFCS levels (r=0.247, p=0,014). Finally, a moderate relationship was observed between GMFCS and CFCS levels (r=0,574, p=000). Table VII demonstrate the ndings of regression models. For the upper limb functions, MACS, GMFCS, and CFCS explained % 65, % 23, and % 8 of variance, respectively. Also, it was found out that the MACS was the strongest predictor of participation in home, school, and home and community, explaining % 62, % 56, and % 48 of variance. On the other hand, the MACS and GMFCS contribute equally to the scaled score of community participation, with a % 48 variance. As a result, the manual ability level as de ned by MACS is the strongest predictor of upper limb functions and participation in home and school environments compared to both gross motor and communication performance on GMFCS and CFCS (R 2 of MACS > R 2 of GMFCS and CFCS). At the same time, GMFCS is an identical predictor to MACS for participation in community environment where mobility would be more needed.

Discussion
This current study demonstrated signi cant differences among each of three levels of functional motor skills, as measured by the MACS and GMFCS, for upper limb functions and participation in different life situations, as measured by ABILHAND-kids and CASP-subtests in children with hemiplegic CP. For CFCS, a statistically meaningful difference was found only between high and moderate levels as to the same parameters mentioned above. On the other hand, no signi cant difference was described between communication performance levels closer to each other. In other words, upper limb functions and participation scaled scores clearly differed between each pair of MACS and GMFCS levels, while it was found out difference only between high and moderate levels of CFCS concerning same outcomes. Furthermore, this study has demonstrated weak to moderate inter-relationships among high to moderate functional ability representatives, as measured by the rst three levels of MACS, GMFCS, and CFCS. Our study ndings suggested that manual ability level was the strongest predictor of manual performance in daily life and participation in home and school environments. Furthermore, manual ability and mobility level were analogous to explain variance in scaled scores of participations in the community environment, and as well as GMFCS was found a second important predictor factor for other outcomes.
Consequently, MACS and GMFCS tools were complementary in predicting upper limb functions and the extent of participation in different life situations. Therefore, functional motor classi cation system tools (MACS and GMFCS) can be used as a concise and precise language by health professions, families, and researchers to describe a child's actual performance in daily life. In other words, both GMFCS and MACS can be used to predict the extent of participation in different life situations and performance in daily activities. These implications are in line with a previously published study' s results, reporting a strong relationship between MACS and ABILHAND-Kids [9]. As reported previously, the actual performance of children with CP is intimately related to gross and ne motor functions [14; 30] Although the homogeneous sample size was a strong aspect of our study, this may also have led to a limitation for the current study. Because our study participants' functional pro les changed between level I and Level III on the MACS, GMFCS, and CFCS, our raw data could include participants with lower functional status (level IV-V). Thus, further study is needed to investigate difference among all levels of classi cation systems for upper limb functions and participation in different life situations.

Conclusion
In conclusion, this study is novel in investigating        Mean Scores on ABILHAND-Kids with %95 Con dence Interval for each MACS level Mean Scores on ABILHAND-Kids with %95 Con dence Interval for each GMFCS level Mean Scores on ABILHAND-Kids with %95 Con dence Interval for each CFCS level Mean Scores on CASP subdomains with %95 Con dence Interval for each MACS level Page 19/19

Figure 5
Mean Scores on CASP subdomains with %95 Con dence Interval for each GMFCS level Mean Scores on CASP subdomains with %95 Con dence Interval for each CFCS level