Assisting Hand Assessment and Children's Hand-Use Experience Questionnaire –Observed Versus Perceived Bimanual Performance in Children with Unilateral Cerebral Palsy

ABSTRACT

Aims: To explore the differences, relationship, and extent of agreement between the Assisting Hand Assessment (AHA), measuring observed ability to perform bimanual tasks, and the Children's Hand-Use Experience Questionnaire (CHEQ), assessing experienced bimanual performance. Methods: This study investigates a convenience sample of 34 children (16 girls) with unilateral cerebral palsy aged 6–18 years (mean 12.1, SD 3.9) in a cross-sectional design. Results: The AHA and CHEQ subscales share 8–25% of their variance (R²). Bland–Altman plots for AHA and all three CHEQ subscales indicate good average agreement, with a mean difference approaching zero but large 95% confidence intervals. Limits of agreement were extremely wide, indicating considerable disagreement between AHA and CHEQ subscales. Conclusion: AHA and CHEQ seem to measure different though somewhat related constructs of bimanual performance. Results of this investigation reinforce the recommendation to use both instruments to obtain complementary information about bimanual performance including observed and perceived performance of children with unilateral cerebral palsy.

KEYWORDS.:

• Assessments
• bimanual performance
• cerebral palsy
• hand function

Bimanual performance is important in enabling children with unilateral cerebral palsy (CP) to independently manage daily activities (Sköld et al., 2004). Bimanual performance affects children's self-management in everyday life situations such as moving around, self-care (e.g., washing, toileting, dressing, and eating), assisting in household tasks, school education and play, leisure activities, and peer relationships (World Health Organization, 2007; Adolfsson, 2011). To guide treatment planning towards these requirements and evaluate treatment effects, it is important to assess bimanual performance of children in these activities. Standardized assessments of observed performance aim to be objective, comparable, and usable for this purpose; however, such assessments do not take into account the child's perspective. Perceived ability needs to be investigated by asking questions that measure the child's subjective experience. Information about children's perceptions of and satisfaction with their own performance can add valuable information about children's priorities and concerns to observation-based assessments. Which bimanual activities are important to tackle in daily life in order to facilitate participation with peers may differ when asking the child, the therapist or other professionals (Pollock and Stewart, 1998; Missiuna et al., 2006). Incorporating the child's opinion and experience into the treatment planning process is crucial in order to enhance satisfaction and optimize treatment success. Paying attention to the child's perspective in this process can enhance motivation, especially in children or teenagers who lack interest in the therapy (Bandura, 1997). This requires that therapists modify their role to that of co-workers helping children focus on addressing the most worrisome and bothersome aspects.

The Assisting Hand Assessment (AHA) is the only available assessment tool that measures bimanual performance using a standardized assessment of a spontaneous play session, and that has sufficient evidence of psychometric properties (Holmefur et al., 2007; Krumlinde-Sundholm et al., 2007; Holmefur et al., 2009). AHA captures how well the affected hand of children with a unilateral disability is spontaneously used as an assisting hand during bimanual task performance. The Children's Hand-Use Experience Questionnaire (CHEQ) measures the perceived performance of bimanual activities of daily life, reflecting children's experience of and satisfaction with their performance (Sköld et al., 2011; Wallen and Stewart, 2014). CHEQ briefly summarizes how the child performs bimanual activities (independently, bimanually), but more importantly provides insight into how the child judges her/his own performance in terms of efficacy in grasping objects, time taken to perform the activity, and whether hand function is bothersome during performance. CHEQ was considered a promising tool (Wallen and Stewart, 2014) and preliminary studies of the psychometric properties of CHEQ have indicated very good validity and test–retest reliability (Sköld et al., 2011; Amer et al., 2015).

Though AHA and CHEQ share some similarities, a key question is the extent to which they differ. Low positive correlations and wide limits of agreement between both instruments indicating large disagreement are expected, since AHA and CHEQ are thought to capture different constructs. The purpose of this study was to explore the relationship and extent of agreement using correlations and limits of agreement between standardized observations of bimanual performance measured by the AHA and the experience of this bimanual performance assessed by the CHEQ in children with upper limb dysfunction.

METHODS

Assisting Hand Assessment

The Assisting Hand Assessment (AHA) is an observational instrument developed to measure and describe how effectively children with unilateral upper limb dysfunction (obstetric brachial plexus palsy or unilateral CP), aged 18 months–12 years, use their affected hand in bimanual performance (Krumlinde-Sundholm et al., 2007). For children over 12 years old, an adolescent test kit (recently available to purchase) was used for data collection. The assessment scoring criteria are the same for the whole age range (Krumlinde-Sundholm et al., 2007). AHA 4.4 measures 22 items regarding general use of the affected limb, arm use, grasp and release, fine motor adjustment, coordination, and pace of use of the affected limb on a four-point rating scales. The ordinal scale is transformed by Rasch analysis into AHA-units ranging from 0 to 100 (Krumlinde-Sundholm, 2012). AHA was administered by a certified rater, blinded to the CHEQ results.

Children's Hand-Use Experience Questionnaire

CHEQ evaluates the child's experience of using the affected hand in bimanual activities. It was developed for children aged 6–18 years with unilateral CP, upper limb reduction deficiency, and obstetric brachial plexus palsy. CHEQ is a computer-adaptive online questionnaire consisting of 29 items and is available free of charge via www.cheq.se. CHEQ's main features are its three subscales measuring the grasp efficacy when both hands are involved (grasp efficacy), the time taken to perform the activity compared with peers (time consumption), and the experience of feeling bothered while performing the activities independently (feeling bothered). CHEQ also assesses whether children perform the activities independently and whether one or two hands are used to perform them (binary scale level). Higher scores indicate more activities performed in total or better satisfaction on the subscales. The questionnaire can be completed by parents/guardians acting as proxies for their children and/or adolescents below the age of 13 years. After completion of the online questionnaire, a report is automatically generated by the website that graphically summarizes the answers. CHEQ subscales are rated on a four-point ordinal scale and raw scores can be transformed by Rasch analysis to logits and further into a 0-100 scale (CHEQ-units) (Sköld et al., 2011; Amer et al., 2015).

Participants and Data Collection

Children aged 6–18 years diagnosed with unilateral CP who agreed to participate in the study were recruited through convenience sampling. Based on the aim of the study, a sample representing a broad range of abilities was sought (Table 1) in order to include a large variation of bimanual skills in children within the given study time-frame. Children were recruited through their attending occupational therapist from the Astrid Lindgren Children's Hospital in Stockholm over a period of one year. Inclusion criteria were unilateral CP, age of 6–18 years, and no intensive treatment or surgery of the upper limb during data collection. AHA was administered by a certified occupational therapist in the hospital. To gather CHEQ data, participants were asked to complete the online questionnaire either during their hospital visit or via www.cheq.se, returning it either by email, postal mail, or at their next appointment at the hospital. Both assessments were administered in arbitrary order for practical reasons. The time between assessments could vary by approximately 3 months as hand function is expected to be rather stable in this age group (Nordstrand, 2015; Nordstrand and Eliasson, 2013). All children had contact with local rehabilitation services, which provide no or at most one occupational therapy session per month for this age group. Informed consent was obtained from all children and parents/guardians and ethical approval was granted by the Ethics Research Committee of Karolinska Hospital in Stockholm.

TABLE 1.  Characteristics of Participants

		Frequency
Characteristic	Mean (SD)	n	%
Number of participants		34
Age in years	12.1 (3.9)
Girls		16	47
Affected side right		18/16	53/47
MACS level*
I		7	21
II		23	67
III		3	9
Respondent child/parent or guardian/both		18/10/6	53/29/18
Days between AHA and CHEQ assessment	33.2 (82.4)
AHA-units	58 (14.4)

Note: MACS: Manual Ability Classification System.

*MACS level data of one participant was missing.

Manual Ability Classification System

The Manual Ability Classification System (MACS) uses a five-level scale to describe how children with CP handle objects in everyday life (Eliasson et al., 2006). Level I indicates independence in handling objects in everyday life with only minor difficulty, whereas children at level V need assistance in handling all objects in everyday activities. Children in this study were distributed over three MACS levels (Table 1).

Statistical Analysis

AHA-units and the units for each CHEQ subscale (i.e., grasp efficacy, time consumption, and feeling bothered), ranging from 0 to 100, were used in all analyses. The sample of participants was investigated using descriptive analysis. Shapiro–Wilk tests and scatterplots were used to check for normality and outliers in AHA and all three CHEQ subscales. To investigate the strength of the relationship between AHA and the CHEQ subscales, scatter plots including the regression line and the line of identity (y = x) as an indicator of total agreement were inspected. We calculated Pearson correlation coefficients (r), considering r ≤ 0.35 as representing a weak relationship, r = 0.36–0.67 a moderate relationship, r = 0.68–1.0 a strong relationship, and r ≥ 0.9 a very strong relationship (Taylor, 1990). Large correlation coefficients are assumed to suggest greater resemblance between both instruments, whereas small correlation coefficients indicate divergence. Coefficients of determination (R²) were calculated to help interpret the meaning of the correlation coefficients. R² measures the amount of variability in one variable that is shared by another. As correlation coefficients provide no information about the agreement or disagreement between two measurement methods (Bland and Altman, 2003; Vet et al., 2011), we investigated the disagreement, and hence divergence, between both assessments using Bland–Altman plots and limits of agreement. In the Bland–Altman plots, difference scores for the two assessments (i.e., AHA and each of the three CHEQ subscales) were plotted against their mean values. The 95% limits of agreement capture the spread of the observations for each individual (Bland and Altman, 1986; Vet et al., 2011). A two-tailed paired-samples t-test at a significance level of p < 0.05 was performed to determine whether the mean differences between AHA and the CHEQ subscales differ significantly from zero. All statistical analyses were performed using SPSS 22.0, MedCalc, and Excel 2013.

RESULTS

Thirty-four children aged 6–18 years with unilateral CP were included in this study. The MACS levels of the children ranged from I to III with most children at level II. The AHA-units ranged from 22 to 87 on the 0–100 scale. Following the CHEQ age recommendations, half of the questionnaires were completed by the child alone, while the other half were completed by proxy raters, i.e., either a parent (in all cases the mother) or the child and mother jointly. AHA and CHEQ were administered with a mean of 33 days between both assessments. Further information about the participants can be found in Table 1.

Data for AHA and the CHEQ subscales were fairly normally distributed and the Shapiro–Wilk test was non-significant, supporting the assumption of data normality. Observations for AHA and all three CHEQ subscales were randomly scattered. Two outliers were found between AHA and the CHEQ subscale grasp efficacy, one outlier between AHA and the CHEQ subscale time consumption, and three outliers between AHA and the CHEQ subscale feeling bothered. All outliers were kept in the analysis as they did not reach extreme values.

Pearson correlation coefficients were weak to moderate ranging from 0.28 to 0.50 (Table 2). The points in the scatterplots were widely spread around the line of equality (y = x), indicating absence of perfect agreement between both measurement instruments and hence some degree of disagreement. The coefficient of determination (R²) indicated that AHA shares 25% of its variance with the CHEQ subscale grasp efficacy, 14% with the CHEQ subscale time consumption, and only 8% with the CHEQ subscale feeling bothered (Figure 1, Table 2).

TABLE 2.  Descriptive Values for AHA and CHEQ (Table 2A) and Pearson Correlation Coefficients r, Coefficients of Determination R²_, and Bland–Altman Analysis (Table 2B) (n = 34)

Table 2A
Units (0–100)				Mean (Range)
AHA				58  (22–87)
CHEQ Grasp efficacy				56.6  (0–95.9)
CHEQ Time consumption				55.3  (32.8–81.0)
CHEQ Feeling bothered				62.9  (36–99.3)
Table 2B
	Pearson	Coefficient of	Mean
AHA-CHEQ	correlation	determination	difference d	95% Limits
subscale	coefficient r	R^2	(95% CI)	of agreement
AHA—Grasp efficacy	0.50	0.25	1.4  (−4.4 to 7.2)	−31.0 to 33.8
AHA—Time consumption	0.37	0.14	2.7  (−2.5 to 7.9)	−26.6 to 32.0
AHA—Feeling bothered	0.28	0.08	−4.9  (−10.9 to 1.2)	−39.1 to 29.4

Note: AHA: Assisting Hand Assessment, CHEQ: Children's Hand-Use Experience Questionnaire, CI: confidence interval, SD: standard deviation.

Figure 1.  Diagram displaying the shared variability (%) between AHA and CHEQ subscales.

Bland–Altman plots for AHA and the CHEQ subscales (Figure 2) indicate good average agreement, with a mean difference (d) approaching zero being non-significant, but large 95% confidence intervals of mean differences between AHA and the CHEQ subscales (Table 2). This indicates that for 95% of the observations, the “true” mean difference between both instruments may deviate as much from zero as indicated by the 95% confidence interval of d in the same table. According to the a priori hypothesis, the limits of agreement were extremely wide (Figure 2). The difference between the two instruments lies at most ±1.96 times SD, on a scale of 0-100 units, on either side of the mean for 95% of the observations for all CHEQ subscales, indicating very large disagreement. For the CHEQ subscale grasp efficacy this means, that for 95% of the observations AHA measurements would be between 31 units below and 33.8 units above measurements by the CHEQ subscale grasp efficacy, limits of agreement are similarly large for the other two subscales time consumption and feeling bothered (Table 2).

Figure 2.  Bland–Altman plots with limits of agreement for AHA and CHEQ subscales (n = 34).

DISCUSSION

Although both AHA and CHEQ can be thought to assess bimanual performance in a similar context, the relationship and extent of agreement between the assessments is weak. This supports the hypothesis that both instruments are largely based on different and unique constructs, AHA measuring observed performance and CHEQ assessing perceived experience.

The wide limits of agreement in the Bland–Altman plots illustrate large disagreement rather than agreement between AHA and the CHEQ subscales. The smallest mean difference and narrowest limits of agreement were found for the CHEQ subscale grasp efficacy. Similar results were obtained from the Pearson correlation coefficients, indicating that AHA and the CHEQ subscales are weakly correlated, sharing only a small to very small amount of variability with each other. The CHEQ subscale grasp efficacy has the strongest, though still only moderate, correlation with AHA. The stronger correlation and larger variability shared between AHA and the CHEQ subscale grasp efficacy may be explained by the contents of the scale having much more in common with actual performance than do the subscales time consumption or feeling bothered. As demonstrated here, AHA and CHEQ apparently capture different constructs and therefore provide different outcomes.

The particular construct measured needs to be considered by clinicians when choosing an assessment in clinical practice (Wagner and Davids, 2012). In treatment planning, it is especially important to incorporate children's and guardians’ opinions, in line with the focus of client-centered care, to ensure motivation and facilitate participation in therapy (Law, 1998; Rosenbaum et al., 1998; Majnemer et al., 2008). Directing therapies towards objectives identified by children and their guardians is supported by a meta-analysis of efficacy of treatment for upper limb therapies for children with unilateral CP (Sakzewski et al., 2014). To incorporate children's and parents’ opinions into treatment planning, self-reports and perceived self-efficacy are important (Engel-Yeger et al., 2009; Vroland-Nordstrand et al., 2016). CHEQ is one of the few assessments intended to facilitate this process for children with unilateral CP, asking not only about grasp efficacy but also about time consumption and the experience of performing activities that require the use of two hands (Sköld et al., 2011; Hermansson et al., 2013; Amer et al., 2015). Through observation, AHA captures in detail the performance of the affected hand in bimanual activities, though it does not measure activities per se. AHA is mainly useful in helping therapists derive various key components of bimanual performance for use in training. It is known to be an objective measure, sensitive to the evaluation of treatment effects and follow-up investigations (Sakzewski et al., 2014), matters that require further investigation in the case of CHEQ.

As children were included in the study by means of convenience sampling, the results may be affected by selection bias. This heterogeneity of children with unilateral CP could be considered a limitation, though it does reflect the target group. The AHA-units are distributed over a scale extending from 0 to 100 (Krumlinde-Sundholm, 2012), and the variation in the sample covered most affected children, though the most severely affected and the most efficient children were missing. To investigate the full age range of CHEQ, we included children up to 18 years old, although the upper age limit for AHA has so far been 12 years. In this study, we used the recently developed version of AHA for children or adolescents over 12 years of age (personal communication with Krumlinde-Sundholm, 23 February, 2015). This version requires adaptation of the test situation but can be used with the same scoring criteria, as previously demonstrated in several studies, including (Eliasson et al., 2009). The lack of published research into the psychometric properties of the adolescents' version of AHA might be considered a limitation as might be the time lag that accompanied the AHA and CHEQ data collection. AHA results are assumed to be rather stable in this age range, and only minor changes are noted as long as AHA is not administered explicitly to assess treatment effects (Holmefur et al., 2010; Nordstrand, 2015). The reliability of CHEQ is good but so far no data are available on the development of CHEQ over time (Amer et al., 2015). In the present study, we collected ratings from children themselves or their parents/guardians if the children did not meet the required age or developmental level for self-scoring. Though differences in the perception and rating of bimanual performance are known to exist between children and their parents/guardians (Majnemer et al., 2008; Sheffler et al., 2009; Gates et al., 2010), the present study was not designed to investigate such differences. Investigations of using a different study design may also enable differentiation according to age, sex, and diagnosis.

CONCLUSION

AHA and CHEQ seem to measure different though somewhat related constructs of bimanual performance. Although this result might have been expected, it reinforces the recommendation to use both instruments to obtain complementary information about bimanual performance including observed and perceived performance of children with unilateral cerebral palsy.

ACKNOWLEDGMENTS

The authors would like to thank all children who participated in this research project and their parents/guardians. They also wish to thank all occupational therapists who helped to collect the data.

Declaration of Interest: Ann-Christin Eliasson is a stakeholder in the company Handfast AB (Inc.), which supplies training courses and test material for the Assisting Hand Assessment. The company has not been involved in or funded any part of this research. The other authors report no conflict of interest.

FUNDING

This work was funded by Stiftelsen Frimurare Barnhuset i Stockholm.

ABOUT THE AUTHORS

Ulrike C. Ryll is a clinical epidemiologist and physiotherapist working as a research assistant at the Department of Women's and Children's Health, Karolinska Institute, Stockholm. Her particular interests are in neuropediatric rehabilitation and health outcome measures. Carolien H. G. Bastiaenen is a clinical epidemiologist, physiotherapist, and Assistant Professor at the Department of Epidemiology, Maastricht University. Her expertise is in the field of health measures within the ICF framework as well as in evaluation research in the field of rehabilitation from an epidemiological perspective. Ann-Christin Eliasson is an occupational therapist and professor at the Department of Women's and Children's Health, Karolinska Institute, Stockholm. Her research focus is on children with developmental disorders with her main interest in hand function in children with cerebral palsy.