- Top of page
- Supporting Information
Aim To compare the conceptual differences, internal consistency, and validity of the Cerebral Palsy Quality of Life Questionnaire for Children (CP QOL-Child), the Child Health Questionnaire (CHQ), and a European generic health-related quality of life (HRQOL) questionnaire (10-domain version; KIDSCREEN-10) for children with cerebral palsy (CP).
Method Two hundred and four primary caregivers (185 females [91%], 19 males [9%]) of children with CP aged 4 to 12 years (mean 8y 4mo [SD 2.51]; 112 males [55%], 92 females [46%], Gross Motor Function Classification System level I=18%, II=28%, III=14%, IV=11%, V=28%) provided demographic data and completed the CP QOL-Child, CHQ, and KIDSCREEN-10. Fifty-four children with CP aged 9 to 12 years completed the CP QOL-Child and KIDSCREEN-10.
Results The KIDSCREEN-10 and CP QOL-Child were developed to measure general HRQOL and CP-specific QOL respectively, whereas the CHQ was developed to measure functional health and well-being. In terms of internal consistency, KIDSCREEN-10 (Cronbach’s α=0.86) and CP QOL-Child (0.74–0.91) outperformed the CHQ (0.18–0.96). In terms of validity, all instruments were moderately correlated. Floor and ceiling effects, although minimal or not evident for KIDSCREEN-10 and CP QOL-Child (1–4.9%), were apparent for CHQ (0.5–62.9%).
Interpretation Conceptually and psychometrically, KIDSCREEN-10 and CP QOL-Child performed more strongly than the CHQ, for children with CP. The choice between these two instruments will depend on the questions posed and outcomes sought by the researcher or clinician.
Cerebral palsy (CP), the leading cause of physical disability in children, occurring in approximately 2 to 2.5 per 1000 live births,1 is defined as a ‘disorder of movement and posture due to a defect or lesion of the immature brain’.2 In recent years there has been increasing interest in measuring the quality of life (QOL) of children with CP. QOL, defined as ‘an overall assessment of well-being across various domains’,3 is a multidimensional construct including both health (i.e. physical, emotional, social) and nonhealth domains (i.e. finances, school, autonomy).4 Health-related quality of life (HRQOL) is a subdomain of the more global construct of QOL including domains such as physical, mental and social well-being.4
QOL instruments are increasingly being used to evaluate the effectiveness of interventions for children with CP.5 Commonly used instruments include the Cerebral Palsy Quality of Life Questionnaire for Children (CP QOL-Child),6 the Child Health Questionnaire (CHQ),7–12 a European generic health-related quality of life questionnaire (KIDSCREEN),13 the Pediatric Quality of Life Inventory,14 the Caregiver Priorities and Child Health Index of Life with Disabilities,15 the Lifestyle Assessment Questionnaire,16 the modified Caregiver Questionnaire,12 and the Pediatric Outcomes Data Collection Instrument.17 It is now becoming increasingly difficult for researchers and clinicians to synthesize data on the effectiveness of interventions from studies that rely on different QOL questionnaires and to select the most appropriate QOL instrument for their purpose. Choice of a scale should be consistent with the conceptual framework and rationale for the instrument’s development.
Two groups have compared outcome instruments for children with CP.7,12 McCarthy et al. compared health and well-being instruments, including the CHQ,7–12 the Pediatric Evaluation and Disability Inventory,18 and the Pediatric Outcomes Data Collection Instrument 17 in a sample of children with spastic CP (n=115).7 The CHQ had more floor and ceiling effects then the other two instruments; however, the Pediatric Evaluation and Disability Inventory demonstrated higher internal consistency.7 Although that study provided useful information about the CHQ, the investigators did not examine other QOL instruments.
Schneider et al. compared the CHQ7–12 with the Caregiver Questionnaire in 30 children with CP12 and found that the Caregiver Questionnaire total score and the CHQ summary scores were not significantly correlated (r=0.22–0.24).12 That study, clearly limited by a small sample size, failed to compare the conceptual differences of the instruments, including the purpose of the instrument. Additionally, the Caregiver Questionnaire does not directly measure the QOL of the child but rather measures the caregiver’s satisfaction and difficulties with the child’s progress in personal care, positioning or transferring, comfort, and interaction or communication.12
The aim of the present study was to compare the characteristics of three QOL scales for children with CP. These characteristics include conceptual differences (e.g. reason for development, number of items, domains and, reporter), reliability and internal consistency, concurrent validity with previously validated measures, frequency of missing values, and floor and ceiling effects (i.e. the proportion of participants who reported the lowest or highest possible scores). The instruments selected for this study include the CHQ, which is the most commonly used instrument to measure the QOL of children with CP.7–12 Given the consensus that QOL domains should be based on qualitative research with parents and children, the CHQ was compared with two new instruments that are based on qualitative interviews with children: the KIDSCREEN 10-domain version (KIDSCREEN-10) and the CP QOL-Child. For both of these instruments, qualitative interviews were conducted with children and parents to determine the domains of QOL and the wording of the items. KIDSCREEN-10 is a new generic HRQOL instrument,13 and CP QOL-Child is a new condition-specific QOL instrument for children with CP.6 All three of these instruments have been used to measure the QOL of children with CP.6,7,12,19
- Top of page
- Supporting Information
The results of this study provide important information about the conceptual differences, reliability and validity of three instruments used to measure the QOL of children with CP. With respect to internal consistency, and floor and ceiling effects, the CHQ, although it is clearly the most commonly used instrument,7–12 did not perform psychometrically as well as KIDSCREEN-10 (which had no ceiling effects) or CP QOL-Child. Although large numbers of floor and ceiling effects may indicate that an instrument is not sensitive enough to accurately capture the QOL of children with CP, floor and ceiling effects are complex and may be attributed to deficiencies in the wording of the items or the response options, or they may actually reflect an optimal state.
The results highlight that the three instruments are designed to measure three different constructs (refer to Table SI online): condition-specific QOL for children with CP, generic HRQOL for healthy and chronically ill children, and a generic measure of functional health and well-being. Nonetheless, some of the domains are similar, such as role/social limitations – emotional/behavioural (CHQ) and emotional well-being (CP QOL-Child), and bodily pain (CHQ) and pain and feelings about disability (CP QOL-Child). Furthermore, moderate correlations among many of the domains were found for parent-proxy report (r=0.40–0.53) and child self-report (r=0.61–0.71). The moderate correlations between the instruments do not provide insight into which instrument is superior but do provide some support for the validity of the domains.
It is important for researchers and clinicians to understand the original purpose of QOL instruments, because this may influence the direction and focus of the domains and items. For example, items in the CHQ are focused on whether the child is limited in aspects of his or her life, whereas items in the CP QOL-Child are focused on how the child feels about aspects of his or her life. Given that KIDSCREEN-10 is a generic HRQOL instrument and CP QOL-Child is a condition-specific QOL instrument, the choice between these two instruments depends on the research question. If a researcher is interested in comparing the QOL of children with CP with that of children with other conditions, KIDSCREEN-10 is suitable. If a researcher is interested in examining the effectiveness of an intervention or gaining insight into the issues that children with CP face, the CP QOL-Child may be better suited. However, it may be limiting to base a decision on the construct that the instrument purports to measure, given the overlap among the domains.
The performance of the instruments varied by reporter. The correlations between CP QOL-Child and KIDSCREEN-10 were higher for child self-report than for parent-proxy report. Furthermore, there were considerably fewer floor and ceiling effects for child self-report than for parent-proxy reported data. These differences, although clearly limited by a small sample size for child self-reports, may be attributed to the context, experience, and expectations of the reporters.
This study did not measure test–retest reliability, but this should be examined in further studies, because one of the most pressing information needs is an understanding of the instruments’ sensitivity to change. Sensitivity to change is necessary if instruments are to be used with confidence to assess the effectiveness of interventions.27 This is particularly important given that researchers have suggested that children with chronic conditions adapt to their current state.28
This study has some limitations. First, this study used KIDSCREEN-10, which produces only one summary score. Although the 10-item measure was derived from the 27-item measure (which has five domains), the results of this study might have been different if the 27-item or the 52-item (10-domain) version had been used; however, the KIDSCREEN-10 is a valid generic HRQOL instrument. Second, the sample included in this study was used to develop and validate the CP QOL-Child. Items of the CP QOL-Child that had large numbers of missing values were deleted, in order to develop an instrument that is applicable to children across the spectrum of functioning, and thus the missing values are lower than for other instruments. Third, the size of the sample of self-report data was limited, because children were required to be between the ages of 9 and 12 years and to be able to understand and respond to the questions. Fourth, the GMFCS was used to classify children with CP, but, because of ethical considerations, we obtained only basic details from the Victoria Cerebral Palsy Register, which did not contain information on motor types or typology.