Introduction

Cerebral palsy (CP) is a childhood condition in which there is a motor disability (palsy) caused by a static, nonprogressive lesion in the brain (cerebral) [1]. A population-based study of CP in the USA estimated a prevalence of 2 per 1,000 live births with 1-year survivorship [2]. CP is the most common severe physical disability affecting children. CP cannot be cured, but a host of interventions can improve functional abilities, participation, and quality of life [3]. Health-related quality of life (HRQOL) has emerged as an essential health outcome in clinical trials and health service research [4]. Although there are a number of generic HRQOL instruments applicable to children with CP [5], assessment has relied exclusively on proxy-report [6], with the exception of a recent study [7]. However, there are significant limitations to relying solely on proxy-report. Although parent proxy-report of their child’s physical functioning often correlates reasonably well with their child’s perspectives on their physical functioning, when it comes to more internal states such as emotional functioning, parent proxy-report and child self-report correlate modestly at best [8]. In recent years, there has been an increasing focus on HRQOL assessments that include child self-report [5].

The Pediatric Quality of Life Inventory (PedsQL) Measurement Model was designed to integrate the relative merits of a generic core instrument with disease-specific modules [9–16]. It has been an explicit goal of the PedsQL Measurement Model to develop and test brief measures for the broadest age group empirically feasible, specifically including child self-report for the youngest children possible [17, 18].

The PedsQL 4.0 Generic Core Scales was specifically designed for application in both healthy and patient populations [19]. The PedsQL 3.0 CP Module was designed to measure HRQOL dimensions specific to CP. Studies with PedsQL indicate that both healthy children and children with CP aged 5–18 years can self-report their HRQOL [20, 21]. A recent study supported the reliability, validity, and sensitivity of the PedsQL in pediatric CP, and the PedsQL is the only validated HRQOL instrument to span ages 5–18 years for child self-report and ages 2–18 years for parent proxy-report while maintaining item and scale construct consistency [11]. Another study supported the reliability and validity of Chinese version PedsQL Generic Core Scales in healthy children [22]. However, there is no study with the reliability and validity of Chinese version PedsQL 4.0 Generic Core Scales and 3.0 CP Module in pediatric patient with CP.

This study investigates the reliability and validity of the Chinese version PedsQL 4.0 Generic Core Scales and 3.0 CP Module in pediatric CP. We hypothesized that the Generic Core Scales would distinguish between healthy children and children with CP, and greater CP-specific symptoms or problems would be correlated with lower generic HRQOL, based on previous PedsQL findings in CP [11].

Methods

CP sample

Children with CP and their parents were recruited from the Rehabilitation Centre at the Children’s Hospital in Chongqing. Hundred and thirty families were approached for the study, and 126 families agreed to participate. Written parental informed consent and child assent were obtained.

Although self-report is considered the standard in HRQOL measurement [4], we anticipated that there would be children who would not be able to complete the PedsQL because of the severity of their cognitive impairment and/or physical impairment. PedsQL research in CP demonstrated the feasibility of interviewer-administration for children with physical impairments who had the capacity to complete the PedsQL cognitively [21]. To determine whether a child had the physical and cognitive capacity to self-report, a research assistant first asked the child’s parents whether the child was capable of completing a self-administered questionnaire. If a child was unable to read or write, as a consequence of either physical or cognitive impairment, a research assistant assessed whether the child could respond to questions verbally, through nodding, or through pointing. In these cases, the research assistant administered the questionnaire verbally and recorded the results. In no case was a child whom the parent deemed capable of comprehension unable to respond to the questionnaire. Additionally, children with comorbid diagnoses, including those that may interfere with cognitive functioning (e.g., Down syndrome), were screened from participation. Thus, for all children aged 5–18 years who did not have the capacity for self-administration but did have the capacity for self-report, the PedsQL was interviewer-administered. Otherwise, the instrument was self-administered for children aged 8–18 years and their parents. Parents (83.6% mothers) and children completed the instruments separately.

Parents were also given the option to complete the PedsQL through self-administration or interviewer-administration. Because some parents were likely to be illiterate and would need help in reading the items. Twelve parents requested an interviewer-administered PedsQL. Parents and children completed the instruments separately. The PedsQL was completed in the waiting room or in the clinical examination room. This research protocol was approved by the Institutional Review Board at Children’s Hospital in Chongqing.

Healthy children sample

Participants were recruited from a primary school. The sample consisted of the 2- to 7-year-old healthy children and their parents. The children in the healthy subgroup were selected from primary schools from children with no history of development disorders and neurological disease within the last month.

A total of 250 questionnaires were distributed to caregivers and 218 were returned, which gave a response rate of 87.2%. And self-report forms are completed by 106 children. The proxy-reports were completed by 96 (44%) mothers, by 62 (28.4%) fathers, or by other caregivers such as grandparents 21 (9.7%), with 39(17.9%) missing. The average age of the 115 boys (52.8%) and 89 girls (40.8%) was 4 years 5 months (SD 1 years 1 month; missing = 14). Self-report forms were completed by 106 5- to 7-year-old children with the help of the research assistant.

PedsQL 3.0 CP Module

The 35-item PedsQL 3.0 CP Module encompasses seven scales: (1) Daily Activities (9 items); (2) School Activities (4 items); (3) Movement and Balance (5 items); (4) Pain and Hurt (4 items); (5) Fatigue (4 items); (6) Eating Activities (5 items); and (7) Speech and Communication (4 items). For the parent report for toddlers form (aged 2–4 years), there are no School Activities or Speech and Communication scales. Furthermore, the Daily Activities and Eating Activities scales were modified to include fewer items (not all items were applicable for toddlers). The format, instructions, Likert response scale, and scoring method are identical to the PedsQL 4.0 Generic Core Scales, with higher scores indicating better HRQOL (fewer symptoms or problems). The Module Scales were developed through focus groups, cognitive interviews, pretesting, and field testing protocols. The parent proxy-report items are listed in “Appendix I”.

The Scales are comprised of parallel child self-report and parent proxy-report formats. Child self-report includes ages 5–7, 8–12, and 13–18 years. Parent proxy-report includes ages 2–4 years (toddler), 5–7 years (young child), 8–12 years (child), and 13–18 years (adolescent) and assesses parents’ perceptions of their child’s HRQOL. The instructions are formed to assess how many problem each item have during the past 1 month. A 5-point response scale is utilized across child self-report for ages 8–18 years and parent proxy-report (0 = never a problem; 1 = almost never a problem; 2 = sometimes a problem; 3 = often a problem; 4 = almost always a problem). To further increase the ease of use for the young child self-report (ages 5–7 years), the response scale is reworded and simplified to a 3-point scale (0 = not at all a problem; 2 = sometimes a problem; 4 = a lot of a problem). Items are reverse-scored and linearly transformed to a 0–100 scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, 4 = 0), so that higher scores indicate better HRQOL. Scale Scores are computed as the sum of the items divided by the number of items answered (this accounts for missing data). If more than 50% of the items in the scale are missing, the Scale Score is not computed [23]. Although there are other strategies for imputing missing values, this computation is consistent with the previous PedsQL peer-reviewed publications as well as with other well-established HRQOL measures. For this study, 100% of child respondents and 99% of parent respondents were included in the CP Module Scale Score analyses after imputing missing values.

The PedsQL 3.0 CP Module Scales were translated independently into Chinese by an English teacher and a clinical psychologist fluent in English and translated back into English by a bilingual English native speaker. After review and comments by the instrument author, the second Chinese translations of the PedsQL 3.0 CP Module Scales were tested on a panel of five child self-report and parent proxy-report respondents with cognitive interviewing methods. The cognitive interviews were conducted by two certified clinical psychologists at the participant’s home, and revisions in the translation were conducted to rectify the identified problems. Finally, the third versions were produced and proofread to be considered as final. All the results of phases were reported to the instrument author and Mapi Research Institute, which were reviewed and accepted by them.

PedsQL 4.0 Generic Core Scales

The 23-item PedsQL 4.0 Generic Core Scales encompass (1) Physical Functioning (8 items); (2) Emotional Functioning (5 items); (3) Social Functioning (5 items); and (4) School Functioning (5 items). To create the Psychosocial Health summary score, the mean is computed as the sum of the items divided by the number of items answered in the Emotional, Social, and School Functioning Scales. For this study, 100% of child respondents and 99.2% of parent respondents were included in the Generic Core Scale score analyses after imputing missing values.

Statistical analyses

Feasibility was determined from the percentage of missing values [24]. Scale internal consistency reliability was determined by calculating Cronbach’s coefficient alpha [25]. Scales with reliabilities of 0.70 or greater are recommended for comparing patient groups, while a reliability criterion of 0.90 is recommended for analyzing individual patient scale scores [26].

Construct validity for the Generic Core Scales was determined utilizing the known-groups method [4]. The known-groups method compares scale scores across groups known to differ in the health construct being investigated. Generic Core Scales scores in groups differing in known health condition (healthy children and children with CP) were computed using independent sample t-tests. To determine the magnitude of the differences, effect sizes were calculated [27]. Effect size as utilized in these analyses was calculated by taking the difference between the healthy sample mean and the CP sample mean divided by the healthy sample standard deviation. Effect sizes for differences in means are designated as small (0.20–0.49), medium (0.50–0.79), and large (≥0.80) in magnitude [27]. Construct validity for the CP Module was examined through an analysis of the intercorrelations between the Generic Core Total Scale scores and the CP Module Scale scores. Computing the intercorrelations between scales provides initial information on the construct validity of an instrument [28]. Effect size measured as the correlation between the independent variable classification and the individual scores on the dependent variable is called the effect size correlation [29]. And effect size correlations are designated as small (0.10–0.29), medium (0.30–0.49), and large (≥0.50) [27]. Intercorrelations were expected to demonstrate medium to large effect sizes.

Exploratory factor analysis was performed on the items to test the PedsQL underlying dimensions [30]. Principal component analysis with oblique rotation was performed to extract the factors. Factors with an eigenvalue less than 1.0 were disregarded.

The concordance between patient self-report and parent proxy-report was determined through correlation coefficients. Intraclass correlations (ICC) were computed, designated as <0.40 poor to fair agreement, 0.41–0.60 moderate agreement, 0.61–0.80 good agreement, and 0.81–1.00 excellent agreement [31]. ICC values were derived using a random effects model, with type absolute agreement rather than type consistency for the reliability of absolute questionnaire scores [32, 33]. Difference between proxy-report and self-report was determined by paired t-test.

Statistical analyses were conducted with SPSS (version 13.0). Response equivalence has been previously demonstrated across language for the PedsQL, as well as across the mode of administration [19].

Results

Sample characteristics

Participants were a convenience sample of children aged 5–12 years (n = 32) and parents of children aged 2–12 years (n = 126), with 130 families accrued overall. For 18 children (56.2%) aged 5–12 years, both child self-report and parent proxy-report were available. (We did not have Children with CP older than 13 years.) Fourteen children were unable to complete self-report measures due to cognitive or intelligence impairment.

For all forms combined, the mean age of the 87 boys (69.0%) and 39 girls (31.0%) was 4 years 1 month (SD 2 years 2 month) with a range of 2–13 years. For child self-report, the mean age of the 12 boys (72.2%) and 5 girls (27.8%) was 7 years 2 months (SD 1 years 7 month) with a range of 5–12 years. For all forms combined, the PedsQL was completed in Chinese.

Missing item responses

For child self-report, no item was missed. For parent proxy-report, the percentage of missing item responses was 0.79%, for all scales except the parent proxy-report School Functioning Scale. The percentage of missing items for the proxy-report School Functioning Scale was 28.1% (ages 5–12 years) and 55.3% (ages 2–4 years). This large percentage for toddlers (ages 2–4 years) may exist because many toddlers do not attend school.

Means and standard deviations

Table 1 presents the means, standard deviations, effect sizes, and t-test results of the Generic Core Scales. Table 2 presents the means and standard deviations of the CP Module. Of the children with CP who were an eligible age to self-report, 56.2% (18 children) were able to self-report. Of the 18 children who did self-report, 14 were aged 5–7 years, and per PedsQL protocol, the instrument was interviewer-administered. There were four children aged 8–12 years who required assistance in completing the self-report forms. All 18 children completed the Generic Core Scales and the CP Module.

Table 1 Scale descriptives for PedsQL 4.0 generic core scales parent proxy-report and child self-report, and comparisons between pediatric CP patients and healthy children scores
Full size table
Table 2 Scale descriptives for PedsQL CP module parent proxy-report and child self-report
Full size table

Internal consistency reliability

Internal consistency reliability coefficients are presented in Table 3. Most scales exceeded the minimum reliability standard of 0.70, and a number of scales approached or met the reliability criterion of 0.90 recommended for analyzing individual patient scores.

Table 3 PedsQL 4.0 generic core scales and CP module scales internal consistency reliability for parent proxy-report and child self-report by summary score
Full size table

Construct validity

Table 1 demonstrates the differences between healthy children and children with CP. The intercorrelations between the CP Module and the Generic Core Scales total score are shown in Table 4.

Table 4 Intercorrelations among pediatric quality of life inventory (PedsQL) scales
Full size table

Factor analysis

The results of the factor analysis for proxy-report of the PedsQL 4.0 Generic Core Scales and 3.0 CP Module are presented in Tables 5 and 6. For the PedsQL 3.0 CP Module, an eigenvalue cutoff of 1.0 resulted in a seven-factor solution for proxy-report, accounting for 87.2% of the variance. And for the Generic Core Scales, a six-factor solution for proxy-report was resulted, accounting for 82.1% of the variance.

Table 5 PedsQL 3.0 CP module scales factor loadings for parent proxy-report
Full size table
Table 6 PedsQL 4.0 generic core scales factor loadings for parent proxy-report in children with CP
Full size table

Parent/child concordance

The parent/child concordance intercorrelations matrix is shown in Table 4, with ICCs in the poor to fair agreement range. The total scores of self-report were higher than those of proxy-report, both for the Generic Core Scales (n = 18, difference = 14.64, t = 5.169, P < 0.01) and the CP Module (n = 18, difference = 11.06, t = 2.991, P < 0.01).

Discussion

The analyses support the reliability and validity of Chinese version PedsQL 4.0 Generic Core Scales and 3.0 CP Module in pediatric CP.

There were minimal missing item responses, indicating that children and their parents were able to provide good-quality data regarding the child’s HRQOL.

Both the PedsQL 4.0 Generic Core Scales and 3.0 CP Module Scales reliabilities approached or exceeded the alpha coefficient standard of 0.70 for most scales. The PedsQL 3.0 CP Module Total Score for both child self-report and parent proxy-report exceeded an alpha of 0.90, recommended for individual patient analysis [26], making the total scale score suitable as a summary score for the primary analysis of HRQOL outcome in clinical trials and other group comparisons. The total score of the Emotional Functioning subscales for parent proxy-report, Social Functioning, Fatigue, Speech and Communication subscales for self-report did not approach or exceed 0.70. Although Cronbach’s alpha represents the lower boundary of the reliability of a measurement instrument and is a conservative estimate of actual reliability [34], scales that did not achieve the 0.70 standard should be used only for descriptive analyses.

The PedsQL 4.0 Generic Core Scales distinguished HRQOL between children with CP and matched healthy children, with most effect sizes in the large range, except the Emotional Functioning subscales for child self-report. Our finding that the parent/child concordance demonstrated poor to fair agreement is consistent with the adult and child literature [35, 36], suggesting information provided by proxy-respondents is not equivalent to that reported by the patient. The total scores of proxy-report were higher than those of self-report in this study. In the HRQOL measurement of children with and without chronic illness, imperfect agreement between self-report and proxy-report has been consistently documented, particularly for less observable or internal symptoms, such as fatigue [37]. While pediatric patient self-report should be considered the standard for measuring perceived HRQOL, there may be circumstances when the child is too young, too cognitively impaired, or too ill to complete an instrument, and parent proxy-report may be needed in such cases. Further, it is typically parents’ perceptions of their children’s HRQOL and symptoms that influences health care utilization [38–40]. In cases in which pediatric patients are not able to provide self-report, reliable and valid parent proxy-report instruments are needed [41].

Most intercorrelations between the CP Module and Generic Core Scales were demonstrated medium to large effect sizes, which are consistent with the conceptualization of disease-specific symptoms as causal indicators of HRQOL.

The results of the factor analysis in general support the hypothesized factor structure of the PedsQL. For the PedsQL 4.0 Generic Core Scales and 3.0 CP Module, the total variance explained was 87.2 and 82.1% for proxy-report, respectively. Most subscales items were loading on one factor for PedsQL 3.0 CP Module, but Pain and Hurt or Eating activities items split into two different factors, and Fatigue items even seem to split into three factor loadings. For PedsQL 4.0 Generic Core Scales, all items split into two different factors. The results are quite different for self-report. All items split into two or three factors except Pain and Hurt. Poor ICCs may contribute to the difference. But the results may be unreliable because of low sample sizes of child self-report. Further study should test it with larger sample sizes. There is no previous study with factor analysis of PedsQL 3.0 CP Module items, so we cannot compare the results. For the PedsQL 4.0 Generic Core Scales, the results do not resemble Varni’s five-factor structure in the original PedsQL version [19]. In their study, only School Functioning items split into two different factors. But the findings of factor analysis may be sample-specific, another study showed different results too [42], that is why the factor structure should be reinvestigated in clinical samples. Further testing with other samples will be conducted before reconfiguring the conceptually derived Summary Scores.

The present findings have several limitations. Information on nonparticipants was not available, and the study was conducted at one pediatric hospital with a small sample size, and the healthy sample did not appear to be matched to the CP sample, which may limit the generalizability of the findings. Associations between PedsQL scores and treatment status and other medical variables were not explored due to the small number of child self-report. Test–retest reliability of the PedsQL was not conducted. External measures of construct validity were not utilized in this study. Sample sizes were not large enough, and the subject-to-item ratio was quite low for exploratory factor analysis. At least a 5-to-1 or more ideally a 10-to-1 subject-to-item ratio is recommended [43]. Since the PedsQL 3.0 CP Module was tested in this investigation with only ages 2–12, the measurement properties of the module are unknown for ages 13–18. Nevertheless, instrument validation is an iterative process and consistent with this paradigm, the PedsQL instruments are currently being further field tested nationally and internationally in pediatric CP with larger populations of children across the full age range.

This study supports the reliability and validity of Chinese version PedsQL 4.0 Generic Core Scales and 3.0 CP Module in pediatric CP. For the purposes of a clinical trial, the PedsQL 3.0 CP Module Scale in combination with 4.0 Generic Core Scales would provide an integrated HRQOL measurement model with the advantages of both generic scales and a CP-specific symptom scales, as a practical and reliable screening instrument for identifying physical and psychosocial health concerns from the perspective of both child and parent that will guide interventions to improve the outcomes of healthcare for children with CP.