The present study describes the psychometric properties of the Brazilian version of the PedsQL™ 3.0 Cardiac Module. The results attest to the reliability, reproducibility, and validity of that version for the assessment of the HRQOL of children 5 to 18 years old and of their parents or caregivers within the context of pediatric cardiology.
The contextualization of the problems (domains) was understood by both the patients and their parents. Despite the low degree of schooling of some patients and their caregivers, the self-reporting method was selected, and assistance was given to patients or parents in case they had doubts, as in the original study .
Among the universe of heart diseases that affect children, a homogeneous group of patients with rheumatic heart disease was chosen for the validation study of the questionnaire because it is a disabling disease  with negative impacts on the QOL of patients and their relatives [5–8] due to the chronic valve disease, whose tertiary prophylaxis represents 35% of the surgical procedures for cardiovascular diseases in Brazil .
The patients’ scores on the Generic Scale were lower compared with a small sample of healthy Brazilian children  and the normative US population  relative to the physical and psychosocial dimensions perceived by both parents and children. Although the PedsQL 4.0 was not applied to local healthy control groups, the previously mentioned differences suggest that the PedsQL 4.0 Generic Scale can distinguish between the HRQOL of healthy children and those with heart disease, as in the original version .
Comparison of the Generic Scale scores of the various groups of children with heart disease showed that the physical functioning of the present sample was similar to that observed by Uzark et al.  and slightly better than the sample described by Berkes et al. . This observation is most likely related to the large number of children with mild heart disease who did not need specific medication. However, the psychosocial functioning scores were lower compared with the scores of children from the US  and Hungary .
The same tendency was found in the HRQOL scores from the specific scale (PedsQL™ 3.0 Cardiac Module) compared with children from the US16 and Hungary .
Just as in the case of the Generic Scale, the average score from the physical domain might be related to the fact that 67.9% of the sample included patients with mild disease. The physical functioning of children with mild heart disease does not significantly differ from the scores of healthy children, as shown by Uzark et al. [5, 16].
Although heart disease exerts direct influence on physical health from the medical point of view, most studies on HRQOL also report low scores in the psychosocial dimensions [5, 6, 16, 17, 32], as were found in the present study. These findings might point to the “new hidden morbidity” , meaning problems that affect psychosocial health, whose identification is crucial for the promotion of integrated healthcare for children.
High scores in the social functioning dimension possibly indicate successful integration of children and adolescents with heart disease with their peers . Low scores in the emotional dimension reflect the suffering of children from their chronic condition .
Although all the patients in the present study attend school, the school functioning dimension was the most affected by disease. However, it should be noted that the low scores in the school functioning dimension might also be the result of the educational conditions of the Brazilian population.
The Brazilian version of the PedsQL™ 3.0 Cardiac Module was shown to be reliable, similar to the Hungarian version . The reliability of the internal consistency between the parents’ perceptions and the children 5 to 18 years old was higher than the standard minimum value of the alpha coefficient for group comparisons. The total internal consistency of the PedsQL™ 3.0 Cardiac Module was considered good based on Cronbach’s alphas of 0.84 and 0.85 relative to the patients and the parents, respectively.
The cognitive problems subscale as assessed by the patients from all three age ranges, the treatment problems subscale as assessed by patients who were 8 to 18 years old, and the treatment problems subscale as assessed by the parents of patients from all three age ranges exhibited poor reliability and internal consistency. The small size of the sample most likely interfered with the precision of the results [5, 16]. In addition, the Cronbach’s alpha values might be influenced by the educational level of the sample . Cronbach’s alpha was higher than 0.70 in the remainder of the scales, which means that each separate subscale might be used to measure specific domains of the HRQOL of patients with heart disease.
Analysis of the item-total correlation using Spearman’s correlation coefficient showed that the instrument’s homogeneity is satisfactory. The value of the coefficient correlation was lower than 0.30 in only two cases: the physical domain (heart problems and treatment) and problems with treatment.
In the physical domain, the item “My lips turn blue when I run” (r = 0.23, patients) exhibited low correlation because cyanosis is not a component of the investigated disease. In the problems with treatment, the item “My heart medication makes me feel sick” (r = 0.26, children, r = 0.09, parents of children 5 to 7 years old) exhibited low correlation because 67.9% of the sample had mild disease and did not use specific medication but only performed secondary prophylaxis.
The intraclass correlation used to assess the reproducibility of all scales and the total score was good; its value varied from 0.76 and 0.94. This measurement property is important to establish whether an instrument is reproducible over time, i.e., whether the scores are similar in the same individual at different moments, provided his or her clinical state has not changed.
The interval between any two measurements must be long enough to reduce the memory artifacts and short enough to reduce possible systemic alterations. Even in an arbitrary manner, intervals of 7 to 14 days are recommended [27, 28].
Other studies have used analysis of discriminant validity as a model to distinguish between groups known to be different [16, 17, 36]. The results of the present study strengthen the hypothesis that children with more severe heart disease exhibit lower scores on the PedsQL™ 3.0 Cardiac Module compared with children with milder disease. Therefore, greater severity of disease was associated with poorer QOL and greater limitations and difficulties.
The symptoms scale was able to distinguish significantly between both groups of patients in the patients’ self-assessment and the parents’ proxy assessment; the average score for this item was lower (65.3) in the patients with moderate/severe disease, according to their parents’ judgment. Parents tend to express lower expectations, underestimate physical abilities, and overestimate the impact of physical functioning on the psychosocial wellbeing of their children, as shown by Casey et al. , after comparing the results of ergometric tests with the parents’ estimates of their children’s tolerance to exercise.
The patients with moderate and severe heart disease exhibited lower HRQOL scores on the PedsQL™ 3.0 Cardiac Module as a function of the limitations imposed by the disease on everyday activities. In addition, the non-physical domains were also affected, such as problems with physical appearance and cognitive and communication problems, which might be due to neuropsychological deficits associated with the disease itself or its treatment [8, 37, 38].
The correlations between the Cardiac Module Scales and the Generic Scale were significant, which agrees with the formulated hypothesis, and the convergent validity of the PedsQL™ 3.0 Cardiac Module was demonstrated, as the original16 study and other studies [17, 36] have also shown.
Despite its statistical significance, the correlation among the patients between problems with perceived physical appearance and the emotional dimension was low, which might indicate the presence of hidden anxiety or distress. The feeling of being different from peers or of experiencing discrimination by peers due to the presence of scars on the chest might contribute to such findings.
Analysis of the correlation between the patients’ and caregivers’ scores revealed strong and significant agreement in all the Cardiac Module domains, which agrees with the reports in the literature [16, 17, 39, 40]. The highest correlation corresponded to the physical domain (r = 0.86), and the lowest correlation corresponded to the non-physical domains, varying from r = 0.50 (treatment anxiety) to r = 0.71 (problems with perceived physical appearance).
The results show that the parents’ perceptions relative to the HRQOL of children with heart disease as assessed by the PedsQL™ 3.0 Cardiac Module are reliable. However, the assessments of non-physical domains, such as treatment anxiety, communication problems, cognitive problems, and problems with perceived physical appearance, are difficult to measure due to their subjective nature. This fact strengthens the need to apply specific questionnaires on QOL to children and adolescents, rather than to their parents or caregivers alone.
The limitations of the present study must be emphasized. The inclusion criteria resulted in a small sample size because the homogeneity of the sample was prioritized by selecting one single type of heart disease. The results would have been richer and improved the reliability of some domains, such as problems with treatment, if other types of heart disease, including congenital types, would have been included.
The lack of validation studies for instruments that assess the HRQOL of children and adolescents with heart disease in Brazil hindered comparison of the results. In addition, only the Hungarian version of the PedsQL™ 3.0 Cardiac Module has been validated.
Finally, the Brazilian version of the PedsQL™ 3.0 Cardiac Module exhibited adequate reliability, reproducibility, and construct validity. The results of the present study suggest that the PedsQL™ 3.0 Cardiac Module might be used as a parameter to measure the impact of heart disease on the QOL of children and adolescents.
The scores on the physical and non-physical domains confirm the need to subject children and adolescents with rheumatic heart disease to global assessment of their HRQOL. In cases where the scores from the physical domain do not agree with the severity of disease, communication to elucidate misconceptions and soothe fears, ergometric tests to provide reassurance, and prescriptions of exercise or rehabilitation might improve psychosocial QOL.
Identification of problems in the domains of perceived physical appearance, treatment anxiety, and communication problems involves referral to psychosocial treatment programs and the need to enhance participation in social activities, such as camps.
Early identification of low scores in the cognitive domain, including attention problems, indicates the need for more thorough assessment, including neuropsychological evaluation.