A general consensus has been reached by authors in regard to the three dimensions that include most aspects of QoL: physical functioning, psychological functioning, and social functioning. However, many authors define more specific dimensions. For example, the WHO has identified six respective dimensions for QoL [12]. These dimensions integrate the concepts of health status and subjective QoL:

Empirical findings in pediatric oncology have shown that the QoL of children is based on (1) the child’s symptoms, (2) the level of participation in usual activities, (3) social and family interactions, (4) physical health, as well as (5) emotions and (6) the perception of being ill [14]. These results come from two pilot studies. The first had a cross-sectional design, and the sample included 23 patients who were between 8 and 15 years of age. The second, a longitudinal study with four times of assessment spanning over 2 years, included 13 patients who were 10–18 years old [14]. Both studies were conducted according to a semistructured interview consisting of four questions defined by the researchers (“What makes a good day for you? What makes a bad day for you? Are there some things you like to do that you cannot do now? How has being sick been for you?”), and participants’ responses were analyzed using a semantic-content method [14]. This study also showed that the definition of QoL differs depending on whether you ask the children or caregivers, which thereby confirms the need to interview key stakeholders.

Who is the best person to determine the QoL of an individual? In the context of pediatric neuro-oncology, it is often inevitable that the QoL of the child is rated by a third party, preferably by one of the parents, given the child’s suffering and physical and cognitive difficulties related to their illness. Some major questions arise and this raises ethical issues. Is self-reported QoL equivalent to the assessment of QoL by a third party?

To what extent can a parent’s assessment of their child’s QoL replace the child’s own self-reported QoL? Such questions are particularly relevant in neuro-oncology where the ability to respond to verbal material is not always guaranteed. There are two opposing positions on this issue. On the one hand, some researchers attribute a great deal of importance to the individual and subjective nature of QoL and suggest that the child is the best source to evaluate their own QoL, and parental assessments can never replace that of the child [8]. This is why some questionnaires were developed to allow the child to respond in a practical, nonverbal manner (cf. PedsQL preschool version).

On the other hand, other researchers are skeptical about the reliability of self-report measures that are reported by a child, and they perceive the child as being limited in their ability to understand and report their own well-being [8]. Within this framework, the proposed solution is the assessment of the QoL by a third party, preferably the parents or person responsible for the child. This is referred to as a “proxy” [8]. In this way, the point of view of the parent would be perceived as more reliable, valid, and objective than that of the child. However, the assessment of QoL by a third party suggests that the evaluator has the ability to put themselves in the place of the child in order to capture the thoughts and emotions of the child so that they can report the child’s perception of QoL and the impact of the illness in their life. This calls on sophisticated skills and a good knowledge of the child’s perception of himself or herself. Everyone does not have these abilities, which are necessary in the evaluation of the QoL of the child. This is why, whenever possible, it is recommended to assess the QoL of the child using questionnaires reported by the child himself or herself and by the parent. In this way, the two viewpoints are considered, and issues with both methods may be taken into account.

Studies assessing the QoL of children with cancer using self-report and parent-report versions allow for observation of the differences between child and parent reports. Parents generally tend to report higher scores on distress, suffering, and deficits than the child himself or herself [8, 15]. A systematic review measuring the QoL of children with chronic illness, which differentiated between self-reports and parent reports, observed that the areas related to physical functioning, such as physical activity, symptoms, and physical complaints, are evaluated similarly by parents and children, while areas affecting psychological and social dimensions, such as emotions, autonomy, appearance, communication, and mood, give rise to more discrepancies between respondents [15]. Questions regarding physical functioning have correlational effects of r ≥ 0.50 between self-reports and parent reports, while questions on psychological and social domains result in lower correlations r < 0.30 [15]. This suggests that it will be more difficult for parents to accurately assess psychological and social aspects of their child than to assess physical functioning, which is probably more easily observable. Some researchers also speculate that parents tend to project their own distress and symptoms in their assessment of the QoL of their child, reporting a lower perceived QoL for their child than it is in reality. However, this has not been shown systematically yet. A study that assessed psychological functioning of adolescents with brain tumors using self-report, parent-report, and teacher-report versions of the Behavioral Assessment System for Children (BASC) obtained significant differences between QoL scores of different respondents [16]. For the depression dimension, 7 % of adolescents reported a clinically significant score compared to 16 % when considering parent reports and 10 % in teacher reports. For the anxiety measure, 3 % of adolescents report a clinically significant score, while the parents and teachers reports were 12 % and 19 %, respectively [16]. These results demonstrate that there is a substantial difference between the QoL perception of the child assessed by the child himself or herself and that perceived by the parent or a third party, such as a teacher.

The assessment of QoL raises another issue related to adjustment to health status across time. This adaptation is referred to as response shift. Response shift is defined as a change in the way in which an individual evaluates his or her own QoL [17]. This change in attitude is the result of a change in the three aspects of the QoL perception of the individual. It can affect (1) the individual’s internal standards, (2) personal values, and (3) the conceptualization of QoL over time [17]. In fact, the evaluation of QoL with questionnaires is based on standards, standards of what constitutes a “good” QoL according to the authors of the test, and standards that participants refer to in evaluating their own QoL [17]. An individual who is treated for an illness, refers to his or her life before illness, refers to his or her life before, more or less prescriptively, in order to determine his or her current QoL. The response shift phenomenon is especially present in long-term follow-up in stable conditions, such as remitted cancer. With time, some individuals compare their current QoL to the QoL they had during their treatments, which in turn leads to assess their QoL as better than it was during their treatment. It is a process of reassessing their own personal standards. Response shift may explain the results of some studies in which participants with brain tumors or brain tumor survivors evaluate their QoL as good or sometimes better than the control group of healthy individuals even though their health status has deteriorated [6, 9, 16]. Among these studies, that of Nathan et al., (2007) evaluates the QoL of adult survivors of neuroblastomas and Wilms’ tumors in childhood [2]. The results of this study indicate that the QoL of survivors does not differ from that of the normative population. However, these counterintuitive results on the QoL of brain tumors survivors are found in a limited number of studies, compared to the numerous studies demonstrating a poorer QoL in this population. If QoL is used as an important outcome in treatment, it is necessary to take the response shift phenomenon into account. This can be measured with different methods and is grouped into six types of measures. First, response shift can be measured using an individualized method, which integrates the perception and feedback of the patient on his or her definition of QoL, and it includes interviews or paper-and-pencil questionnaires [17]. Second, the preference–based method focuses on the importance and the value attributed to the patient’s state of health and its various dimensions through the use of questionnaires. Third, the successive comparison approach is used to evaluate the patient’s judgment on the order of importance of different psychological, physical, and social concepts. These concepts are scored on a continuum, and it must be ordered according to the value the patient attributes to them. Fourth, the design approach measures the response shift with questionnaires on the patient’s retrospective perception of changes in their standards, ideals, and values perceived after treatment, compared with those before the diagnosis. Fifth, the statistical approach evaluates longitudinal data on the QoL of patients in order to observe the evolution of response shift [17]. Finally, the last approach is the qualitative method for assessing the response shift, and this involves using semistructured interviews. To our knowledge, this phenomenon has not been systematically studied in pediatric oncology. Future advances in the assessment of response shift will account for this phenomenon in measuring QoL of cancer survivors in order to study the adaptive and nonadaptive long-term implications of response shift. This is a central issue in the use of QoL as a long-term outcome in oncology. In children, the analysis of response shift will also be complicated by developmental changes.

There are different ways to assess QoL in children. Cremeens et al. conducted a literature review of QoL measures in children between 3 and 8 years of age [18]. Some of these measures have been used in pediatric neuro-oncology. Four different formats of QoL instruments were identified and included questionnaires or Likert-based scales, graphical measures, measures based on facial expressions including photographs, cartoons, and visual measures based on a visual, linear scale of QoL [18]. Despite children’s difficulties in responding to those questionnaires, the QoL of children is currently primarily evaluated using questionnaires based on Likert-type scales [18]. There are several QoL questionnaires, some of which are generic in nature. Others are specific to pediatric oncology populations. Table 25.2 presents five QoL instruments that have been used in pediatric oncology. References of these five instruments are in Table 25.2. To date, only the PedsQL Brain Tumor Module has been validated specifically with a neuro-oncology population of children. The other instruments have been used in pediatric oncology, but have not been developed for children with cancer.