Measuring health-related quality of life in children: difficulties and challenges

Symone B Detmar

Jeanet Bruil

Over recent decades, healthcare systems have increasingly had to operate with limited resources to ensure an adequate level of care for a consistently increasing number of individuals. Consequently, decision makers in the field of health and health-service researchers have become increasingly interested in the evaluation of health and economic outcomes of diseases and care strategies to enhance cost-effective health interventions. The effect of prevention and treatment in healthcare is traditionally measured in terms of morbidity and mortality. Health-related quality of life (HRQOL) has, however, become more important as a measure of effectiveness.

HRQOL instruments aim to establish the personal evaluation of health, illness, treatment and circumstances. HRQOL is defined as: ‘individuals’ perceptions of their position in life in the context of culture and value systems in which they live, and in relation to their goals, expectations, standards and concerns’ [1].

Despite the increasing attention for HRQOL among adults, the recognition that HRQOL assessment is of major importance in chronic illness, and the fact that approximately 15–20% of children are chronically ill, research on HRQOL among children with a chronic illness has only recently began.

The use of HRQOL measures in children becomes even more important when one realizes that the number of children with a chronic illness, for example, asthma or diabetes, is increasing. Second, the treatment and survival of children with serious diagnoses (e.g., cystic fibrosis [CF] and leukemia) have improved significantly. Third, the survival rate for prematurely-born children has improved, with a corresponding increase of significant health problems in some children. Finally, it should be realized that illness and treatment not only affect children’s current HRQOL, they can also complicate the development and transition to adulthood because of physical, social or psychologic stress during childhood. Furthermore, many interventions are used for patients with nonlife-threatening illnesses such as asthma or epilepsy, in which the central aim is to improve quality of life.

In adult healthcare, there has been major progress in developing brief, easy-to-complete, self-reporting questionnaires that yield valid and reliable information about patient HRQOL. The last 10 years has seen a burgeoning interest in HRQOL assessment for children. It was in the mid-1990s that the first HRQOL instruments for children became available. Eiser and Morse performed a search in electronic databases from 1980 to July 1999 for articles relating to HRQOL measures. Additional information was sought on the internet, by hand searching and through contacts with key workers. This group identified 43 measures, 19 of which were generic and 24 disease-specific. A total of 16 measures allowed for completion by children and caregivers, and only seven allowed for completion by a proxy; the remainder (n = 17) had to be completed by the child. Three studies were identified that used adult measures with very few changes for children. Few measures were available for young children [2]. Dependent on the aim of the study and the target group, researchers now have the opportunity to choose one of the available instruments instead of having to develop a new measure.

Use of health-related quality-of-life instruments

The most common application of HRQOL assessment has been in descriptive research at a group level to characterize the burden of illness and monitor intervention outcomes. In addition, HRQOL measures can be used to provide information relevant to the distribution of scarce health service resources.

More recently, HRQOL assessments have been used for individual patients as an aid to communications during healthcare consultations. Using instruments as a guideline, topics may be raised that would otherwise be ignored, such as the influence of asthma or medication on school performance or bullying. Studies among adults indicate the positive effects of the clinical use of HRQOL instruments on both the content and evaluation of doctor–patient communications [3]. As yet, however, there have been no evaluations of the effects of use on pediatric consultations.

Generic or disease-specific

There are two types of HRQOL measures. Generic instruments have a wide scope and allow comparisons between groups (including children with various chronic illnesses and healthy children). Typically, such questionnaires address the current levels of physical, psychologic and social functioning, and a global assessment of wellbeing. Disease-specific measures are usually more sensitive to specific problems relating to a specific disease and are often related to the issues discussed in the clinic. These measures often include recent symptom experience and effects of medication. Currently, the modified categoric approach is preferred, in which both generic and disease-specific features are incorporated. While disease-specific measures are beneficial for clinical use and the assessment of the effectiveness of interventions within single patient groups, comparisons with other groups as well as the assessment of effectiveness/cost–effectiveness compared with other interventions in other groups are only possible with generic measures.

Multidimensional

Despite the multidimensional character of HRQOL, some researchers prefer a single HRQOL score. For purposes of health service distribution, a single score is clearly preferable. The situation is, however, less clear-cut in clinical areas. A summary score might conceal variations in effects on HRQOL at different stages of the disease and treatment process. In addition, contradictory trends for different aspects of HRQOL are missed. In particular, it has been argued that children’s HRQOL should be represented in a profile as children have a greater ability than adults to overachieve in some areas and to make gains in other functions or abilities within a developmental framework.

Requirements for instruments for children

Adult measures are not automatically applicable to children as they may include irrelevant domains, such as economic independence, and the wording within domains may be inappropriate. There are special requirements for children’s HRQOL measures (BOX 1). It is important they are developmentally appropriate in terms of the questions asked and by reflecting features of quality of life that change with age. Child HRQOL measures should be child-centered, employ self-report where possible, be age-related, have a generic core and disease-specific modules, and put an emphasis on HRQOL-enhancing aspects [2].

Developmentally appropriate

The problem with the development of HRQOL measures in children is the rapidly changing physical and mental maturity of children and, therefore, a host of changes in social behavior and emotional experiences during childhood. It is a challenge to give a valid characterization of the three domains for different developmental stages. A child with a chronic illness faces the same developmental tasks as a healthy child, but is also confronted with stress associated with the illness and treatment. It may also experience limitations in the mastery of developmental tasks. Both issues are of importance in developing an HRQOL instrument. In addition, the instrument should recognize the child’s understanding of health and treatment. During adolescence, illness may represent a barrier to the attainment of autonomy and independence. In addition, these children are able to understand the implications of illness more easily. The basic domains of HRQOL for adults are also assumed to be relevant for children. The development of KIDSCREEN has shown the extent to which this is justified [101].

Box 1. Requirements for a health-related quality-of-life instrument for children/adolescents

•	The instrument covers the definition of health-related quality of life
•	Instrument is based on interviews with experts, parents and children
•	Wording and format is appropriate for the children’s language and cognitive skills
•	Physical, social and psychologic domains are included
•	Self-report available from 8 years of age onwards
•	Proxy version available for parents
•	Sound psychometric qualities
•	Short and easy
•	No training needed to administer and score
•	Profile and, where appropriate, summary score

Self-report or proxy version

There is a growing consensus, at least in adult patient care, that individuals themselves are the only proper judges of their HRQOL. Despite this, assessment among children is often conducted using parental reports and children’s own judgments are often ignored.

It is important to consider the role of proxy ratings. First, they can be used as substitutes for ratings made by children. This is particularly important when children are too young or ill to provide their own ratings. Second, they can be used to complement information from children. In our opinion, children’s own reports are preferable. There are, however, circumstances where such direct assessments are not possible. Children may be too young or too ill to answer questions or complete questionnaires by themselves. Proxy ratings should be taken into account to resolve the problem of missing HRQOL data for highly vulnerable groups in clinical studies. Caution is, however, advisable when interpreting these results.

Studies reveal a low-to-moderate correlation between parental and child ratings. Parents tend to report lower HRQOL scores than their children, but the accuracy of proxy ratings also appear to be partly dependent on the specific domains of quality of life being considered. Concordance might be expected to increase with the child’s age. Better child–parent correlation is also observed for more concrete items. Given the increased dependence that may be established between parents and sick children, we might expect parent–child concordance to be greater for ill than for healthy children. This is confirmed by our own study in which the intraclass correlations were higher for the group of children with a chronic illness than for healthy children. A possible explanation might be that chronically ill children are more closely supervised by their parents. We found that parents scored more limitations in physical, cognitive and social activities, and less happy feelings. However, on the contrary, parents also reported fewer physical complaints (e.g., stomach ache, headache and tiredness) than their children. These results may indicate that, in general, parents have a more pessimistic view of their children’s functioning. In addition, the results may also indicate that internal processes, such as physical complaints, are more difficult to detect for parents. Whenever possible, dependent on the aim of the study, multirespondent assessment of HRQOL should be considered.

Research has shown that children aged between 7 and 11 years provide self-descriptions that incorporate psychologic characteristics and social comparisons. They have a situationally-based awareness of different components of self and they differentiate between mental and physical aspects of self. Children aged between 7 and 8 years begin to adjust their self-perceptions according to feedback and past experiences with success or failure. From the age of 8 years onwards, children’s self-reports become considerably more meaningful as they are more capable of reporting on their thoughts and feelings and providing more accurate information about diverse experiences and situations. A prerequisite in this case, however, is that items should be worded in language a child uses, and employ time frames that are compatible with the child’s understanding.

Based on interviews with the target group

It can be concluded that instrument development should be based on the exploration of domains and items in the target groups. The KIDSCREEN-52 item generic HRQOL questionnaire for children and adolescents was developed as a self-report measure appropriate for healthy and chronically ill children and adolescents between 8 and 18 years. The generation of domains and items was based on expert opinion [4] and children’s focus groups in seven European countries. In addition, parental questionnaires were distributed to explore parents’ opinions. The focus groups with children went smoothly with many bright discussions. It emerged that the most important aspect of HRQOL for the youngest group was family functioning. In both adolescent groups, social functioning, including the relationship with peers, was most important. In all groups, children considered physical and cognitive functioning to be less important than social functioning. These results were taken into account in designing the KIDSCREEN questionnaire for both healthy children and adolescents.

Conclusion

HRQOL measures are of major importance in the evaluation of the effects of illness and treatment in childhood. Until now, the main focus of HRQOL research in children and adolescents has been on the development of HRQOL measures. As several instruments are currently available, the time has come to study HRQOL in children to understand the effects of health and illness, to explore determinants and to enhance our understanding of evidence-based interventions. Users of HRQOL instruments should review existing questionnaires to determine whether the instrument to be used is applicable for the age group in question, whether it focuses on areas or domains relevant for the target group, and whether it includes generic and disease-specific modules. In addition, the instrument should facilitate child and parent reporting. We need to move beyond the development of new measures: this step should only be considered when no instrument is available and researchers must take into account the theoretical knowledge of children’s understanding of illness, emotions and the ability to use rating scales. Instruments should be based on interviews with parents and children, and experts, and they should be psychometrically sound and short. When parent reports are used, it is important to be aware that parent reports are valuable but not identical to the reports of children.