Health related quality of life in Dutch infants, toddlers, and young children

Health-Related Quality of Life (HRQoL) has become an important part of patient outcome assessment in pediatrics. A HRQoL instrument should be “multidimensional, consisting at the minimum of the physical, mental, and social health dimensions delineated by the WHO” [1]. Generic HRQoL questionnaires provide a comprehensive overview of HRQoL and allow for comparisons between healthy and disease populations [2]. Questionnaires about children’s HRQoL can be self- or proxy- reported. Self-report is usually preferred, because the child’s report is the best source of information about what he or she is experiencing [3, 4]. However, when children are too young (under the age of eight [4, 5]), too cognitively impaired, too ill or too fatigued to complete a HRQoL instrument, parent/caregiver proxy-report is recommended [6].

To reliably measure the impact of a chronic condition on children’s HRQoL compared to their healthy peers, it is important to use validated questionnaires. So far, several validated age-appropriate questionnaires to measure HRQoL in infants (0–1 years), toddlers (2–4 years), and young children (5–7 years) are available, such as the Infant and Toddler Quality Of Life questionnaire (ITQOL) [7], TNO AZL Preschool Children Quality of Life (TAPQOL) [8], the Child Health Questionnaire (CHQ) [9], and the Pediatric Quality of Life Inventory 4.0 (PedsQL) [10].

Electronic administration of measures is becoming more common and offers several advantages over pencil and paper formats. For example, data can be automatically scored and extracted, and it can be stored in the electronic medical record. The electronic questionnaires are a literal reproduction of the original paper-pencil questionnaires. Research to date suggests that data obtained by paper versus electronic rating scales are equivalent, and there are no systematic differences in outcomes with the two approaches [11]. Furthermore, patients and healthcare providers generally prefer electronic measures above paper-pencil measures [11, 12].

As part of a project on the implementation of an electronic system for the routine monitoring of patient-reported outcomes in children with a chronic disease or cancer [13, 14], electronic versions of the TAPQOL and the PedsQL 4.0 (both internationally widely used questionnaires to measure HRQoL in children) have been developed. However, for infants aged 0–1 years (TAPQOL), toddlers (PedsQL 4.0, 2–4 years), and young children (PedsQL 4.0, 5–7 years) no (sufficiently reliable) Dutch normative data was available. For the PedsQL 4.0, normed within the Netherlands for ages 5–18 [15] and 18–30 years [16], no Dutch reference data existed for 2–4 year-old children and the reference sample of the 5–7 year old group was rather small.

Therefore, the aim of the current study was to provide Dutch reference data for the TAPQOL in infants aged 0–1 years and for the PedsQL 4.0 in toddlers aged 2–4 years, and in young children aged 5–7 years. Possible gender differences in the HRQoL of infants, toddlers, and young children were explored. Finally, we aimed to assess the internal consistency and to explore known-groups validity of the TAPQOL and PedsQL 4.0 in the concerning age groups.

The current study aimed to provide Dutch HRQoL normative data for the TAPQOL for infants (aged 0-1 years), the PedsQL 4.0 for toddlers (aged 2–4 years), and the PedsQL 4.0 for young children (aged 5–7 years). Furthermore, we examined the influence of gender on parent-reported HRQoL and we assessed the internal consistency and known-groups validity of the TAPQOL and PedsQL 4.0 in the concerning age groups.

Overall, no gender effect was found in the HRQoL of infants, toddlers, and young children. Yet, infant girls had less behavioral problems than infant boys on the TAPQOL ‘problem behavior scale’. This in line with a study that looked at sex differences in children up to 8 years of age, where parents reported that infant girls showed less behavioral problems than infant boys [21].

Moderate to good reliability was found for all TAPQOL domains, except for the TAPQOL ‘stomach’ scale (α = .39). The original study conducted by Fekkes et al. [8] on the psychometric properties of the TAPQOL in 1–5 year-old healthy children, also found a rather weak reliability for the ‘stomach’ scale (α = .51). The items included in this scale measure cramps, stomach or belly pain, and nausea. These items do not have to necessarily correlate with each other, which might have accounted for the low reliability we found for this scale. For example, infants experiencing cramps do not have to be nauseous per definition. Thus, cautiousness is warranted with using this scale to measure stomach problems in infants.

In our sample of toddlers aged 2–4 years, all PedsQL 4.0 scales reached moderate to good reliability. The lowest Cronbach’s alpha was found for the PedsQL 4.0 ‘school functioning’ domain (α = .60). Varni, Limbers, and Burwinkle [22] also found the lowest alpha for the PedsQL 4.0 ‘school functioning’ domain in the 2–4 year-old age group (α = .59–.69). Relatively lower Cronbach’s alphas might be related to the small number of items included in a scale [22, 23]. In the 5–7 year-old age group, the PedsQL 4.0 had satisfactory to good reliability for all domains. Across the different age groups, Cronbach’s alphas seemed to increase slightly with the child’s age. Similar results were found in a study comparing children’s proxy-reported HRQoL across different age groups, using the PedsQL 4.0 [22]. In general, this study showed that the TAPQOL and PedsQL 4.0 are reliable instruments to measure HRQoL in infants, toddlers and young children in the Dutch normative population. Scales with lower reliability coefficients should only be used for explorative purposes [22].

Relating to the known-groups validity, for young children (aged 5–7 years), differences in scores between healthy and chronically ill children were found on all PedsQL 4.0 domains, indicating the PedsQL 4.0 to be a valid instrument to distinguish HRQoL in healthy children and in children with a chronic condition in this age category. The TAPQOL only differentiated between children with and without a chronic health condition in 3 out of 12 domains. The initial study performed by Fekkes et al. [8] among children 1–5 years also found that several scales of the TAPQOL did not differentiate between children with and without a chronic health condition. Toddlers with a chronic health condition differed from healthy toddlers on 2 out of 6 domains on the PedsQL 4.0. The study conducted by Varni, Limble and Burwinkle [22] found significant differences for all PedsQL 4.0 domains in this age group. Yet, analyses in their study were conducted on a larger sample (N = 617) of toddlers with a chronic health condition, which might have enabled them to find significant differences. It is therefore the question whether, in very young children (i.e., infants or toddlers), a valid difference can be made in HRQoL between children with and without a chronic health condition. Despite the fact that average HRQoL scores may not always differentiate between children with and without a chronic health condition, poor HRQoL scores may still provide clinically meaningful information [24].

This study has some limitations to note. First, even though we did our best to deliver a sample that was as representative for the general Dutch population as possible, participants and non-participants still slightly differed with regard to the child’s age, parental gender, and parental educational level. It is widely known that women and highly educated people are often more willing to participate in research. Second, as mentioned earlier, the sample size of children with a chronic health condition in the sample of infants and toddlers was small and it might have been the case that we did not have enough power to be able to find significant differences between children with and without a chronic health condition. Future studies should include larger age subgroups of infants and toddlers with a chronic health condition, such that can be sufficiently determined whether the TAPQOL and PedsQL 4.0 differentiate HRQoL in children with and without a chronic health condition. Third, parents judged themselves whether their child had a chronic health condition. It could therefore be the case that our sample reflected a slightly different group of children with a chronic condition than when this would have been directly diagnosed by a physician. Finally, since this study was related to a larger project on the implementation of patient-reported outcomes in clinical practice [14], we chose to provide Dutch normative data for brief infant, toddler, and young child electronic HRQoL questionnaires that were already available in Dutch and in use by our hospital at that time. Since 94% of households in the Netherlands have access to the internet, we think it is safe to say that using electronic measures did not have a tremendous impact on the generalizability of the findings [25]. Yet, other possible reliable and valid paper-pencil HRQoL measures (e.g., the PedsQL 4.0 infants scales [24], or the Infant and Toddler Quality of Life questionnaire [7]) are worth studying as well. Future research could determine which questionnaires are best in assessing HRQoL in infants and toddlers.