Exploring the environmental context effect on the HRQoL reported with the EQ-5D-Y: A pilot study on a sample of ill children at school vs. at hospital.

Introduction: In our validation study of the EQ-5D-Y [Scalone et al, 2011], participants were asked to specify for each domain why they reported problems, or figure out why another child could have. The analysis of these open-ended questions revealed that children attributed difficulties to their health conditions, but also to events related with school (e.g., demanding homework to do), sport (e.g., intensive muscular effort), or age-related developmental tasks (e.g., problems with schoolmates). Recent literature suggests that the retrieving of information from episodic memory can depend upon contextual information or cues [Lee S & Schwarz N, 2013; Aslan A et al., 2015]. Accordingly, we hypothesized that the context in which children were during the completion of the instrument (school versus hospital), rather than a pure health perception, could have influenced their responses. Objective: To investigate the environmental context-dependent effect on the retrieval of information used to explain the reported problems with the EQ-5D-Y in a sample of ill children interviewed at school versus at hospital. Methods: We compared existing data of 8-13 years old children: 25 with Acute Lymphoblastic Leukemia (ALL) recruited at hospital, and 68 children recruited at school, who specified they had a long-term disease or a disease diagnosed by a physician. We compared their HRQoL reported with the descriptive system and with the VAS, and the open-ended answers on why they reported problems in each domain. Open answers were coded by two independent coders using the following categories: a) health condition, b) lack of capability/will, c) everyday life events (e.g. family- or school-related), d) other. Results: The ALL patients (versus ill schoolchildren) reported statistically significant (chi-square test, p<0.05) higher problems in self-care (20.0% vs 4.4%) and pain/discomfort (48.0% vs 42.6%) domains. With the analysis of open-ended answers, while we did not find significant differences on the other domains, anxiety/depression revealed a significant difference (chi-square test, p<0.05): 36.4% (n=4) of ALL patients explained that they felt worried/sad/unhappy for their health condition (e.g. I have a disease, I have to come to hospital), and 6 (54.5%) attributed their feeling to everyday life events (e.g. my mother’s scolding). Instead, 82.8% (n=24) of ill schoolchildren attributed their problems to everyday/school life setting (e.g. a friend make fun of me; math test), and only one child (3.4%) attributed his problems with his disease. Discussion: The context in which children are asked to report their own health perception could influence their responses, which could result not only from different real health perceptions, as aimed and expected from the heading in the questionnaire “your health TODAY”, but also from the salience of cues in the surrounding environment. This could imply erroneous comparisons between sub-populations that are in different contexts during the reporting of their HRQoL, confirming the recent literature alert. The results of this pilot study suggest that further research is advisable to clarify the relevance of environmental context effect, and the possible necessity of improving the wording in the EQ-5D-Y, to enhance the measurement properties of the instrument.