Background Health-related quality of life is an important outcome. Self-report is the gold standard, but in the paediatric setting we often rely on proxy reporting. Our understanding of the differences between self- and proxy reports and the factors that influence them is limited. These differences can impact on treatment choices and the patient–doctor relationship.
Objective To evaluate differences between children's, parents' and doctors' perceptions of health states and health-related quality of life in children with chronic illness and explore factors which explain these differences.
Methods Consecutive families attending eligible clinics at a tertiary paediatric centre were invited to complete the Health Utilities Index (HUI) 23 questionnaire. Percentage agreement and kappas were calculated as a measure of the agreement between pairs. Chi-squared tests or Fisher's exact test, if appropriate, were performed to determine if there was an association between level of agreement and participant variables.
Results Data were collected for 130 parent–doctor pairs, 59 child–parent pairs and 59 child–doctor pairs. Overall health-related quality of life scores did not differ between responders, but there was poorer agreement for subjective domains. Doctor–child agreement was lower than parent–child agreement. Children with a diagnosis of cerebral palsy or chronic neurological condition were more likely to have lower inter-rater agreement for both subjective and objective domains. On the HUI2, agreement was lower for parent–child pairs when the father was the respondent. For child–doctor pairs, an increased frequency of patient–doctor visits and doctors' seniority were predictors of poorer agreement on the HUI3 and HUI2 respectively.
Conclusions We identified factors associated with level of agreement for self- and proxy reporting on the HUI23. Parent–child agreement was higher than doctor–child agreement. Patients with significant pain or emotional distress and patients with a diagnosis of severe cerebral palsy or chronic neurological conditions were more susceptible to under-reporting of subjective aspects of well-being by doctors and parents and may benefit from formal assessment of health-related quality of life in the clinical setting.