Historically, health-related quality of life (HRQL) assessment in pediatrics, including the few validated instruments in pediatric oncology, has been based on proxy reporting, relying primarily on parental assessment. Children have been deemed incapable of providing consistent and reliable information about their level of functioning or state of well-being. Previous studies have been hampered by either limited or poor correlation among the proxy reporters, i.e., teachers, parents and physicians, and in comparisons to disease severity. Simply stated, proxy reporters have greater agreement about what the child can do vs. what the child thinks or feels. Comparisons among proxy reporters have been hindered also by a lack of parallel content in the instruments used, which may result in poorly congruent assessments simply because the instruments measure different constructs. In addition to the measurement issues, the emotional milieu of the parent, particularly the mother, has been shown to influence assessments of the child's functioning. Maternal distress, marital adjustment and health locus of control all co-vary with reports of the child's behavior. What, then, is the proxy reporter telling us about the child? We conducted a cross-sectional study of school-aged pediatric bone marrow transplant (BMT) patients at our institution to evaluate children's self-reported HRQL and functional status. We formally tested the Child Health Rating Inventories (CHRIs), a recently developed generic health-status measure, with its companion measure, the Disease Impairment Inventories-Bone Marrow Transplant (DSII-BMT). Separate questionnaires were administered to patients, parents and physicians at a scheduled outpatient visit after BMT. The questionnaires were designed to have parallel content. All responses were confidential. The psychometric properties of the CHRIs and DSII-BMT are reported elsewhere. In brief, the responses of all raters were reliable, based on measurements of internal consistency. The children's self-reported health status was correlated significantly with the physicians' disease severity rating (DSR) across all generic and disease-specific domains. In contrast, parental reports of child health status were not correlated significantly with the DSR for disease-specific problems or the child's pain. Parental ratings deviated most from the children's ratings within the dimensions of mental health and quality of life (p < 0.001). For the entire sample, parental ratings were significantly lower than the children's ratings. Within the subgroup "early after transplant (<6 months)", parental ratings were significantly lower than the children's self-reports in all categories. In the subgroup ">12 months after transplant", with the exception of mental health and quality of life, parental scores were the same as or higher than the children's ratings. Our results confirm previous studies that the parental reporting of children's health status is a complex construct and that valuable information can be elicited directly from the children. Further research is needed to substantiate these findings, particularly in longitudinal applications with adequate sample sizes.
Copyright 1999 Wiley-Liss, Inc.