Social adjustment in Chinese women following breast cancer surgery

Abstract

BACKGROUND: Most childhood cancers emergebefore age 60 months. Assessment of disease andtreatment impact in these children currently relieson proxy respondents such as parents and nurses.However, parents are not objective observers andnurses are often unfamiliar with the child. Theseproblems introduce ‘cross-informant variance’, orbias, producing inaccurate reporting of the child’sstate. No direct assessment of cancer impact onchildren aged 3–5 years old exists. We thereforedeveloped and tested one. METHOD: We com-pared parent and nurse (proxy) administeredPedsQL (standard quality of life measure) withan interviewer administered novel storybook-based, direct assessment in 91 Chinese cancerpatients aged 30–71 months, and self and proxycompleted PedsQL assessments in 155 cancerpatients aged 60–109 months of age, recruitedfrom hospitals in the Pearl Estuary region ofSouthern China: 30 children completed both story-book and PedsQL. All respondents completed atleast two assessments at different time points.Analyses were stratified by treatment status.RESULTS: The Storybook displayed five ‘feelings’factors (Procedural/separation anxiety, Symptoms,Nausea, Treatment anxiety and Home issues) (55%of variance) and three ‘Intensity’ factors (Symp-tom, Nausea and Procedural intensity) (65% ofvariance). The storybook/PedsQL gavesuperior correlations in children self-completingboth than between child storybook and proxyPedsQL of the same children. Intra-proxyPedsQL correlations were low. Proxy andchild PedsQL correlated more in older thanyounger children, whilst Storybook and PedsQLdifferentiated treatment status in younger childrenbut not older children, either self or proxyassessed. Storybook test-retest reliability withintreatment/non-treatment groups exceeded 0.8. CON-CLUSIONS: Proxy assessments are inferior todirect assessments when evaluating cancer impactson children; different proxy assessors have lowagreement. Children as young as 30 months of agecan reliably indicate their quality of life whenassessed in an age-appropriate manner, withclinically valid score patterns indicating adequatesensitivity. RESEARCH IMPLICATIONS: Inyoung children with cancer, assessment shouldwherever possible be direct. Cross-variant bias issignificant and inter-proxy ratings are poor.CLINICAL IMPLICATIONS: The Storybookillustrates the feasibility of direct assessment ofQuality of life in children as young as 30 monthsold. This method should now be adopted as thegold standard. This will enable more preciseassessment of symptom and comfort status in thesechildren. ACKNOWLEDGE MENT OF FUND -ING: Fundi ng was provided by Healt h CarePromotion Fund of the Hong Kong Governmentgrant 02030161.