Giving voice to the child perspective: psychometrics and relative precision findings for the Child Health Questionnaire self-report short form (CHQ-CF45)

Jeanne M. Landgraf, Amy van Grieken, Hein Raat

To derive and evaluate a shorter self-report Child Health Questionnaire (CHQ) legacy measure for use in research and clinical trials/care. Stepwise regression, factor analysis, and item scaling principles were used to derive and guide item selection, using data from a large general sample in the Netherlands (n = 933). Feasibility was assessed in a school sample (n = 114) and item internal consistency, discriminant validity, floor, and ceiling effects were evaluated using an external larger validation sample in the US (n = 1468). Reliabilities were estimated using Cronbach's alpha. Relative precision (RP), the ability to distinguish between clinical subgroups, was computed by comparing the proportion of variance explained by the short-form scales vs. respective full-length scales. The CHQ-CF was reduced from 87 to 45 items. The median alpha coefficient was 0.89. Ninety-seven to 100% scaling successes for item discriminant validity were observed. Floor effects were not observed; some ceiling effects were detected. RP estimates ranged from 0.73 to 1.37. The CHQ-SF45 is reliable and valid and exceeds item level scaling criteria.