Adjusting for Misclassification in the Analysis of Self-Rated Health: Evidence from Italy

Eleonora Trappolini, Sapienza University of Rome
Maria Felice Arezzo , Università di Roma "La Sapienza"
Giuseppina Guagnano, Sapienza University of Rome

Population health evaluation is vital, especially in aging populations. Self-rated health (SRH) is a widely-used tool in epidemiology, predicting health outcomes. However, concerns about its subjectivity and misclassification have been raised, particularly among disadvantaged groups. Measurement reliability is crucial for valid analysis, especially in health disparities research. This study proposes an approach to adjust regression estimates related to SRH and assess the relationship between SRH and predictors among the elderly in Italy. We used 2019 Italian data from the European Health Interview Survey for individuals aged 50 and above, and we conduct two separate analyses. We first model the association between SRH and a set of typical predictors by applying a standard ordered logit models, and then an adjusted one to account for potential misclassification in SRH and we compare the results. In particular, we expect the association between the outcome variable and the independent variables will be weakened in the naive model when misclassification is present. In addition, higher education and stable relationship are expected to be associated with better SRH. This study addresses the subjectivity and potential misclassification in SRH. By applying adjusted models, we aim to obtain more accurate estimates in understanding SRH and its relationship with various predictors, contributing to better population health assessment.

See extended abstract

 Presented in Session 105. Measuring Health, Wellbeing and Morbidity