Use of an IRT-based latent variable model to link different forms of the CES-D from the Health and Retirement Study

Abstract

The goal of this study was to calibrate depressive symptoms collected using different versions of the Centers for Epidemiologic Studies - Depression (CES-D) instrument in different waves of the Health and Retirement Study (HRS). The HRS is a prospective and nationally representative cohort study. This analysis included a sample of HRS participants, adults aged 23-85 years in 1992 who had complete data on depressive symptoms at initial 2- and 4-year follow-up interviews (N= 5,734). Depressive symptoms were assessed with the CES-D. Symptom coverage and response categories varied across study wave. The first wave (1992) used a four-category response, whereas subsequent waves (1994 and 1996) used a two-category response. A structural equations model (SEM) based in Item Response Theory (IRT) was used to calibrate symptoms and generate linked depressive symptom burden scores. Linked depressive symptom burden scores, derived from calibrated symptoms, were distributed similarly across HRS wave. Our results demonstrate the applicability of an IRT-based SEM to address a common challenge in prospective studies: changes in the content and context of symptom assessment. Future investigations may make use of our linked syndrome scores to further explore other aspects of depression from a longitudinal perspective.