Methodological aspects of the lidA cohort study – pretest analyses of response and psychometric properties

Abstract

Introduction: High job demands (e.g. flexibility, continuous qualification) and rapidly changing work environments are challenges for an aging workforce, thus demanding concepts for maintaining work ability and health of older employees. The lidA-study is a commencing German longitudinal study investigating occupational, individual and socioeconomic factors for health and professional transitions of aging employees thus requiring careful consideration of study design, sampling, and selection of indicators and confounders. A cohort sequential design with two age cohorts followed over 12 years (intervals: 3 years) should allow identification of age, time and generation effects. A random sample of 6.000 employed and unemployed persons in 200 sample points allows representativity for the working population. Methods: Primary data (CAPI) on work characteristics, individual factors and health combined with secondary data of a social security registry and the statutory health insurance will provide a work-health matrix with indicators of expositions, confounders and outcomes. The pilot study (n=200) analyzes the applicability of the CAPI instrument, aspects of possible (e.g. social) bias in response and acceptance of conjoining secondary data. Psychometric properties of applied items and scales will be analyzed by classical and probabilistic test theory approaches (e.g. internal consistency, construct validity). Expected Results: The pilot study will provide results on response and acceptance for the data collection strategy in the source population. We expect varying psychometric properties of the selected survey instruments and possible differing response patterns for items and scales. Results of the pilot study will be available at the time of the conference. Discussion Large cohort studies have to be designed carefully, ideally taking future developments into account. Selection of survey instruments has to be based on adequate psychometric properties and reasonable observed variance within expected predictors and confounders for a longitudinal study. The results of the pilot study will show whether the assembled survey instruments can fulfill this criteria.