This study analyzes the impact of nurse stang (patient-to-nurse ratio) and nurse work environment on patient experiences with hospital care and its possible moderation by nurse education. A Bayesian three-level moderated mediation model is fitted to the data collected during a large European nurse workforce survey, the RN4CAST study, whose multi-level structure is taken into account.
The measure of nursing care undone behaves as a mediator in this relationship. We treat the ordinal outcome as an observation coming from a latent continuous normal distribution with unknown threshold specification.