10.4225/03/58b3a7937ee34 Sanagou, Masoumeh Masoumeh Sanagou Investigation of hospital-level associations with adverse outcomes following cardiac surgery Monash University 2017 Generalized estimating equations Cardiac surgery Open access and full embargo thesis(doctorate) monash:130839 Hospital-level characteristics Adverse outcomes ethesis-20141008-100630 Hierarchical modelling 2014 1959.1/1047087 2017-02-27 04:14:09 Thesis https://bridges.monash.edu/articles/thesis/Investigation_of_hospital-level_associations_with_adverse_outcomes_following_cardiac_surgery/4697218 While decreases in rates of mortality after cardiac surgery have been observed due to improvements in technology and health management, rates of adverse outcomes and complications following cardiac surgery are less clear in Australia and in particular how these rates relate to system factors. The aims of this thesis were to investigate the association of hospital-level characteristics with adverse outcomes after cardiac surgery after extending knowledge on the necessary first steps of patient-level prediction of these outcomes and the methodology of hierarchical modelling. This thesis aimed to find associations between hospital-level characteristics and patient outcome following cardiac surgery in Australia during 2001 – 2011. The patient-level information was provided from the Australian and New Zealand Society of Cardiac and Thoracic Surgeons registry and hospital-level information for 16 hospitals was provided by that registry and also the Australian and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation. In this thesis, in order to estimate the associations between hospital-level characteristics and adverse outcomes following cardiac surgery with the hierarchically structured data, patients within hospitals, marginal logistic regression using generalized estimating equations with an independent working correlation structure was used. The main findings of the studies carried out in this thesis were that hospital-level characteristics are not associated with renal failure and pneumonia following cardiac surgery in Australia in ways that we might expect. Of the available hospital factors measured, most did not exhibit an association with adverse outcomes. Of those that did, the risk of renal failure was higher among hospitals with the presence of nephrology expertise and the risk of pneumonia was higher among hospitals with more registered nurses per 100 ICU admissions and per ICU available bed. These findings were somewhat counter-intuitive and, in the context of the finding that no other measured markers of quality of care accounted for differences in renal failure or pneumonia incidence, suggesting that overall there is little evidence of hospital-level factors relating to adverse outcome risks in cardiac surgery in Australia.