Effective use of diagnostics: application of machine learning solutions in clinical practice

The main goal of this thesis was to provide more insight into strategies to improve the effectiveness of diagnostics in the hospital setting, focusing mainly on reducing unnecessary laboratory and microbiology tests. The first two chapters describe the effectiveness and sustainability of already implemented strategies. The other chapters present new strategies such as machine learning algorithms to reduce unnecessary blood and urine cultures. In addition, we outline the required steps to implement our machine learning algorithm predicting blood culture outcomes and pursue crossing the gap between the development of machine learning algorithms and implementing these algorithms in day-to-day practice.