Research synthesis, translation and implementation of non-invasive liver tests

The development of medical diagnostic tests has been an exciting field of research, filled with enormous potential and, at the same time, uncertainty. Despite the great enthusiasm and substantial investment in the biomarker development, there is a paucity of approved diagnostic tests in the clinic. One explanation for this imbalance can be the complicated path for biomarkers from bench to bedside. This lengthy and costly process can be challenging and even frustrating, not only for the test developers, but also for researchers and clinicians. This thesis report studies that highlighted the challenges in the biomarker evaluation field. We observed that in addition to the common methodological challenges in health research, evidence synthesis for evaluating medical tests may face added challenges in many cases. For instance, when more than one threshold is reported by each primary study in a meta-analysis of accuracy studies or when a reference standard is imperfect, or a gold standard is missing (Chapter 2-4). We also observed that in addition to the limitations of risk-of-bias assessments in comparative accuracy studies, these studies commonly miss or incompletely report essential information about comparisons, which may mislead the interpretation of the findings (Chapter 5-7). Moreover, our findings showed that implementation of the medical diagnostic tests in clinical practice is a complex and challenging process, which can be promoted and hindered by different factors (Chapter 8). These resulting findings offer a starting point for improving current research practices in evaluation and implementation of biomarker-based tests