Heart failure is a leading cause of death around the world and is often managed using implanted ventricular assist devices (VADs) to assist with blood circulation. Despite their wide use, VADs cause many complications that lead to dangerous outcomes such as stroke, gastrointestinal bleeding, and pulmonary embolism, which typically begin with thrombi, or pathological blood clots. VAD thrombosis is caused by one or a combination of three factors: alteration of blood flow characteristics, blood-material interactions, biochemical changes. In this work, I employ microfluidic devices to image shear-induced thrombus formation using fluorescence microscopy. In chapter 2, I detail the impacts of clinically relevant sub-micron and microscale surface roughness on shear-driven thrombus formation on titanium alloy surfaces. I visualize platelet adhesion and aggregation as a representation of thrombosis. The results show that over a range of sub-micron scale rough surfaces, at physiological shear rates, platelet adhesion is limited to relatively innocuous monolayers and small aggregates. Under pathological shear, there are large and unstable aggregates connected by fibrillar structures that are hundreds of microns in length. At the micro-scale, platelets deposit in crevices. More strikingly, platelet aggregates form even at physiological shear rates. Platelet aggregate height increases with surface roughness level between sub-micron and scale levels, whereas the number of aggregates nucleated is not correlated with surface roughness level or shear rate. In chapter 3, I focus on red blood cell (RBC) entrapment in thrombi at a pathological shear rate. I develop methodology to simultaneously image platelets and RBCs at high temporal resolution. I probe the effects of RBC deformability and surface antigen binding on entrapment location, thrombus (aggregate) growth and RBC accumulation. I demonstrate that RBC deformability enables their entrapment in thrombus core regions and the clustering of RBCs within thrombi. Surface antigens GPIIb/IIIa and integrin-associated protein inhibition negatively affect thrombus stability and sustained growth. Finally, I suggest future directions to further our understanding of shear-induced thrombus formation in implanted devices, including mechanistic studies to probe thrombosis at the micro-scale and the study of immunothrombosis, an under-explored phenomenon.