Cardiovascular disease remains one of the world’s leading causes of death. Myocardial infarction (heart attack) is triggered by occlusion of coronary arteries by platelet-rich thrombi (clots). The development of new anti-platelet drugs to prevent myocardial infarction continues to be an active area of research and is dependent on accurately modelling the process of clot formation. Occlusive thrombi can be generated in vivo in a range of species, but these models are limited by variability and lack of relevance to human disease. Although in vitro models using human blood can overcome species- specific differences and improve translatability, many models do not generate occlusive thrombi. In those models that do achieve occlusion, time to occlusion is difficult to measure in an unbiased and objective manner. This thesis describes the development of a novel microfluidic device that reliably produces occlusive thrombi in vitro. The microfluidic device is a custom-designed PDMS-based chip, that triggers thrombosis with a collagen and tissue factor spot. These two substrates are exposed in vivo when an atherosclerotic plaque ruptures, and thus represent appropriate biological stimuli to trigger occlusive clot formation within an in vitro model. To allow the ‘time to occlude’ of the chip to be measured, I developed a simple and robust approach using a balance. This approach allows quantitative data to be collected regarding the efficacy of compounds in preventing occlusive clot formation, and subsequent statistical analysis to assess significance. Early stages of the project highlighted the potential for occlusion to occur in thrombosis microfluidic devices through off-site coagulation, obscuring the effect of anti-platelet drugs. I therefore redesigned the device in order to incorporate a stream of high-concentration ethylenediaminetetraacetic acid (EDTA) to quench coagulation downstream from the collagen and tissue factor patch. This EDTA solution was mixed into the blood by an on-chip chaotic mixer. To validate the device, I tested the approved anti-platelet drug, eptifibatide in both quenched and unquenched devices. In quenched devices, I measured a significant difference in the ‘time to occlude’ in treated devices compared to control conditions. These results were not replicated in unquenched devices, despite significant differences in the levels of platelet aggregation on the collagen and tissue factor patch. These results demonstrated that in unquenched devices, ‘off-site’ activity can mask the efficacy of antiplatelet compounds, but these erroneous effects were removed by the addition of downstream EDTA-solution. I then showed that the EDTA-quenched design is sensitive to differences in concentration of eptifibatide, further supporting it as an effective tool for drug testing. With the design of the device finalised, I tested a number of anti-thrombotic medications. Dual antiplatelet therapy composed of a P2Y12 inhibitor plus aspirin is currently the most commonly prescribed treatment for people at risk from adverse cardiovascular events. To assess this approach, I tested cangrelor and aspirin using my device. I found the treatment to be effective when collagen alone was used as a trigger for thrombosis, but when tissue factor was also used, as would occur in vivo, treatment with cangrelor and aspirin was ineffective. I tested the PAR inhibitors vorapaxar and BMS 986120, and found that neither PAR inhibitor on their own or in combination with one another effectively prevented thrombosis when triggered by both collagen and tissue factor. However, treating blood with a combination of cangrelor, aspirin, vorapaxar and BMS 986120 effectively prevented occlusion in my device. Finally, I showed that aspirin was not necessary for this to be the case: treating blood with cangrelor, vorapaxar and BMS 986120 effectively prevented occlusion in all donors tested. These results demonstrate that the device can be used to monitor the effect of antithrombotic drugs on time to occlude in vitro, and delivers this essential data in an unbiased and objective manner. The data gained concerning simultaneous inhibition of multiple platelet receptors sheds light on the interaction and redundancy between these receptors, and can be used to inform subsequent drug development initiatives. The relative simplicity of set-up and low cost of the developed system makes replication by other labs eminently achievable, and thus offers a strong alternative to the murine carotid artery injury model commonly used by the field.