Approximately 250,000 valve replacement operations occur annually around the world and more than two thirds of these operations use mechanical heart valves (MHV). These valves are subject to complications such: pannus and/or thrombus formation. Another potential complication is a malfunction in one of the valve leaflets. Although the occurrence of such malfunctions is low, they are life-threatening events that require emergency surgery. It is, therefore, important to develop parameters that will allow an early non-invasive diagnosis of such valve malfunction. In the present study, we performed numerical simulations of the flow through a defective mechanical valve under several flow and malfunction severity conditions. Our results show that the flow upstream and downstream of the defective valve is highly influenced by malfunction severity and this resulted in a misleading improvement in the correlation between simulated Doppler echocardiographic and catheter transvalvular pressure gradients. In this study, we were also able to propose and test two potential non-invasive parameters, using Doppler echocardiography and phase contrast magnetic resonance imaging, for an early detection of mechanical heart valve malfunction. Finally, we showed that valve malfunction has a significant impact on platelet activation and therefore on thrombus formation.