Our goal is to accurately predict all types of cardiovascular events in patients utilising nonclinical cardiovascular safety data. In the past two decades, cardiovascular safety science has primarily focused on events associated with the electrocardiogram. Broadening out to other cardiovascular parameters, we share real-life case studies that highlight our progress towards improved and better-informed project progression based upon use of disease models, mechanism-based translation and structure–function relationships. To fulfil this goal, further advances in patient-relevant humanised models will be required to enable cardiovascular safety science to keep pace with the ever-changing landscape of novel therapeutic paradigms.