A method is proposed to determine the optimal achieved intensity of oral anticoagulant therapy by calculating INR-specific incidence rates of thromboembolic and hemorrhagic events. This method uses person-time at different INR levels to assess the risk of adverse events at various intensities. The approach involves categorizing observation time into cells based on INR values and calculating incidence rates by dividing the number of events by the total person-time at each intensity. This method allows for the assessment of the effects of covariates through multivariate regression analysis and can be applied in clinical trials or routine patient care. The method helps identify the optimal anticoagulation intensity that balances thrombosis prevention and bleeding risk. It is particularly useful in determining target levels for future trials and analyzing risk factors for bleeding. The method assumes a linear change in INR between measurements, which may introduce some bias at extreme values but is considered minor since the optimal level is unlikely to be at these extremes. The method can be applied retrospectively using data from thrombosis services and is supported by data from studies on patients with mechanical heart valves. The results suggest a U-shaped relationship between anticoagulation intensity and stroke risk, with the lowest risk at an optimal intensity. This method provides a rational basis for setting target levels in future trials and helps evaluate the impact of various factors on bleeding risk.A method is proposed to determine the optimal achieved intensity of oral anticoagulant therapy by calculating INR-specific incidence rates of thromboembolic and hemorrhagic events. This method uses person-time at different INR levels to assess the risk of adverse events at various intensities. The approach involves categorizing observation time into cells based on INR values and calculating incidence rates by dividing the number of events by the total person-time at each intensity. This method allows for the assessment of the effects of covariates through multivariate regression analysis and can be applied in clinical trials or routine patient care. The method helps identify the optimal anticoagulation intensity that balances thrombosis prevention and bleeding risk. It is particularly useful in determining target levels for future trials and analyzing risk factors for bleeding. The method assumes a linear change in INR between measurements, which may introduce some bias at extreme values but is considered minor since the optimal level is unlikely to be at these extremes. The method can be applied retrospectively using data from thrombosis services and is supported by data from studies on patients with mechanical heart valves. The results suggest a U-shaped relationship between anticoagulation intensity and stroke risk, with the lowest risk at an optimal intensity. This method provides a rational basis for setting target levels in future trials and helps evaluate the impact of various factors on bleeding risk.