This thesis combines five published research papers investigating methods of improving the safety and control of oral anticoagulation, with the use of computer assisted dosage and the standardisation of the International Normalised Ratio (INR). The INR is a conventional measurement derived from the time it takes blood of a patient to form a clot and is used to monitor the effects of widely used oral anticoagulants such as warfarin for the prevention of stroke and other related disorders.The first paper investigates whether the use of computer-assisted programs was as safe and effective as medical staff manual dosage in the prevention of bleeding or thrombotic complications during oral anticoagulant treatment. This was an international multi-centre randomised study conducted by the European Action on Anticoagulation (EAA) investigating the clinical benefit of two computer programs, PARMA 5 (Italy) and DAWN AC (UK). Composite clinical events were reduced by 7.6% using computer programs, though not achieving statistical significance (p=0.1), showing computer programs to be not dissimilar to medical staff dosage.The second paper recommends guidelines for screening safety and effectiveness of other marketed computer programs based on the results of the EAA study. A process for a candidate computer program to achieve non-inferiority relative to the medical staff dosage arm from the EAA study is explained.The third paper introduces a modified approach to the 'Direct INR' method for the standardisation of INR termed the 'Prothrombin Time/INR Line' (PT/INR). This was directly compared to the local International Sensitivity Index (ISI) calibration procedure originally approved by the World Health Organisation and later by the United States Food and Drug Administration (FDA). Using manually certified lyophilised plasmas tested by specialist centres, the PT/INR Line using a set of 5 calibrant plasmas to establish a fitted line to estimate local INR was shown to be as effective as the FDA procedure.The fourth paper investigates the PT/INR Line further by using simulated sets of calibrant plasmas across the therapeutic range of 2.0-4.5 INR and determining the PT/INR Line. Local INR of five validation plasmas, certified by 3 centres using the manual PT technique, was determined using the estimated PT/INR Lines and compared with local ISI calibration. Using 4 or 5 calibrant plasmas to determine the PT/INR Line was shown to be as accurate as local ISI calibrations for reliable local INR.The fifth and final paper assessed INR variability and control in oral anticoagulant therapy using a method termed the Variance Growth Rate (VGR), and compared its predictive ability of adverse events with the Time in Target INR range (TIR), the conventional method used in evaluating the quality of oral anticoagulant therapy. The VGR method was shown to be a better predictor of adverse bleeding or thrombotic episodes in the short term period prior to an event (3 and 6 months) compared with TIR.