The prothrombin time (PT) assay is the most often requested coagulation test and used primarily for monitoring Vitamin K antagonist therapy, where results may be expressed as an International Normalised Ratio (INR). The INR is the patient's PT 'mathematically adjusted' or 'standardised' to take into account the specific test system used (i.e., comprising the test reagent/instrument combination). This standardisation or 'adjustment' is achieved by applying two 'correction factors', respectively defined by the 'International Sensitivity Index' (ISI) and the 'mean normal prothrombin time' (MNPT), according to the formula: INR=[patient PT/MNPT](ISI). While some manufacturers provide assigned ISI values for specific PT reagents and instrumentation, the vast number of possible reagent/instrument combinations usually precludes this in most situations. Even when an ISI is provided by the manufacturer, laboratories need to check or validate the assigned value. When a manufacturer does not provide an ISI, the laboratory needs to define its own (local ISI) value. The MNPT usually has to be locally defined, based on the population being tested. Current recommendations for defining ISI values include the classical, but prohibitively complex, World Health Organization (WHO) recommended procedure, and more recently the use of commercial reference-plasma calibration sets. The MNPT can also be defined using the WHO recommended procedure or with calibration sets. However, there is limited information to validate the performance of these in laboratory practice, and there are several unrecognised problems that limit the validity and utility of the ISI and MNPT values that are determined. Thus, it is perhaps time to start thinking outside the box, and to utilise additional methods for determining and/or validating ISI and MNPT values. This may include the use of regression analysis to assess ongoing peer-related performance in external quality assurance programmes, and to compare the behaviour of proposed replacement reagents with that of existing reagents. Such strategies have proven of considerable benefit to local laboratory practice, and should therefore enable other laboratories to optimise their practice in order to provide INRs that better reflect a patient's anticoagulant status, and thus assist in their clinical therapeutic management.