AIM: To assess the reliability of local international normalised ratios (INR) using a set of three international reference preparation (IRP) certified freeze dried plasmas. METHODS: 55 centres in the United Kingdom and the Republic of Ireland participated; 36 centres employed coagulometers and 19 a manual prothrombin time technique, all with the same batch of routine commercial thromboplastin. The plasmas had certified INR with the manual technique using a thromboplastin IRP, and results were provided graphically to participants for self assessment. An INR deviation of more than +/- 10% from the certified INR with any of the screening plasmas was regarded as unsatisfactory and clinically significant. Sets of 20 freeze dried plasmas were provided for local ISI calibrations and sets of seven freeze dried normals were provided for supplementary exercises where screening results were unsatisfactory. RESULTS: 15 of 38 coagulometers, but only three of the 19 manual prothrombin time test techniques, gave unsatisfactory results. With 10 of the 15 unsatisfactory coagulometer results the problem was resolved by local ISI calibrations with plasma calibrant sets provided. Unsatisfactory results with manual technique in all four instances were corrected by substitution of the mean result with freeze dried normal plasmas provided. CONCLUSIONS: The freeze dried plasma screening set was useful in detecting incorrect INR in a high proportion of coagulometer users and a smaller number of participants using the manual technique.