Blood transfusion scientists have different definitions of ‘clinical significance’. For instance, is clinical significance judged by shortened RBC survival, laboratory evidence of haemolysis (e.g. increased bilirubin, LDH, etc.), and/or clinical signs (e.g. jaundice) of a haemolytic transfusion reaction (HTR)? Are one, two, or all three of these criteria needed to make an antibody significant? For instance, is a shortened RBC survival with no obvious laboratory abnormalities and/or or clinical signs really clinically significant? The answer may be different for a haematology patient requiring regular transfusions and a surgical patient with normal erythropoiesis. It is important to answer these questions in one’s own institution as they are the basis for our choice of compatibility tests and needed for interpretation of assays predicting clinical significance.Several approaches are valuable in predicting clinical significance: (1) specificity and thermal amplitude of the antibody; (2) 1 h survival of 51Cr‐labelled incompatible RBCs; (3) functional cellular assays. Only antibodies reactive at 37°C in vitro are of potential significance. The 51Cr survival and cellular assays [e.g. monocyte monolayer assay (MMA)] are usually used to help decide if incompatible RBCs can be transfused to patients with antibodies to high‐incidence antigens that are sometimes significant but often are not. When the MMA = 5·1–20%, 33% of patients showed no clinical reactions, but 67%, with no clinical signs showed laboratory signs of a reaction. If the MMA is >20%, most patients show clinical (64%) or laboratory (75%) reactions. It is of interest to note that regardless of MMA or 1 h 51Cr results, most patients with antibodies to high‐incidence antigens have abnormal full RBC survival studies (T50Cr), but only one‐third show any clinical signs of a reaction. Thus, when interpreting results for clinicians, it is important to clearly define what you mean by ‘clinical significance’ so that they can balance decisions of transfusing incompatible blood versus time and expense of obtaining rare units lacking the cognate antigen. Cellular assays are not available for many people, so historical data on whether certain specificities have caused HTRs are the only practical approach for many.