The viral safety of the blood supply provided by serological tests alone decreased the residual risk of viral transmission to less than 1:250,000 for hepatitis C virus (HCV) and 1:1.3 M for HIV in the EU and the USA in 2000. This was further improved to 1:2-4 M by the introduction of nucleic acid testing (NAT) for HCV and HIV RNA that considerably reduced the risk of window period transmission. However, over the past 20 years, the successive introduction of up to 10 direct or surrogate viral markers enormously complicated the screening process and testing errors have become the main residual risk of viral transmission by transfusion. At over $ 2 M per QALY, the very low cost-effectiveness of NAT and some other tests overburdens limited funds that might be better used for other health care priorities. At the same time, haemovigilance programmes have shown that blood transfused to the wrong patient and a range of immunological consequences of transfusion caused two deaths per million transfusions and little is done to prevent them. There are means of limiting these serious hazards of transfusion that should become the priority in blood safety.