The ultimate test of a vaccine is whether it protects against disease. However, avian flu is so dangerous that it would be unethical to test this by administering an experimental vaccine and then exposing the subject to the virus. Instead, we have to be satisfied by giving the vaccine and then measuring the antibodies produced as a result. We know how much antibody seemed to be protective in an earlier outbreak of H5N1. So, that was the measure used in this vaccine study. But it is only an assumption.