Frank Shann argues that the basic expanded programme on immunisation (EPI) schedule (Bacillus Calmette–Guerin (BCG) at birth, diphtheria, pertusis, tetanus (DPT)– oral polio vaccine at 6, 10 and 14 weeks, measles at 9 months) may not be optimal under certain circumstances. He suggests that some of these vaccines may be associated with timing-related ‘non-specific effects’ on the immune system which influence mortality attributable to conditions other than the specific targets of the vaccines.1 He reviews evidence that BCG and measles vaccines may have non-specific beneficial effects, and on the other hand, DPT might, under certain unusual circumstances (high background infectious disease risks but low pertussis risk) appear to have detrimental effects on all cause mortality. This complex subject needs to be considered in its proper context. This article is part of a long and complex …