Despite more than 30 years' experience with the pill, being by far the most thoroughly studied drug ever, we must conclude that there still is a remarkable paucity of data that would allow us to assess unambiguously its margins of efficacy. The physiological studies on which we must rely encompass limited numbers of subjects and are unlikely to include sufficient representatives of the vulnerable minority of women that really matter. Even though we realise that this vulnerable minority is there, we still cannot do better than to hypothesise about their characteristics, let alone identify them in advance. This lack of knowledge has contributed to the existence of diverging views on how to advise the general population of pill takers about missed tablets. Against this background, we felt there was a need to make an inventory of the existing data and, subsequently, to incorporate them in advice that in our opinion is most appropriate in the current state of knowledge. We have come to the conclusion, in contrast to what is often held, that it is not the number of tablets missed, but rather the timing relative to the pill-free interval that determines the impact of noncompliance. We further conclude that shortening of the pill-free interval to five or six days could substantially improve the efficacy of the pill: at the low doses currently used in oral contraceptives the total steroid burden would not be substantially increased, while still allowing withdrawal bleeding to occur.(ABSTRACT TRUNCATED AT 250 WORDS)