The probability of conception after discontinuance of oral contraception: A note on “oral contraception, coital frequency, and the time required to conceive,” by Westoff, Bumpass, and Ryder

Westoff Bumpass and Ryder (1969) conclude that no difference exists in the time required to conceive between women who have used oral contraception (OC) and women who have used other contraceptive methods. Yet the data on which this conclusion is based suggest a different interpretation. A table summarizes the data. Essentially the table is an extract of a life table in which only the probability of survival without becoming pregnant to the end of each ordinal month is shown. If the inversion of determining the monthly probabilities of conception among the monprgnant survivors of the preceding month is performed an intriguing and suggestive pattern emerges. Because of the progressive selecting out of women with the highest fecundability a progressive secular reduction in the proportion becoming pregnant in each successive month is expected and this is the case for women who have discontinued other forms of contraception than OC. The peak at 6 months is presumably no more than an example of round number preference. With the OC users this pattern does not emerge. A marked bimodality is present strongly suggesting that the use of OC divides women into 2 population with distinct characteristics. The characteristics of these populations must be such as to explain the very much lower conception rate of the OC users in the first 2 months. The authors explain this difference as being due in part to a higher rate of accidental pregnancy among users of less efficient methods and partly to lack of sophistication among users of other methods in interpreting the question of how long it takes to become pregnant. Neither of these explanations accounts for the finding that pregnancy appears more likely in the 3rd month than the 2nd and more likely in the 4th month than the 3rd after discontinuance of OC. It is the OC users pattern which is abnormal and unexpected so that it seems more likely that some factor is acting to depress their fecundability than the results for other users are distorted. It is assumed that there are 2 populations of OC users 1 of which (about 80% of users) is unaffected by OC use. The other population experiences a carryover of the contraceptive effect of the pill either for 2 or for 3 months after discontinuance. A table shows the monthly probabilities of conception deduced from this model and compares them with those actually found. In every month (except the 6th) the fit is so close that no more than 1 pregnancy in the study group of 110 women separates the 2 figures. It is concluded that for 20% of OC users there is a delay of either 2 or 3 months before their capacity to conceive is restored. This is most likely due to the persistence for this time of anovulatory cycles.