Further analysis of contraceptive failure of the ovulation method
Trussell, J.; Grummer-Strawn, L.
American Journal of Obstetrics and Gynecology 165(6 Part 2): 2054-2059
1991
ISSN/ISBN: 0002-9378 PMID: 1755470 DOI: 10.1016/s0002-9378(11)90581-xDocument Number: 316791
Re-analysis of data collected from clinical trials of the ovulation method (OM) of natural family planning (NFP) delineated failure rates for both perfect and imperfect use during the teaching phase, the effectiveness phase, and the actual 1st year of use, and offered computed failure rates for various types of use errors. The data, originally published in 1981, covered 765 women in the teaching phase (2701 cycles), and 725 in the effectiveness phase (7514 cycles), from New Zealand, India, Ireland, Philippines and El Salvador. In this study rules for the teaching phase differed significantly from the effectiveness phase, with total abstinence required for the 1st cycle, and more restrictions in the teaching phase generally. The data suggested that errors were underreported, and that failure rates were higher than those typically reported in other studies. Except for the 1st cycle, cycles with no intercourse were eliminated from analysis. 2.0% conceived in the 1st cycle, as 46% had coitus. Life table probabilities of pregnancy were 22.0% in the 1st year, 3.2% in perfect use and 84.1% in imperfect use. Failure rates per cycle in the teaching phase (cycles 2-6), 0.22% failure with perfect use, and 9.9% with imperfect use, were similar to those in the effectiveness phase, 0.24% for perfect use and 14.4% with imperfect use. The most risky behaviors were having intercourse on wet/mucus days, within 3 days after the peak, during stress, on the peak day or around a 2nd peak (90% failure rate). Moderately risky behaviors were intercourse during mischarted cycles, spotting/bleeding days, having genital contact on wet days (50%). Intercourse on consecutive early days was least risky (16.6%). By plotting the percent failing in the 1st year as a function of per cycle failure rates for perfect and imperfect use, it was concluded that the OM is fairly effective if used perfectly, extremely unforgiving of imperfect use, and moderately difficult to use perfectly. If underreporting is taken into account, the conclusion would be that OM is moderately effective if used perfectly, and very difficult to use perfectly.