- The multivariate logistic regression included all variables shown above: age, race, education, current smoking, current cocaine user, new partner, and use if each method of contraception versus no use of that method.
- All types of contraception were categorized on the basis of use in the past 4 weeks.

Our data from the PEACH study, and these have been published, suggest pretty much the same thing. Oral contraceptives certainly don’t reduce the risk of upper genital tract infection/endometritis, but they actually seem to be reducing the tenderness score, which is the amount of tenderness. But obviously this is not statistically significant. We did have the opportunity to look at barrier methods of contraception and other methods. We found that there was no method of contraception that actually reduced the risk of UGTI/endometritis, so that was the bad news.