OBJECTIVE: To examine the prevalence of withdrawal (coitus interruptus) use among a cohort of U.S. females aged 15–24 years, to describe characteristics of withdrawal users, and to evaluate whether withdrawal users exhibit a higher risk of unintended pregnancy.
METHODS: We analyzed the 2006–2008 National Survey of Family Growth, estimating with a Cox proportional hazards model the risk of an unintended pregnancy over and up to a 47-month retrospective period among females aged 15–24 years who used withdrawal relative to females who used only other methods of contraception. We also examined correlates of withdrawal use using a logit model.
RESULTS: During the study period, 31.0% of females in our cohort used withdrawal. Of withdrawal users, 21.4% experienced an unintended pregnancy compared with 13.2% of females who used only other contraceptive methods (adjusted hazard ratio 1.75, 95% confidence interval [CI] 1.23–2.49). Withdrawal users were also 7.5% more likely to have used emergency contraception (adjusted odds ratio [OR] 1.57, 95% CI 1.13–2.20). Married females were 14.8% less likely than single females to use withdrawal (adjusted OR 0.58, 95% CI 0.35–0.96).
CONCLUSION: Use of withdrawal as contraception is common and might place females at higher risk of unintended pregnancy. Health care providers should be aware that many patients may use withdrawal, should consider the need for emergency contraception among these females, and should encourage them to use more effective methods of contraception.
LEVEL OF EVIDENCE: II
Use of withdrawal is common among young females and may be a marker for those at high risk of unintended pregnancy.
Department of Obstetrics & Gynecology, Duke University Medical Center, Durham, North Carolina; the Department of Psychiatry, University of Colorado Denver, Denver, Colorado; and the Department of Obstetrics and Gynecology, University of Chicago, Chicago, Illinois.
Corresponding author: Annie M. Dude, MD, PhD, P.O. Box 3616, Durham, NC 27710; e-mail: firstname.lastname@example.org.
Presented at the 2009 Population Association of America Annual Meeting, April 29–May 1, 2009, Detroit, Michigan.
Financial Disclosure The authors did not report any potential conflicts of interest.