THE ROLES OF ABSTINENCE AND CONTRACEPTION IN DECLINING PREGNANCY RATES

Santelli et al.1 appear to have underestimated the contribution of current sexual activity to declining pregnancy rates among adolescents aged 15 to 17 years. The authors calculated a contraceptive risk index (the weighted average of estimated “typical use” failure rates for contraceptive methods reportedly used at last intercourse), assuming this index accurately reflects the risk of pregnancy among sexually active adolescents. But a review of pregnancy data for 1995 and 2002 reveals that the calculated index does not consistently correspond with actual pregnancy risk for sexually active adolescents aged 15 to 17 years.

Pregnancy risk among sexually active adolescents can be calculated more accurately by dividing the pregnancy rate by the proportion of adolescents who are sexually active. In 1995, the pregnancy rate for girls aged 15 to 17 years was 64.1 per 1000.2 Because 28.2% reported recent sexual activity in 1995, the pregnancy risk among sexually active girls was actually 0.227 (0.0641/0.282), not 0.344 as estimated by the authors. In 2002, the pregnancy rate among girls aged 15 to 17 years was 0.0423. Because 23.4% reported recent sexual activity in that year, the pregnancy risk among sexually active girls was 0.181 (0.0423/0.234), similar to the authors’ estimate of 0.186.

Repeating the authors’ calculations with the more accurate estimates of pregnancy risk among sexually active adolescents aged 15 to 17 years, we found that reduced pregnancy risk among sexually active adolescents accounts for 54.7% of the decline in pregnancy rates, whereas the reduced proportion of adolescents who are sexually active accounts for 45.3%. An alternative approach is to calculate the anticipated change in pregnancy rate were sexual activity to decline without any change in pregnancy risk for sexually active adolescents. Had this occurred, the 2002 pregnancy rate would have been 0.053 (0.234 × 0.227). By contrast, had there been no change in recent sexual activity, and pregnancy risk for sexually active adolescents had declined, the 2002 pregnancy rate would have been 0.051 (0.282 × 0.181). Using either approach, the contributions of declining sexual activity and reduced risk of pregnancy among sexually active adolescents are actually quite similar.

Additional research is needed to determine why standard contraceptive effectiveness rates may not apply to adolescents (differences in the frequency of intercourse and the consistency and correctness of use compared with adults are possibilities). Meanwhile, researchers should not assume that pregnancy risk among sexually active adolescents is only a function of contraceptive choice.