SANTELLI ET AL. RESPOND

Abstract

We thank Mann and Stine for their interest in our recent article. We examined data on sexual activity and contraceptive use from the National Survey on Family Growth (NSFG) and found that most of the declines in US adolescent pregnancy from 1995 to 2002 could be attributed to improved contraceptive use: for 77% of the decline among those aged 15-17 years and for all of the decline among those aged 18-19 years. The methodological improvement in calculating attribution-in our Journal article and a previous article - was to calculate pregnancy risk by combining data on sexual activity contraceptive use and contraceptive failure rates. Mann and Stine suggest an alternative method for calculating attribution that uses data for a single variable: sexual intercourse in the past 3 months. We would note that the decline in sexual intercourse in the past 3 months for 15- to 17-year-olds (from 28.2% to 23.4% a decline of 17%) does not reach statistical significance (P=.065). By contrast the decline in pregnancy risk among sexually active adolescents (45.9%) was highly significant (P<.001). Mann and Stine describe a calculation of pregnancy risk that should more properly be called a pregnancy rate among sexually active teens. Their calculation is very different from our measure of pregnancy risk which is based on behavioral data. Moreover our overall pregnancy risk index calculates the risk of becoming pregnant not the risk of completing a pregnancy. Given these differences the absolute value of our pregnancy risk index will not equate with actual pregnancy rates; however changes over time in pregnancy risk and pregnancy rates should be similar as we described in our article. In addition Mann and Stine suggest that standard contraceptive failure rates may not apply to adolescents. This is not true. In fact typical-use contraceptive failure rates from the NSFG are calculated using data from both young women (aged 15-19 years) and adult women (aged 20-44 years). Method-specific failure rates from the NSFG are virtually identical for adolescents and adult women and did not change between 1988 and 1995. Calculations on the basis of typical-use contraceptive failure rates and contraceptive use at last intercourse inherently reflect correctness and consistency in use as well as coital frequency. Finally could other factors be influencing trends in pregnancy risk among sexually active adolescents? Certainly! In our article we discussed the possibility of changes in other factors driving fertility trends; however we are not aware of any evidence supporting such a change from this country or elsewhere. (full text)