NAEA - National Abstinence Education Association

A Critique of the American Journal of Public Health,
January 2007 article:

“Explaining recent declines in adolescent pregnancy in the United States:
The contribution of abstinence
and contraceptive use”

Review commissioned by the
National Abstinence Education Association (NAEA)

Consultants: Lynne Tingle, Ph.D. and Lisa Rue, Ph.D.

Overview

Dr. John S. Santelli and colleagues from the Guttmacher Institute published an article in the January 2007 American Journal of Public Health suggesting that 86% of the decline in pregnancy risk is due to improved contraceptive use and only 14% is attributable to reduction in sexual activity, commonly known as abstinence. This study’s results challenge findings from an earlier study conducted by the same author, which attributed 53% of the decline in teen pregnancy to abstinence. (2004 Journal of Adolescent Health article Can changes in sexual behaviors among high school students explain the decline in teen pregnancy rates in the 1990s, v35).

Critique

The reviewers identified three important  points of discussion and two contradictions in the Santelli research. They are highlighted below:

Discussion Point #1: Different datasets are used

• The 2004 study results better reflect the strong impact of increased abstinence on the risk index because a more appropriate, school-based dataset was used for the first study.

Details to support #1:
The results of the two studies vary because the datasets used were different. The 2004 study used the YRBS dataset, whereas the 2007 study used the NSFG dataset. The YRBS survey is a school based study whereas the NSFB includes data from teens who may not be in school, some of whom are married. Both datasets are credible; however, because the authors’ intent was to draw conclusions about abstinence education, we do not believe the NSFG was the most appropriate. The YRBS is a school-based dataset more reflective of the target audience of abstinence education programs.

Discussion Point #2: Reductions in sexual activity are underestimated

• Both studies (2004, 2007) fail to account for factors important to abstinence education such as delayed initiation of intercourse. This oversight potentially underestimated the contribution of abstinence — which according to the authors is synonymous with abstinence education — to the risk index.

Details to support #2:
The biggest problem in methodology is the Contraceptive Risk Index (CRI). This index includes more than 20 variables to calculate risk associated with contraceptive use. Of concern is the failure to include reduced risk for delayed onset of sexual activity in adolescents — the authors acknowledge trends toward later initiation. Failure to account for this biased the final risk estimates because, for example, equal risks are attributed to sexually active 15 year olds and sexually active 19 year olds. Taking this example a step further, a 19 year old could have been exposed to abstinence education at an earlier age, and subsequently delayed her onset of sexual activity. According to the CRI, once she became sexually active at the age of 19, her risk score in the analysis would be based on whether contraceptives were used “at most recent sexual intercourse” and what type, completely disregarding the delayed onset. In the calculations of the CRI, this person is no longer contributing anything to the abstinence component (which in the authors’ conclusions is reflective of abstinence education), but rather to the contraceptive component.

Discussion Point #3: Conclusions are not adequately supported

• There are fundamental flaws with the conclusions drawn from the results. The authors’ conclusions focus primarily on abstinence education programs and policies; however, the data used in the analyses represent a variety of teenagers who may or may not be part of the abstinence education target population.

Detail Support for #3:
There are fundamental flaws with the conclusions drawn from the results. The authors’ conclusions focus primarily on abstinence education programs and policies; however, the data used in the analyses represent a variety of teenagers who may not be part of the abstinence education target population. For example, the NSFG includes: 1) teens who are not in school — abstinence education is predominantly implemented in schools, 2) teens from States that have minimal numbers of abstinence education programs, and 3) married teens. It is unknown how these errors biased the results, but it stands to reason that including those teens in the dataset would affect the percent attributable to reductions in sexual activity. In essence, while the dataset used, and the manner in which it was used, provides a picture of the nation’s teenage sexual behavior, it does not necessarily provide a picture of evaluation results on abstinence education.

Contradiction #1: Cost benefits of abstinence education are underestimated

• The authors own findings scientifically document a 23% decrease attributable to a reduction of sexual activity.  Given the disparity between abstinence education and contraceptive funding levels, this indicator is promising. The government is getting a good return on a small investment.

Detailed Support for contradiction #1:
The authors state that “Federal government funding for abstinence-only education in the United States has grown rapidly since 1998, despite lack of scientific evidence supporting these programs…” This statement contradicts the authors’ own findings, where the results demonstrate movement in the desired direction on outcomes related to adolescent sexual abstinence. For example, there was a 55.2% decrease in pregnancy risk among 15-17 year-olds with 23% of the decrease attributable to a reduction in sexual activity. Given that the CRI calculations do not include age of onset, and the dataset is not wholly reflective of adolescents who are known to receive abstinence education, these findings are promising.

It is inappropriate that the authors lead the reader to believe that the federal government funds “abstinence-only” education at the expense of other interventions. Annual spending is in the billions for secondary and tertiary interventions versus approximately $2.5 million for primary prevention programs which promote abstinence. Given this significant funding disparity, the government’s investment in primary prevention initiatives is achieving adequate results on outcomes of interest. 

Contradiction #2: Abstinence education does not appear to reduce contraceptive use among sexually active teens

• The authors include several editorial comments which are contradicted by the data reported. They imply that the promotion of one abstinence strategy, “virginity pledges,” reduces the likelihood of contraceptive use. However, these data are contradictory to this argument. The 2007 article demonstrates that the promotion of abstinence education has not decreased access to or use of contraceptive by adolescents. 

Detail to support for contradiction #2:
The authors imply that the promotion of abstinence strategies such as “virginity pledges” reduces the likelihood of contraceptive use; and that current US domestic policy stresses “abstinence-only sex education to the exclusion of accurate information on contraception.” If this is true, whereby the US is completely under abstinence education influence, and abstinence education reduces the likelihood of contraceptive use, then why are the numbers of contraceptive use increasing? In our view, the authors have demonstrated that the promotion of abstinence as a primary prevention strategy during the last 10 years has not caused a decrease of contraceptive use in sexually active youth.

Conclusion

Although sound bites are often all that is remembered from published articles, it is important that abstinence educators familiarize themselves with research that impacts their ability to successfully implement their programs in communities across America.

• 

March 14, 2007