Review: sexual abstinence only programmes do not affect STIs or HIV risk behaviours in high-income countries
Correspondence to: Ms K Underhill, University of Oxford, Oxford, UK;
Do sexual abstinence (SA) only programmes prevent HIV infection in high-income countries?
Medline, CINAHL, EMBASE/Excerpta Medica, CENTRAL, Catalogue of US Government Publications, and 25 other databases (1980 to February 2007); libraries of agencies involved with HIV prevention (eg, WHO); relevant conference proceedings (after 2000); experts; and cross-referencing articles on pregnancy prevention and HIV prevention.
Study selection and assessment:
randomised controlled trials (RCTs) or quasi-RCTs in any language that evaluated any intervention or programme for SA as the only means of HIV prevention in high-income countries (ie, gross national income/capita ⩾$10 726, £5450, or €8035). Trials of pregnancy and HIV prevention, or HIV prevention alone were included. Trials of people who were HIV positive, programmes explicitly promoting condom use or safe sex, SA only programmes not evaluating HIV prevention, and RCTs that did not report biological or behavioural outcomes were excluded. Quality assessment of individual studies was based on Cochrane criteria (eg, allocation concealment, blinding, and follow-up). 13 RCTs (n = 15 940, mean age range 10–19 y) met the selection criteria; control groups included usual care, no treatment, non-enhanced SA programmes, time-matched safe sex programmes, and time-matched programmes that promoted both safe sex and abstinence.
self-reported biological outcomes (eg, HIV incidence, sexually transmitted infections [STIs], or pregnancy) and behavioural outcomes (eg, incidence or frequency of any protected or unprotected vaginal sex, number of sex partners, and sexual initiation).
Meta-analysis of studies was not done because of a lack of intention-to-treat analyses and heterogeneity of trial and intervention designs. No RCTs evaluated HIV incidence. In 6 of 7 RCTs (n = 9779), groups did not differ for incidence of STIs; in 1 RCT, the SA group had more STIs than the usual care group at 3 and 17 months (table). In 7 of 8 RCTs (n = 9417), groups did not differ for self-reported pregnancy; in 1 RCT, the SA group had more pregnancies than the usual care group at 17 months (table). In 6 of 7 RCTs (n = 3454), groups did not differ for incidence of any vaginal sex; in 1 RCT, the SA group had a lower incidence of vaginal sex than the usual care group at 2 months (table). In 3 of 4 RCTs (n = 2376), groups did not differ for frequency of vaginal sex; in 1 RCT (n = 338), the SA group had a higher frequency of vaginal sex than the usual care group at 3 months (mean increase from baseline 0.9 v 0.3, p = 0.02). Groups did not differ for incidence of unprotected vaginal sex (5 RCTs, n = 2892), number of sex partners (8 RCTs, n = 4483), or sexual initiation (10 RCTs, n = 11 298).
Sexual abstinence only programmes do not affect self-reported sexually transmitted infections or HIV risk behaviours in high-income countries.
Underhill K, Montgomery P, Operario D. Sexual abstinence only programmes to prevent HIV infection in high income countries: systematic review. BMJ 2007;335:248.
Clinical impact ratings: Health promotion 6/7; Paediatrics 5/7; Public/Community health 6/7; Sexual health 5/7
The review by Underhill et al used Cochrane methodology and included 13 RCTs that evaluated whether SA only programmes prevent HIV infection in high-income countries. Meta-analysis was not done because of the diversity of study designs. However, all but 1 of the studies in the review showed that SA only programmes did not differ from control for prevalence of sexual risk behaviours. In 1 study, the SA only programme that was associated with decreased vaginal sex did not include information on the prevalence of condom use. These findings highlight the danger of using SA only programmes because they do not provide tools for adolescents to protect themselves against STIs when they become sexually active.
Adolescents’ accounts of their first intercourse experiences show that this is often not a cognitive choice but arises out of gender roles, hormonal and sexual desires, and socioeconomic, socialisation, and power issues.1 2 When SA is a component of comprehensive sex education programmes, adolescents are more likely to abstain from sexual intercourse longer and to use safer sex when they begin to engage in sexual intercourse.3 4 Although all of the studies in the review were from the US, this review is relevant to any region where funding for HIV prevention programmes is limited to SA only programmes, leaving adolescents in countries with a high prevalence of HIV at increased risk of unsafe sexual behaviours.