Secondary Logo

Journal Logo

Reporting deficiencies in trials of abstinence-only programmes for HIV prevention

Underhill, Kristen; Operario, Don; Montgomery, Paul

doi: 10.1097/QAD.0b013e32801199fe
Correspondence
Free
SDC

Centre for Evidence-Based Intervention, University of Oxford, Oxford, UK.

Received 24 August, 2006

Accepted 5 October, 2006

Abstinence-only programmes for HIV prevention have long inspired debate on the grounds of public health, politics, and morality [1–9]. Amid this discord, evidence from experimental trials is frequently neglected. A forthcoming Cochrane review of the clinical effectiveness of abstinence-only interventions in high-income settings [10] has identified implications for policy and practice, but reporting shortcomings may limit the application of trial evidence.

On the premise that sexual abstinence is the only certain way to prevent HIV transmission, abstinence-only programmes encourage participants to refrain from sexual activity. Unlike comprehensive or ‘abstinence-plus’ programmes, abstinence-only interventions do not promote safe-sex strategies such as condom use. This approach has been widely criticized, especially because ‘abstinence-until-marriage’ interventions receive 33% of the prevention funding in the United States President's Emergency Plan for AIDS Relief [11].

Recent meetings of the International AIDS Conference have highlighted not only the political controversy surrounding abstinence-only programmes, but also the evidence for their effectiveness. The International AIDS Conferences in 2004 and 2006 featured a systematic review of abstinence-only programmes for HIV prevention in developing countries [12,13], which found little to no evidence of behavioural effects.

This year's conference also featured a Cochrane systematic review of abstinence-only programmes for HIV prevention in high-income countries [10,14,15]. Given the severe need for evidence in this debate, our review exposes recurring reporting deficiencies that may limit the extent to which programme trials can make specific recommendations for policy and practice. Despite the use of randomized controlled designs, no abstinence-only programme trial has specified all information recommended in the complied CONSORT reporting guidelines [16], and missing data hindered meta-analysis. Clinical trial evidence has striking implications, but the valuable lessons of abstinence-only programme evaluations may be hampered by incomplete reporting.

With a view towards making evidence more applicable to practice, the following deficits warrant attention:

Specifying procedures for random assignment: Despite searching 30 databases, 16 765 abstracts, and unpublished literature, the review encompasses only 13 191 participants from randomized controlled trials (a small figure for a widespread public health intervention). No trial described procedures for randomly assigning participants or concealing the allocation sequence; one report failed to state that random assignment was used. These details are critical for appraising trials, and deficient reporting may obscure methodological strengths.

Reporting clinically meaningful outcomes: Biological and behavioural outcomes are the best indicators of HIV risk. The majority of abstinence-only evaluations avoided these outcomes or failed to define them specifically. No study measured HIV incidence, three assessed self-reports of sexually transmitted disease diagnosis, and only one assessed the incidence of unprotected sex. It was often unclear whether oral and anal sex acts were included in measures of ‘sex’, ‘partners’, or ‘condom use’, although these acts carry unique HIV risks and should be measured separately. Terms such as ‘intercourse’ and ‘virginity’ were imprecisely defined, despite various possible understandings [17,18]. Resistance to measuring behavioural and biological outcomes in youth is common, but using and unambiguously defining these measures is essential to evaluate the risk of HIV.

Conducting appropriate analyses: No study conducted intention-to-treat analyses, which minimize attrition bias by accounting for dropouts. Several studies failed to accommodate clustered randomization statistically, and only three reported sufficient data for meta-analysis on all measured outcomes. Inappropriate analyses and selective reporting threaten the internal and external validity of trial results.

Providing implementation and fidelity details: Because abstinence-only programmes are complex interventions, our review also attempted to investigate how different programme components affected behaviour across studies. Programmes may have used similar strategies to promote abstinence, but incomplete reporting made it difficult to ascertain exactly what programme activities, dosages, settings, facilitators, and contexts were measured. Without rigorous fidelity monitoring, few studies could reliably indicate whether interventions were actually delivered as planned.

Including comparison details: A major source of uncertainty in the abstinence-only policy debate is how abstinence-only interventions compare with programmes that emphasize abstinence, but also promote safe sex. To address this question, studies comparing abstinence-only against abstinence-plus programmes are most useful. Only one included trial made this comparison. Several other trials compared abstinence-only programmes against ‘usual care’ control groups, which may have received abstinence-plus or condom-promotion programmes. However, unclear reporting made this impossible to discern, limiting the relevance of trial evidence.

In the interests of making the debate over abstinence-only interventions more evidence-based, our Cochrane review has identified trial characteristics that merit fuller reporting in primary studies. Methodological critiques of clinical trials are not new, but debates over abstinence-only policies often prioritize political or moral claims over rigorous evidence. Our findings suggest that abstinence-only programme evaluations could make valuable and specific contributions to these debates, but must report key data more completely.

Back to Top | Article Outline

Acknowledgements

The authors are grateful to the University of Oxford and the Centre for Evidence-Based Intervention for internal support.

Back to Top | Article Outline

References

1. DiClemente RJ. Preventing sexually transmitted infections among adolescents: a clash of ideology and science. JAMA 1998; 279:1574–1575.
2. Editorial. Abstinence, monogamy, and sex. Lancet 2002; 360:97.
3. Human Rights Watch. Ignorance only: HIV/AIDS, human rights, and abstinence-only programs in the United States. New York, NY: Human Rights Watch; 2002.
4. Sinding S, Green E. Debate: CNN vs ABC (CNN = condoms, needles and negotiating skills/ABC = abstinence, be faithful, condoms). In: XVth International AIDS Conference. Bangkok, Thailand, 11–16 July 2004.
5. Smith W. Abstinence-only as prevention: science, ideology and details from the US experience. In: XVth International AIDS Conference. Bangkok, Thailand, 11–16 July 2004 [Abstract MoPeE3995].
6. Walgate R. Bush's AIDS plan criticised for emphasising abstinence and forbidding condoms. BMJ 2004; 329:192.
7. Editorial. Is it churlish to criticise Bush over his spending on AIDS?Lancet 2004; 364:303–304.
8. Garrett L, Osotimehin B, moderators. ABC in Africa – what is the evidence? In: XVIth International AIDS Conference. Toronto, Canada, 13–18 August 2006.
9. Editorial. HIV prevention policy needs an urgent cure. Lancet 2006; 367:1213.
10. Underhill K, Montgomery P, Operario D. Abstinence-based programs for HIV infection prevention in high-income countries. Cochrane Database Syst Rev 2005; 3:CD005421.
11. United States Leadership Against HIV/AIDS, Tuberculosis, and Malaria Act of 2003. House International Relations Committee, Washington, DC, USA. PL 108-25 (S.250, HR 1298). 2 April 2003.
12. O'Reilly K, Medley A, Dennison J, Schmid GP, Sweat MD. Systematic review of the impact of abstinence-only programmes on risk behavior in developing countries. In: XVth International AIDS Conference. Bangkok, Thailand, 11–16 July 2004 [Abstract TuPeC4899].
13. O'Reilly K, Medley A, Dennison J, Sweat MD. Systematic review of the impact of abstinence-only programmes on risk behavior in developing countries (1990–2005). In: XVIth International AIDS Conference. Toronto, Canada, 13–18 August 2006 [Abstract THAX0301].
14. Underhill K, Montgomery P, Operario D. Systematic review of abstinence-only and abstinence-plus programs for the prevention of HIV infection in high-income countries. In: XVIth International AIDS Conference. Toronto, Canada, 13–18 August 2006 [Abstract THPE0382].
15. Underhill K, Montgomery P, Operario D. Systematic review of abstinence-only programs for the prevention of HIV infection in high-income countries. In: 1st International Workshop on HIV Transmission. Toronto, Canada, 11–12 August 2006.
16. Moher D, Schulz KF, Altman D, for the CONSORT Group. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA 2001; 285:1987–1991.
17. Sanders S, Reinisch J. Would you say you “had sex” if? JAMA 1999; 281:275–277.
18. Barnett T, Parkhurst J. HIV/AIDS: sex, abstinence, and behaviour change. Lancet Infect Dis 2005; 5:590–593.
© 2007 Lippincott Williams & Wilkins, Inc.