Article Text

Download PDFPDF

Effective Sexual Health Interventions: Issues In Experimental Evaluation
  1. J Green
  1. St Mary’s and Imperial College Hospital, London, UK;

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Ed Judith M Stephenson, John Imrie, and Chris Bonell. Pp 232; £55. Oxford: Oxford University Press, 2003. ISBN 0-19-850849-2.

    HIV spreads more every day and there are epidemics of other STIs in both the developed and developing world at least in part because the fear of HIV appears to be receding in the population. Our current strategies to contain these problems are meeting with limited success and treatment of people who are already infected, important though that is in controlling bacterial infections, is much less effective with continuing viral infections. There is an urgent need to develop and to test better methods of helping people to reduce their risky sexual behaviour.

    This book is excellent, brief, fairly comprehensive, and very readable. Its focus is designing studies on the effectiveness of sexual health interventions. If we are to get anywhere in improving behavioural interventions it is essential that what is done is carefully evaluated.

    The first three chapters of the book are concerned with methodology, particularly whether randomised controlled (RCTs) trials are an appropriate method for evaluating interventions in this area. While this section of the book is well argued on all sides it doesn’t really break any new ground. The strengths and weaknesses of RCTs in behaviour change are pretty much what they are in any other area of medicine. Methodologies don’t exist as stand alone phenomena, whether an RCT or some other methodology is appropriate depends simply on what question one is seeking to answer.

    The second section of the book covers models of behaviour change and the choice of design and outcome measures. It is clear that one of the main problems in intervening in sexual health is the poor quality of the available psychological models and our real lack of understanding about why people behave as they do. Without understanding why people behave as they do it is difficult to help them to change. It is interesting that models of health behaviour never seem to get discarded, even the ones that are known to be weak. There are particularly strong chapters on cluster randomisation, an approach which probably gives rise to more inappropriate statistics than any other and on complex behavioural measures. The latter should be required reading for anyone measuring any aspect of risky sexual behaviour simply because it highlights how weak many studies of sexual behaviour—and not just of behaviour change—are in this respect.

    The book ends by looking at generalisability in its broadest sense. Generalisability is an area that tends to get overlooked. Even a highly successful behaviour change programme would be of no use in developing countries if it was labour intensive and dependent on highly skilled staff for its delivery.

    I would recommend this book to anyone planning a trial or simply seeking to understand the existing literature. I would however caution that to make sense of it you will have to look at some of the available reviews of the behaviour change literature since the book assumes some knowledge, or willingness to acquire knowledge, of these.