Jonathan Elford

The TREND statement: Social science, communication and HIV/AIDS

The editors of AIDS Care support the TREND statement as a means of improving the reporting of nonrandomized evaluations of behavioural and public health interventions (Des Jarlais et al. , 2004; http://www.TREND-statement.org). The TREND statement provides an evolving checklist of critical descriptors for assessing such evaluative studies. We support it for two reasons. First, because it offers clarity and transparency for the reporting of interventions with nonrandomized methodologies. Second, because it raises important issues about scientific communication in the third decade of the HIV pandemic and the first decade of the 21st century. The philosophy of science has traditionally regarded non-significant (i.e. negative) findings as weaker than statistically significant outcomes, partly because negative evidence is more difficult to interpret. It is difficult to determine if non-significant outcomes are a function of methodological weaknesses or a genuine lack of difference. However, the increasing importance of meta-analyses and the biases related to non-publication of ‘negative’ results require a re-assessment of the reporting of all well-conducted studies without privileging those with significant outcomes. The TREND statement emphasizes the importance of fully reporting data in intervention studies while noting the costs of excluding some designs or negative results. The elements in the TREND checklist will enable the scientific community to more comprehensively assess the strengths and weaknesses of published interventions. In addition, they will serve to emphasize the fact that studies rarely stand on their own, but constitute part of a larger fabric of scientific discourse in which individual studies are the datum. With such an emphasis, full reporting of negative data is potentially as important as full reporting of positive data. The growing importance of interventions in behavioural science and public health has drawn attention to the fact that process issues are as important as intervention outcomes. Elford and Hart (2003) have noted that without an appreciation of how outcomes occur, we can make little progress in understanding and improving subsequent interventions. The transparent reporting outlined in the TREND statement offers a welcome opportunity to assess the processes that lead to the success or otherwise of interventions. This in turn will allow a fully informed debate as to the importance of negative or positive results from a methodological, as well as a philosophic standpoint. In AIDS Care , we have always stressed the need for triangulation of knowledge and the equal importance of qualitative and quantitative, experimental, quasi-experimental, and naturalistic approaches to understanding HIV/AIDS. While some research designs may be more powerful in allowing causal interpretation than others, our support of the TREND statement does not mean that we favour a methodological hierarchy. On the contrary we strongly believe in the value of all approaches of inquiry. We welcome carefully conducted and AIDS CARE (August 2004), VOL. 16, NO. 6, pp. 667 /668