Standards for reporting non‐randomized evaluations of behavioral and public health interventions: the TREND statement

On 24–25 July 2003, the US Centers for Disease Control (CDC) convened a meeting of editors of journals that publish papers on HIV behavioral interventions to discuss ways to improve reporting of these studies. The meeting was attended by editors and representatives of 18 journals, including Addiction (for a complete list of journals go to http://www.trendstatement.org). The meeting stemmed from work carried out by the CDC's HIV/AIDS Prevention Research Synthesis (PRS) team, which has been synthesizing evidence from HIV behavioral intervention studies involving randomized and non-randomized clinical trials. During this work, the PRS team realized that many published and unpublished studies failed to provide information such as intervention timing, dosage and effect size data, which are critical for research synthesis. The meeting was also influenced by the movement for evidence-based medicine [see, for example, the Agency for Healthcare Research and Quality (accessed 9 April 2004), http://www.ahrq.gov; New York Academy of Medicine (accessed 9 April 2004), http://www.ebmny.org/thecentr2.html], the initiative on evidence-based public health practice (Victora et al. 2004) and the work represented by the Consolidated Standards of Reporting Trials (CONSORT) (http://www.consort-statement.org). The latter has become a standard for the reporting of randomized clinical trials, having been adopted by a large number of journals in the medical, clinical and psychology fields. It provides a list of 22 items and a subject flow chart for reporting the methodology of randomized trials. More specifically, the main goals of the meeting were: (a) communicate the usefulness and importance of adequate reporting standards; (b) reach consensus on reporting standards for behavioral interventions; (c) develop a checklist of reporting standards to guide authors and journal reviewers; and (d) develop strategies to disseminate the resulting reporting standards. During the meeting there was rapid uniform agreement on the importance of the standards. Discussion also led to a broadening of the focus for the standards to cover public health interventions in general, and not only HIV behavioral interventions. This was particularly so in the case of research using designs other than randomized allocation. There was also a consensus that any standard, to be successful, would have to be adopted by a large number of journals as well as embraced by researchers in the field, which could only be achieved by broad participation in the continuous development of the standards. The initial checklist constituting the standards, presently known as the Transparent Reporting of Evaluations with Non-randomized Designs (TREND), was recently published in an editorial in the American Journal of Public Health (Des Jarlais et al. 2004). The list is in the Appendix to this editorial and can also be found at http://www.trendstatement.org. Items in bold in the appendix are especially important to behavioral and public health intervention studies. The standards are intended for studies that necessarily include a defined intervention whose efficacy and effectiveness is being evaluated. They emphasize description of the intervention and its theoretical base, description of the comparison condition, full reporting of outcomes and description of design characteristics needed to assess possible biases in the outcome data. The Addiction editorial team agrees with the intent and content of the TREND statement. Also, in agreement with the original intentions of the statement, we do not see the list as a set of criteria by which papers on non-randomized intervention trials should be reviewed for evaluation. Rather, we see it as a working document that, if improved and followed, will increase comparability of reporting across studies, will assist in creating a cumulative body of knowledge and will help as a reminder of the types of information that should be reported. Various challenges lie ahead until the statement becomes widely accepted. As seen by all in the meeting and by the Addiction editorial team, the statement is a set of guidelines, which will not apply equally well to all types of non-randomized studies. Strict adherence to its suggestions will likely create strong pressure on the restrictive page limitations imposed by many journals. While Addiction has a flexible approach to such limitations, they are a reality in the day-to-day publishing of scientific papers. Potential solutions to this problem include the posting of additional information specified in the TREND statement on the Addiction website, on authors' websites or sending it to a central repository. The main purpose of this editorial is to call the attention of Addiction's readership to the TREND statement. The TREND group invites all of those interested in non-randomized intervention studies and in the HIV/AIDS and addiction fields to send comments on the standards to [email protected]. The TREND group will take all comments into consideration and periodically revise the checklist accordingly. The editorial team of Addiction will be participating actively in this process, and we have agreed to reference the statement in our publication guidelines for authors. We urge our readers and authors to familiarize themselves with the standards, and to participate in its refinement. We see these standards as a positive development in the field that should be supported.