in treatment (IOM, 1990). For drug abuse treatment to be most effective in reducing HIV transmission, it is essential that social policies encourage rather than discourage participation in treatment. In addition, consistent encouragement or even pressure to enter and remain in treatment from family members, friends, community leaders, or authorities can increase treatment engagement and contribute to successful treatment outcomes (IOM, 1990). The criminal justice system can also play an important role in getting drug users into treatment and outcomes by providing treatment as an alternative to incarceration, or as a condition of probation or parole (IOM, 1990).

THE COMMITTEE’S APPROACH TO EVALUATING THE STRENGTH OF THE EVIDENCE

A large number of systematic approaches to evaluating the quality and strength of scientific evidence are available. Many of these approaches, such as that used by the U.S. Preventive Services Task Force,2 rely on explicit criteria to assign a “grade” to the evidence. Others use a more qualitative approach. In evaluating the strength of the evidence on the effectiveness of HIV prevention strategies for IDUs, the Committee used a structured qualitative method based on an approach developed by the GRADE Working Group—a collaboration of researchers that aims to address problems with rating systems (GRADE Working Group, 2004).

The GRADE approach takes into account strength of study design, study quality, consistency of findings across studies, directness/relevance of outcome measures, and populations. In establishing causality, this approach grades randomized trials as strong, prospective cohort and case-control studies as moderate, and other observational studies and reports as modest (e.g., serial cross-sectional, ecologic) or weak (e.g., cross-sectional) (see Box 2.1 for a description of common research study designs). The approach downgrades studies for serious limitations on quality; important inconsistencies; sparse, indirect, or imprecise data; low follow-up rates; and a high probability of reporting bias. Studies are upgraded for strong evidence of an association, consistency of effect, demonstration of dose-response, and good analytic control of confounders. Combining these elements, the approach assigns evidence to one of four quality categories:

  • Strong: Further research is very unlikely to change confidence in the estimate of the effect.

2

The task force’s rating system for the strength of the evidence is available at: http://www.ahrq.gov/clinic/3rduspstf/ratings.htm.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement