Skip to main content


Relative paucity of high-quality or moderate-quality evidence to inform World Health Organization guidelines on antiretroviral therapy

Date and Location




Monday 23 September 2013 - 13:30 - 15:00


Presenting author and contact person

Presenting author

George W Rutherford

Contact person

Tara Horvath
Abstract text
BACKGROUND: Guidelines development hinges on identifying evidence for specific interventions. When the World Health Organization (WHO) updates its guidelines, the process requires narrowly constructed PICO questions, which are designed to update specific recommendations. Because randomized controlled trials (RCTs) are generally designed and conducted without reference to WHO’s specific concerns, it is often not possible to obtain high-quality evidence directly addressing WHO’s questions of interest. OBJECTIVES: To inform the 2013 update of WHO’s antiretroviral therapy (ART) guidelines, we were asked to conduct systematic reviews with 44 PICO questions specifying 306 outcomes. METHODS: We searched the Cochrane Central Register, EMBASE, PubMed, Web of Science, and WHO’s Global Index Medicus, as well as abstracts from key scientific conferences. We used standard Cochrane methods for the reviews. In addition to RCTs, observational studies with comparators were eligible for inclusion. We used GRADE methodology to classify evidence quality for each outcome. RESULTS: We identified 145 publications, 68 (49%) of which were reports of RCTs. Of the 306 outcomes for which we gathered evidence (including assessments at different time-points), 37 (12.1%) were addressed by high-quality RCT data, 50 (16.3%) by moderate-quality RCT data, 87 (28.4%) by low-quality RCT data and 72 (23.5%) by very low quality RCT data. Observational studies provided low and very low-quality evidence for 60 (19.6%) outcomes. Indirectness and imprecision were particular problems, which resulted in downgrading of evidence quality for 244 (79.7%) outcomes. CONCLUSIONS: There was relatively little high-quality or even moderate-quality evidence to inform the new ART guidelines. In the absence of large RCTs that specifically inform PICO questions, we suggest that analyzing unpublished cohort data from large registries (such as that of the International Epidemiological Databases to Evaluate AIDS) may be an important adjunct to reviewing data from published RCTs and individual observational studies.