查看文章

Methods

This study is a sub‐analysis of data collected as part of a global systematic review of causes of hospitalization among HIV‐positive patients. The full methodology of the study is explained elsewhere [12]. Briefly, two investigators (NF, ZS) working independently and in duplicate searched Medline via PubMed, Embase, Web of Science, LILACS, AIM, IMEMR and WPIMR from inception up to 31 January 2015. The investigators also screened all International AIDS Society conferences up to Melbourne, 2014. Consensus was achieved to resolve discrepancies. Investigators followed a study protocol available from the corresponding author; the search strategy is provided in the Supplementary file. Studies were considered eligible if they included TB among the reported causes of hospitalization within a cohort of at least 20 HIV‐positive hospitalized patients. In order to focus on data following ART scale‐up, we …