Abstract
Since 2003, >200 systematic reviews (SRs) have been published in the Journal of Antimicrobial Chemotherapy (JAC). Many have been widely cited. Guidelines for proper execution were outlined in 2005. Since then, new threats, challenges and methods have emerged. Data provenance is an emerging threat. There are several assumptions and limitations of both SR methods and the primary studies they include. The pivotal impact of these assumptions is illustrated using 13 SRs of pneumonia prevention in ICU patients on inferences towards the aspirational goal of 'Pneumonia zero' as a case study. Depending on these assumptions, the SRs of antimicrobial-based interventions pivot between two not incompatible inferences of pneumonia prevention for individual ICU patients versus harm for ICU populations. The case study highlights how newer methods for data visualization enhance reader insight into the underlying data beyond the summary effect size. To remain relevant to evidence-based medicine, SRs will need to adapt to emerging challenges and to recognize and validate the key underlying assumptions. This is especially so for antimicrobial-based infection prevention interventions given their potential for spillover effects.