Risk-of-bias assessment of vaccine effectiveness studies: a scoping review of systematic reviews.

Abstract

Observational vaccine effectiveness (VE) studies provide essential real-world evidence but are prone to bias. Valid synthesis relies on rigorous risk-of-bias (RoB) assessment in systematic reviews of VE studies. Following JBI guidance, we mapped and described RoB assessment methodologies in systematic reviews of VE studies. We searched MEDLINE, Embase, and Web of Science from 1 January 2013 to 17 May 2023 and the grey literature from 1 January 2018 to 15 August 2023. Of 367 identified reviews, 38 lacked any RoB assessment, yielding 203 systematic reviews. Of these, 190 used existing tools (NOS (85/190, 44.7%), ROBINS-I (46/190, 24.2%), and JBI (11/190, 5.8%)) and 13 used an author-developed tool (13/203, 6.4%). Tools were adapted in 16.7% (34/203) of reviews and 7.2% (14/203) used multiple tools. Reviews included 20 (±25.7) observational studies, commonly cohorts (175/203, 86.2%), with COVID-19 (66/203, 32.5%) and seasonal influenza (62/203, 30.5%) frequently studied. VE was reported descriptively in 25.1% (51/203) of reviews, while 74.9% (152/203) provided meta-analyzed estimates primarily based on laboratory-confirmed infection (137/203, 67.5%) and symptomatic disease (130/203, 64.0%). Our findings indicate heterogeneous RoB assessment, reflected by use of different/multiple tools, frequent adaptations, author-developed methods, and absence of RoB assessment, highlighting the need for clearer guidance or tailored tools.