Article type
Abstract
Background: Non-randomized studies of interventions (NRSI) provide valuable insights into the real-world performance of interventions, particularly when randomized controlled trials (RCTs) are impractical, unethical, or lack generalizability. However, we do not know whether their inclusion in Cochrane reviews is justified, and whether the analytical methods and reporting of results are adequate and if guidance is being followed.
Objectives and Methods: The guidance for including NRSI in Cochrane reviews was updated in 2019 in version 6.0 of the Cochrane handbook. We will search for reviews in the Cochrane Library including NRSI, published either in 2019 or 2023 to investigate how methods and reporting quality have changed over the last five years. We will focus on the following domains:
1.) Are authors following Cochrane Handbook guidance, including the use of study features instead of study labels, and ensuring justification of NRSI inclusion?
2.) How are NRSI reported in the Summary of Findings tables?
3.) How are NRSI analyzed (meta-analysis, narratively)? Are the unadjusted or adjusted estimates extracted? Are the confounding factors reported for the adjusted estimates?
4.) Are NRSI and RCTs analyzed together (which models are used) or separately?
5.) What common errors are observed when synthesizing NRSI?
6.) Which risk of bias assessment tools are used?
Impact of the Research: We propose to present the findings of our research at the Summit where we will highlight good practice and provide advice for authors who plan to include NRSI in their review. We will be able to show how closely authors are adhering to guidance in the Handbook, and to MOOSE (Meta-analysis Of Observational Studies in Epidemiology) guidance. It will also help to identify areas in which review authors require further guidance to support the production of reviews, including NRSI. Our research will also be relevant to users of reviews that include NRSI, with information about good practice and providing robust evidence. Such reviews are not only important in the field of health research but also other areas in which NRSI are more appropriate than RCTs to answer certain research questions, such as about climate change, education, and crime and justice.
Objectives and Methods: The guidance for including NRSI in Cochrane reviews was updated in 2019 in version 6.0 of the Cochrane handbook. We will search for reviews in the Cochrane Library including NRSI, published either in 2019 or 2023 to investigate how methods and reporting quality have changed over the last five years. We will focus on the following domains:
1.) Are authors following Cochrane Handbook guidance, including the use of study features instead of study labels, and ensuring justification of NRSI inclusion?
2.) How are NRSI reported in the Summary of Findings tables?
3.) How are NRSI analyzed (meta-analysis, narratively)? Are the unadjusted or adjusted estimates extracted? Are the confounding factors reported for the adjusted estimates?
4.) Are NRSI and RCTs analyzed together (which models are used) or separately?
5.) What common errors are observed when synthesizing NRSI?
6.) Which risk of bias assessment tools are used?
Impact of the Research: We propose to present the findings of our research at the Summit where we will highlight good practice and provide advice for authors who plan to include NRSI in their review. We will be able to show how closely authors are adhering to guidance in the Handbook, and to MOOSE (Meta-analysis Of Observational Studies in Epidemiology) guidance. It will also help to identify areas in which review authors require further guidance to support the production of reviews, including NRSI. Our research will also be relevant to users of reviews that include NRSI, with information about good practice and providing robust evidence. Such reviews are not only important in the field of health research but also other areas in which NRSI are more appropriate than RCTs to answer certain research questions, such as about climate change, education, and crime and justice.