Sources of bias in randomized trials and non-randomized studies of interventions: a causal inference perspective
The assessment of risk of bias in randomized trials (RCTs) and non-randomized studies of interventions (NRSI) has evolved substantially in recent years. These changes are reflected in newly available tools (ROB 2.0 for randomized trials and ROBINS-I for NRSI; see www.riskofbias.info). These new tools are based on key concepts in modern causal inference.
In this course Jonathan will:
• Introduce the key features of the RoB 2.0 and ROBINS-I tools (domain-based bias assessments, use of signalling questions and algorithms to guide risk of bias judgements), and derivation of the overall risk of bias in a specified result).
• Describe the domains of bias appropriate to assessment of RCTs and NRSI.
• Describe how modern epidemiological understanding of confounding (including time-varying confounding), selection bias and misclassification bias informs the new tools.
• Explain the importance of specification of whether the effect of interest is that of assignment to intervention (the “intention to treat” effect) or that of starting and adhering to intervention (the “per protocol” effect).
Participants will use one of the new tools to assess risk of bias in either a RCT or an NRSI.
Who should attend?
This workshop is designed for clinical researchers, epidemiologists and biostatisticians who have at least some background in elementary biostatistics and basic epidemiological concepts.
COURSE FEES AND REGISTRATION
Registration now open via Trybooking.