OBJECTIVES: The Target Trial (TT) framework was introduced by Hernan and Robins (2016) as an approach for reducing biases associated with using real-world data for estimating comparative effectiveness, utilising the design principles of randomised controlled trials combined with causal inference statistical methods. This review evaluates applications of the TT framework, examining the extent of its use, compliance to its key components, and key issues encountered by studies using this methodology.

METHODS: Our search strategy aimed to identify relevant studies published after Hernan and Robins’ 2016 paper and before June 1, 2021. The Cochrane CENTRAL, Medline and EMBASE databases were searched using key search terms and a citation search of Hernan and Robins' (2016) paper was conducted. Two reviewers independently screened the search results. Information on the 7 TT components were extracted (eligibility criteria, treatment strategies, assignment procedures, follow-up period, outcome, causal contrasts of interest, analysis plan).

RESULTS: 97 studies were eligible for inclusion. Cancer was the most common disease area (19.9%). Time-to-event outcomes were reported in 75 (77.3%) of studies. Inverse probability weighting and propensity score matching were the most commonly used causal inference statistical methods (56 (57.7%) and 19 (19.6%) studies, respectively), followed by cloning methods and nested trials (14 (14.4%) and 15 (15.5%) studies, respectively). Most TT components were well defined across studies. However, study baselines (often referred to as “time zero”) and causal contrasts of interest were not adequately defined in 13 (13.4%) and 20 (20.6%) studies, respectively. It was unclear whether analysis plans had been pre-specified in 80 (82.5%) studies.

CONCLUSIONS: The TT framework is being used regularly to conduct comparative effectiveness analyses using real-world data. However, studies do not always fully comply with all elements of the framework. Analyses could be improved by reporting on all key TT components and through pre-specification of analysis plans.