#2020 The use of the target trial approach in kidney failure: a systematic methodological review

Abstract

Abstract Background and Aims The concept of target trial emulation has recently been introduced and implemented in nephrology as a complement to when randomized trials may not be feasible. Thereby critical questions in the comparative effectiveness and safety of interventions during kidney failure can be addressed. This methodological systematic review aims to evaluate the scope, validity and robustness of the target trial approach for studying outcomes in patients with kidney disease. Method MEDLINE, Embase, and reference lists were searched up until 1st June 2023 to identify studies on kidney disease patients that use target trial emulation. Dual study selection (AA, PK) was undertaken, with a third reviewer settling disagreements (JP). Interventional observational studies were eligible that used target trial emulation to mimic a hypothetical trial where the conduct of a randomized trial was unfeasible, unethical, or untimely. We included studies across the range of chronic kidney diseases, i.e., predialysis, dialysis and transplantation; studies assessing acute disease states (including acute kidney injury) were excluded as well as non-primary research (i.e., case series/editorials/reports/narrative reviews). The quality of reporting and risk of bias was assessed independently by three reviewers and differences resolved through consensus with a fourth reviewer. The critical appraisal was piloted using a previously published reporting checklist and refined throughout the process. Results The systematic search yielded 2883 studies, of which 649 duplicates were removed. From 2232 studies, 367 full-text studies were assessed for eligibility according to target trial emulation criteria and 74 studies were included. A large number of studies (47) focused on patients with kidney failure undergoing renal replacement therapy, mainly haemodialysis. 30 studies were reported before 2019. Of the 74 studies, 46 focused on pharmacological interventions, six on surgical interventions, four on transplantation and 18 on other interventions. Studies used either registries (41), health administration data (4), insurance data (5), electronic health records (15), study databases (8) or other (1). It should be noted that in multiple instances more than one data source was used. The majority (66) of studies did not explicitly mention target trial emulation in the title or abstract. Only a minority (8) of studies explicitly prespecified a target trial protocol. To mitigate confounding, studies mainly focused on exact matching of the groups by using inverse probability weighting and propensity score matching. Conclusion In the past decade the use of the target trial emulation framework in nephrology has substantially increased. However, reporting and quality of study conduct is variable and shows the clear need for a reporting framework to increase clarity and reliability, and to understand to what extent study findings are generalisable into clinical practice to improve health outcomes