Automated insulin delivery systems (AID) are the standard of care for people with type 1 diabetes (T1D) in the United States (US).1,2 AID algorithms use continuous glucose monitoring (CGM) data to automatically adjust insulin delivery in an effort to optimize glycemia while also decreasing the burden of T1D management.3 Clinicians and people with type 1 diabetes (pwT1D) considering an AID system must be aware of the unique features of each system, including: the control algorithm, settings that can be modified to meaningfully alter insulin delivery, user-inputs needed for optimal system function, and wearability features (e.g. insulin cartridge capacity, presence of tubing, smartphone operability).

Clinical trials investigating AID systems are the gold standard for evaluating the safety and efficacy of AID systems in pwT1D. In clinical trials, racial/ethnic minority groups, publicly insured individuals, and those not attaining glycemic targets are often underrepresented while participants from wealthier and more highly educated households who are highly engaged with T1D care are overrepresented.4 These underrepresented groups have been shown to have higher hemoglobin A1c (HbA1c), lower technology uptake, and higher rates of discontinuation, leading to sampling bias that limits the generalizability of the results.5,6 Real-world evidence (RWE) capturing more representative populations is essential to understand the effects of AID use in people with varying degrees of engagement with T1D management. This review will summarize currently available literature for FDA approved AID systems in the US and their unique features with a focus on RWE.