Abstract
Dietary intake influences glycaemic outcomes and performance of automated insulin delivery (AID) systems; however, dietary assessment and reporting practices in AID research have not been described. This scoping review mapped how dietary intake data are collected and reported in AID studies and used that map alongside an established dietary assessment framework to propose initial framework-derived considerations for terminology, reporting and method selection for dietary data. CINAHL, Embase, Emcare, Ovid MEDLINE, and grey literature were searched for English-language studies published from 2005 to May 2025. Eligible studies involved AID users and reported dietary assessment methodology; studies with investigator-controlled meals and no dietary assessment were excluded. Charted data were summarised narratively and in tabular form; considerations were then derived by applying an established dietary assessment framework to the summarised patterns. Fifty-three studies were included, predominantly from high-income countries and focused on type 1 diabetes. Dietary intake was primarily from technology data (carbohydrate announcements to pumps or algorithms, 75%), of which 11 (28%) used a second dietary assessment method, and seven permitted verification of announced carbohydrate. Traditional dietary assessment tools were less frequently used. Units, terminology, meal definitions, and postprandial observation windows differed across studies, with reporting focused on carbohydrate quantity. AID research is constrained by inconsistent dietary assessment methods, reliance on unverified carbohydrate announcements, and variable reporting practices. The considerations presented here offer a starting point for standardisation.
Open Science Framework: https://doi.org/10.17605/OSF.IO/8P7J2.