Testing a new blood glucose management software tool for people with type 2 diabetes mellitus

This is a single-arm cross-over trial. The intervention is a novel digital glycaemic management tool that allows participants to track their dietary intake using text-based input or automated image-based input. Text-based input involves the participant text-searching a nutritional database. Automated image-based input uses the devices camera and the Passio Nutrition AI tool (Passio Inc., Palo Alto, US) that provides suggested food items based on machine learning recognition of images. The tool also integrates with a continuous glucose monitor to display the participant's blood glucose level alongside their dietary intake. This tool is novel and has been developed for this project. Participants will be instructed and educated on how to interpret the blood glucose curve to inform dietary decisions. Access to this tool and continuous glucose monitors will be provided for 20 days. Participants will be educated on the use of the application at the first study visit, prior to recording their dietary intake. Study visit one will take place immediately after the successful screening visit or at a later date, per the participants preference. This will involve how to download the application, creating an application profile, using the text-based input, using the image-based input, using additional application features such as favouriting items and creating custom food items. Study visit 2 takes place at least ten days after study visit 1. At study visit 2 participants will be shown how to connect, use, and interpret the continuous glucose monitor display. The education and instruction at study visit 1 will take approximately 45 minutes, and approximately 20 minutes at study visit 2. Adherence to the tool is assessed through a secondary objective: measured energy intake through the application compared to objectively measured energy expenditure. Adherence is promoted through the use of a novel prompting system that sends reminders to the users via their device notification system based on user-specific metrics; please see the protocol for full details. A secondary objective, measuring the intra-individual and inter-individual variability of post-prandial glycaemic responses, will be assessed by providing participants with a high GI meal and a low GI meal at three points at the study. This is to determine if participants have consistent post-prandial glycaemic reactions to meals, for example, a consistently larger post-prandial glycaemic response to the high GI meal versus the low GI meal, and if the differences between individuals are larger than the variation within individuals. One set of meals, that is, one low GI meal and high GI meal, will be provided at visit one, and two sets provided at visit two. Participants will be instructed to consume the meals within 10 minutes after an overnight fast without any accompanying food items. The low GI meal are wheat flakes, and the high GI meal is corn flakes. Analysis of inter- and intra-individual variation will proceed with analysis run and reported separately for low-GI and high-GI meals (as this is another source of variation in responding). This will use descriptive statistics (per-participant standard deviations of outcomes) and analysis using linear mixed models with a random intercept for each participant. These models will estimate intra-individual variation (akin to an average standard deviation across all participants) and inter-individual variation, with 95% confidence intervals for each. The same analysis will provide an intraclass correlation coefficient (ICC) expressing the relative contribution of intra-individual variation to the overall variation in outcomes.