Periodic Reporting for period 1 - MELISSA (Mobile Artificial Intelligence Solution for Diabetes Adaptive Care)
Reporting period: 2022-06-01 to 2023-11-30
Achieving near-normal glycaemic control remains highly challenging for most people with type 1 or type 2 diabetes requiring intensive insulin treatment, despite advances in insulin delivery and glucose monitoring technology. Daily insulin needs for people with diabetes (PwD) fluctuate due to factors like carbohydrate – CHO intake, physical activity, health conditions, and variables like mood, temperature, and insulin sensitivity. While the effect of some of the known factors can partly be mitigated by the PwD adjusting their daily insulin dosing, the effect of other (known and unknown) factors remains an obstacle to achievement of optimal glycaemic control and quality of life due to hyper- and hypoglycaemic excursions resulting from ‘erroneous’ insulin dosing. Consequently, many PwD do not reach recommended glycaemic targets and remain at increased risk of developing devastating short- and long-term complications related to poor glycaemic control. At present, systems for decision support with regards to daily insulin dosing for PwD treated with Multiple Daily Injections (MDI) are limited to the coverage of basal insulin requirements and to simple bolus calculators of meal-related insulin administration working with fixed algorithms that rely heavily on self-estimated CHO intake. Enhancement of algorithms by Artificial Intelligence (AI) may have a considerable potential to further improve daily decision-making for many PwD by compensating for the effect of both known and unknown factors affecting insulin needs that the person may not be able to manage. MELISSA’s main objective is to provide a clinically validated, efficient and cost-effective AI-based digital diabetes management solution to support PwD on insulin in providing personalized treatment and care recommendations, with the primary goal of achieving and maintaining normoglycaemia. The introduced digital solution is based on the combined use of already prototyped AI-approaches and innovative tools for quantification of lifestyle and behavioural factors, taking into consideration gender, age, and socio-economic parameters related to the management of diabetes.
MELISSA is structured into work packages (WPs) focused on different project aspects. A governance structure oversees progress and coordination, while a website, social media accounts (X & LinkedIn), and a dissemination toolkit promote the project’s mission and progress, attracting nearly 1000 visits/month (website). A communication committee and a patient advisory committee (PAC) were formed. The PAC meets monthly and provides valuable feedback to the MELISSA researchers. The original version of AI-powered Adaptive Basal-Bolus Algorithm (ABBA) to support personalised insulin treatment was further optimised by considering both current and past blood glucose values, adding the insulin-on-board feature and by personalised tuning of a Reinforcement Learning (RL) algorithm. The new version of ABBA was validated in-silico in simulated adults and adolescents with type 1 and adults with type 2 diabetes, showing ~5% increase in time spent in normoglycaemic range and reductions in times spent above and below range compared to standard bolus calculators. The goFOODTM was improved and refined with respect to food item segmentation and recognition. The application now consists of three forms, i.e. goFOODTMLite (only data collection), goFOODTMMini (data collection, qualitative feedback on CHO content) and the full version of goFOODTM (data collection & quantitative estimation of CHO). The MELISSA app was developed and screens for the functions were designed. User Requirements Questionnaires for PwD and healthcare professionals, developed in multiple languages, are now on RedCap. The PwD version is seeking ethical approval at some institutions and is distributed at others with waived approval. The clinical validation study design for MELISSA has been finalized, with the protocol written and nearing submission for ethics approval across all study sites. Standard Operational Procedures have been drafted, the Data Management Plan has been written and the data management system has been built in the RedCap platform. Contracts with manufacturers like Novo Nordisk (for NovoPen 6 and provision of insulin), Abbott (for FreeStyle Libre sensors), Dexcom (for G7 sensor), and activity tracker makers (Fitbit, Garmin, Withings) have been made for wireless communication with the MELISSA app. A systematic review with meta-analysis on the effect of available Bolus Advisors (BA) on glucose control and quality of life was written and submitted for publication in a scientific journal. The MELISSA solution has been analysed using the business canvas framework, establishing its value proposition and ensure alignment with the true needs of users. The process included a Needs Assessment through a design thinking workshop, literature review, additional interviews with Key Opinion Leaders and members of the PAC. Current workflow has been mapped for the participating countries through interviews with the clinical partners of the project. A scoping review of the existing regulatory framework (Class III medical Device/CE Marking and GDPR) is underway to identify potential issues and develop an action plan. A publication committee has been established and publication guidelines have been developed to ensure fair credits to (co-) authors and to guarantee open access publication. Finally, on request of the EC, an ethics WP was added to the project and two ethics advisors have been appointed for providing feedback on the project.
ABBA is based on an RL-based algorithm that tackles challenges in online RL and personalized insulin recommendations for PwD on MDI, offering quicker convergence and stable suggestions. A patent is being considered due to its innovations.
goFOODTM uniquely utilizes smartphone sensors, supporting Android and iOS, enabling expert-level accuracy in estimating calories and macronutrients.
The combination of ABBA and goFOODTM enables unprecedented personalized insulin adjustment through automatic meal quantification, marking a major innovation in personalized healthcare.
goFOODTM uniquely utilizes smartphone sensors, supporting Android and iOS, enabling expert-level accuracy in estimating calories and macronutrients.
The combination of ABBA and goFOODTM enables unprecedented personalized insulin adjustment through automatic meal quantification, marking a major innovation in personalized healthcare.