Background: Artificial intelligence (AI) is making a paradigm shift in traditional diabetes care in terms of target achieving efficacy, excellent safety, best quality care, effective monitoring and improved quality of life for both patients with Type-1 diabetes (T1D) and Type-2 diabetes (T2D). All guidelines for adult and pediatric T1D suggest that target Glycated hemoglobin (HbA1c) should be less than 53 mmol/mol (7%) but very less percent population with T1D can achieve this target due to many factors namely individual variability of insulin requirements, fear of hypoglycemia, errors in insulin dose calculations, etc. These lead to poor glycemic control, diabetes complications and a major source of stress for families and care givers especially in pediatric populations. All these affects quality of life (QoL), mental well-being and productivity of patients. So, the emergence of AI in diabetes care has revolutionized the management of diabetes care in optimizing the glycemic control by extending the ‘Target In Range’ (TIR), reducing ‘Target Above Range’ (TAR), ‘Target Below Range’ (TBR) and minimizing severe hypoglycemia. Glucose monitoring by Self Monitoring Blood Glucose (SMBG) using smart glucometer or Ambulatory glucose profile (AGP), Continuous Glucose Monitoring System (CGMS), to name a few. AI is being used in delivering insulin by smart pens, Continuous Subcutaneous Insulin Infusion (CSII), recently introduced Hybrid Artificial Pancreas and with prolonged CGMS. Smart insulin pens namely InPen, Novo Pen 6, Novo Pen Echo Plus, etc. can calculate the bolus dose considering current glucose reading, target, carb intake, probable exercise, etc. by algorithm. Furthermore, it can record the amount and timing of each insulin dose, display the last dose and insulin onboard, and wirelessly transmit the information via Bluetooth to a dedicated mobile application and transfers data automatically on the cloud for sharing with health care professionals as well. Novo Pen 6 and Novo Pen Echo Plus use near-field communication technology. CGMS uses sensor to monitor blood glucose continuously for 14 days and shows results in real-time without using a needle which made the life of young children much comfortable. It calculates TIR, TAR, TBR and probable HbA1c in a graphic presentation. Now new models are in pipeline for much longer duration. Smart glucometers are incorporated with bolus advisors to calculate insulin dosages, algorithm-driven message response to each glucose reading. The Hybrid Close Loop artificial pancreas is the new generation of CSII which automatically determines the basal rates, can determine the bolus dose as per CGMS but this feature is still under trial and having other old generation basic and advanced features. The ongoing trials results will guide further improvements and eventually will lead complete AI based diabetes management. Aim: To see the scope, feasibility and critical role of AI in diabetes care. Method: Contemporary literature is being searched to get the latest information on AI use in diabetes care. Results: PubMed, MEDLINE, CINAHL Plus, EMBASE, Cochrane and Google Scholar is being searched with search words 'Artificial Intelligence AND Diabetes', 'Artificial Pancreas', Close Loop system', 'Smart Insulin Pen'. 312 articles were found and finally 34 were selected.
Read full abstract