An adaptive real-time intelligent system to enhance self-care of chronic disease (ARISES)

Abstract

Diabetes mellitus is an increasingly prevalent chronic metabolic condition characterised by impaired glucose homeostasis and raised blood glucose levels (hyperglycaemia). Broadly categorised as either type 1 (T1DM) or type 2 diabetes (T2DM), people with diabetes are largely responsible for self-managing their blood glucose levels. Despite the development of diabetes technologies such as real time continuous glucose monitoring (RT-CGM), many individuals are frequently exposed to iatrogenic low blood glucose levels (hypoglycaemia). Severe hypoglycaemia is associated with an increased risk of recurrent hypoglycaemia, impaired symptomatic awareness of hypoglycaemia, and potentially death if left untreated.

This thesis affirmed the existing clinical impact of severe hypoglycaemia and its recurrent risk in a six-month analysis of severe hypoglycaemia attended by the London Ambulance Service NHS Trust (LAS). Fewer incidents of severe hypoglycaemia observed in a date matched repeat analysis during the 2020 COVID-19 lockdown suggested improved self-management possibly motivated by a proximal fear of hospitalisation and improved structure at home. Finally, a 12-week randomised control trial demonstrating a significant difference in time spent in hypoglycaemia <3mmol/L, is the first study to prove the immediate provision of RT-CGM significantly reduces the risk of recurrent hypoglycaemia. Moreover, it highlighted the impact of socioeconomic disparity as a barrier to effective hypoglycaemia risk modification. This guided the design of an adaptive real time intelligent system to enhance self-care of chronic disease (ARISES) aimed to deliver therapeutic and lifestyle decision support for people with T1DM.

The ARISES graphic user interface (GUI) design was a collaborative process conceived in a series of focus group meetings including people with T1DM. Finally, a 12-week observational study using RT-CGM, a physiological sensor wristband, and a mobile diary app, allowed for a sub-analysis identifying measurable physiological parameters associated with current and impending hypoglycaemia in people with T1DM.