-The impact of mobile technologies on healthcare is particularly evident in the case of self-management of chronic diseases, where they can decrease spending and improve the patient quality of life. In this position paper we propose the adoption of agent-based modelling and simulation techniques as built-in tools to dynamically monitor patient health state and provide recommendations for self-management. To demonstrate the feasibility of our proposal we focus on Type 1 Diabetes Mellitus as our case study, and provide some preliminary simulation results.
I. INTRODUCTION
as self-management interventions [
IoMT holds promise to pave the way for a new era in
medicine, changing the way health-services will be provided
[
A number of improvements are expected in healthcare by the introduction of IoMT: • increasing accessibility of health-services, by guaranteeing a wider coverage; most in fact own a mobile device with which they can access diverse services devoted to improving their health, from the simplest cases of SMS reminders with dates of appointments, SMS or emails for communications from the health professionals, SMS or emails with medical reports, to the more complex cases of data acquisition (via sensors), transmission, and elaboration; • decreasing the cost of healthcare, since people can avoid to frequently move towards healthcare facilities and, caregivers can be automatically updated in case of unexpected changes; • supporting chronic disease self-care, by enabling data collection, sharing and disease tracking, to support diagnosis and personalised treatment —well turn to in more detail on that in the following; • providing suitable tools for timely managing healthcare in emergencies.
III. SELF-MANAGEMENT OF CHRONIC DISEASES
Self-management of chronic diseases is defined as the
active involvement of patients in their treatment with
day-today decisions about different actions to be taken: control of
symptoms, take medicines, make lifestyle changes, undertake
preventive actions. It is thus characterised by an extensive
responsibility that the patients need to take on [
IoMT technologies can significantly improve disease
selfmanagement [
In this paper we propose the adoption of modelling and
simulation tools that are gaining acceptance in medicine as
a valuable support for decision making, since they provide
both, short-term and long-term clinical predictions of patient
health [
In the literature, agent-based systems, and MAS in
particular, are considered an effective paradigm for modelling,
understanding, and engineering complex systems [
In the pioneering work of Bonabeau [
The benefits of ABM over other modeling
techniques can be captured in three statements: (i) ABM
captures emergent phenomena; (ii) ABM provides
a natural description of a system; and (iii) ABM
is flexible. Emergent phenomena result from the
interactions of individual entities. By definition, they
cannot be reduced to the systems parts: the whole is
more than the sum of its parts because of the
interactions between the parts. An emergent phenomenon
can have properties that are decoupled from the
properties of the part. [...] ABM is, by its very
nature, the canonical approach to modeling emergent
phenomena: in ABM, one models and simulates the
behavior of the systems constituent units (the agents)
and their interactions, capturing emergence from the
bottom up when the simulation is run. [
We model the whole healthcare system as an agent-based system composed of two levels, as shown in Figure 1: 1) a high-level model that represents patients and their interactions with a tool supporting diabetes selfmanagement, such as a PDA; 2) a disease model that represents the disease physiopathology and predicts the state of the patient in the near and far future.
The high-level model reproduces the behaviour of patients, and how they respond to feedback received through their personal devices. We assume that PDAs acquire (1) information from the individual about the composition of their meals, (2) information about individual physical activity from embedded sensors, such as accelerometers, and finally (3) information about glycemia values, gathered wirelessly from a wearable device. These data are used as input for the chronic disease model. In the following we demonstrate the feasibility of our proposal by adopting Type 1 Diabetes Mellitus as our case study.
IV. BACKGROUND ON TYPE 1 DIABETES MELLITUS
PHYSIOPATHOLOGY
Diabetes mellitus, diabetes in short, is a chronic metabolic disorder characterised by an excessive amount of sugar circulating in the blood plasma, i.e., hyperglycaemia, for a prolonged period. Diabetes is strictly related to insulin defects. Insulin is a hormone produced by the β-cells of the pancreas; it has a crucial role in the absorption of glucose in twothird of body cells, but mainly in fat, liver, and muscle cells, where glucose is a necessary source of energy for the cells to perform their activities. Diabetes is due to either an autoimmune process, where pancreatic β-cells do not produce enough insulin (Type 1 diabetes mellitus, Type 1 DM in short), or else the cells of the body develop a sort of “resistance” to insulin action, thus not responding properly to the insulin produced (Type 2 DM). In the following, we first describe the metabolic system, then how the physiological functions are affected by the lack of insulin, as in Type 1 DM.
Every cell in our body needs fuel in order to fulfil its specific function. Glucose is the main source of such energy. It is obtained firstly from the food we eat (and drink, possibly) via the intestinal absorption, which ensures food to be broken down into the monosaccharide form of glucose that is then released into the blood. The blood flow ensures that glucose is delivered to all cells of our body. Finally, glucose diffuses from the blood into cells where it is used as an energy source via the aerobic respiration, or where it is stored as glycogen, a polysaccharide of glucose composed of varying numbers of glucose units, depending on the cell type (for instance, glycogen of muscle cells is composed of around 6.000 units of glucose, while liver cells require 30.000 units of glucose to create a glycogen molecule).
Plasma glucose levels are normally maintained within a narrow range (70−100 mg/dl) through the combined antagonistic action of the two pancreatic hormones, insulin and glucagon, which enable the uptake and release of glucose from the blood into cells and vice-versa:
Insulin — It is a hypoglycaemic hormone, i.e., it is responsible for enabling the uptake of glucose, mainly into fat and muscle cells, thus reducing the level of glucose in the blood. It is produced by pancreatic β-cells as a function of glucose concentration in the blood (formally called glycaemia): a basal secretion of insulin from β-cells is always observed, ensuring the availability of glucose at all times; instead, sinks of postprandial secretions are observed, i.e., when the blood glucose levels are high. Secretions finally stop in case of hypoglycaemia.
Glucagon — It is a hyperglycaemic hormone that promotes
the release of glucose from the liver cells into the
blood. It is produced by pancreatic α-cells when
glycaemia is low, normally in fast periods, such as
during the night. Therefore, α-cells behave in the
opposite way than β-cells: they have high secretion
rates when the blood glucose concentrations are low,
and low secretion rates when the glucose levels are
high [
Type 1 DM is characterised by the loss of β-cells. It is mainly caused by an autoimmune process, where T-cells of the immune systems attack and destroy β-cells, thus leading to insulin deficiency. The insulin-dependent uptake of glucose in cells is no more possible, and the main consequences are a high level of blood glucose and low level of fuel for body cells. Typical diabetes complications include cardiovascular disease, stroke, chronic kidney failure, foot ulcers, and damage to the eyes. However, prevention and treatment are possible: Type 1 DM must be firstly managed with insulin injections; however, medical evidence shows that patients affected by Type 1 DM benefit from a healthy diet, sport activity, maintaining a normal body weight, strict controlling glycaemia, and avoiding use of tobacco. For this reason, Type 1 DM is subjected to several initiatives for supporting the self-management of the disease.
Diabetes is nowadays one of the most common and known
chronic disease. Data from the World Health Organization
(WHO) [
In the following we present an ABM of Type 1 DM selfmanagement. The whole model architecture is depicted in Figure 3: from an initial state that reproduces the health condition of a patient, a low level simulation of the metabolic system (described in the following) is performed. From the simulation results, feedbacks are provided to the patient, who could then modify his/her behaviour accordingly, if needed.
We model the metabolic system as a set of interacting agents, where each agent is a set of cells conducting the same activities. In particular we include the main organs (or set of cells) involved in metabolic processes—as shown in Figure 1. Agents then interact via an interaction medium, the environment, that models the bloodstream where agents release and retrieve molecules.
We consider the following agents: • intestine-cells agent — it absorbs and breaks down food substances, releasing the glucose derived by digestion in the bloodstream • β-cells agent — if glucose level in the blood exceeds the threshold of 75 mg/dl it secretes insulin in the bloodstream; this process ends as soon as glycaemia get back into physiologic values • α-cells agent — if glucose level in the blood falls below the threshold of 70 mg/dl it secretes glucagon in the bloodstream; this process ends as soon as glycaemia get back into physiologic values • liver-cells agent — if glucagon is available in the blood, it begins the process of glycogenolysis, breaking down glycogen molecules and realising glucose in the blood; plus, if glucose level in the blood exceeds the threshold of 75 mg/dl, it uptakes glucose from the blood and begins the process of glycogen synthesis • muscle-cells agent — if insulin is available free in the bloodstream, it absorbs glucose and begins the process of glycogen synthesis; plus, during physical exercises, it consumes – according to the type of activity – a balanced quantity of glycogen • brain-cells agent — it continuously absorbs glucose from the blood The model of Type 1 DM is hence obtained by simply stopping the β-cells agent.
The model is implemented on top of the MASON
infrastructure [
Figure 4 shows the dynamic over 3 days of glucose, insulin, and glucagon concentration in blood in a healthy patient.
There, the glycaemia value varies during the day following meals (breakfast, morning snack, lunch, afternoon snack, and dinner). Insulin and glucagon follow these dynamics: (1) insulin increases as a response of glucose increase, while (2) glucagon is secreted mainly during the night when patient does not eat. 425 400 375 350 325 300 275 l250 /gd225 m200 175 150 125 100 75 50 25 0
Figure 5 shows the dynamic over 3 days of glucose, insulin, network and rule-based systems – that autonomously identifies and glucagon concentration in blood in a Type 1 DM affected suggestions on the best behaviour that the patient should patient. There, the glycaemia value is no longer under control, follow in order to contain the Type 1 DM effects. and the patient enters soon into a hyperglycaemia state. Insulin is no longer produced, and glucagon also is no longer secreted, VI. CONCLUSION since glucose concentration is over the threshold. In this position paper we explored some future directions
We plan to interpret these dynamics as predictions on in the field of self-management of chronic diseases. Given the state of the patient in the close future, and to use this the widespread diffusion of chronic diseases, a set of specific information as input for algorithms – such as artificial neural interventions are suggested by the health organisations, such as PAHO and WHO. Among the others, self-management is identified as a promising approach to decrease health spending in chronic diseases and to improve the patients quality of life. IoMT seems to be crucial for implementing the idea constituting the self-management approach in the real world.
In particular here we propose the adoption of ABMS as a built-in tool – within a IoMT infrastructure – that can automatically characterise health state of patients, providing them with feedbacks for their daily life. As our case study, we focus on a specific disease, the Type 1 Diabetes Mellitus, building our first model of chronic disease self-management, thus also showing the general feasibility of our approach. Finally we provide some preliminary simulation results illustrating the behaviour of our model of the metabolic system both in a healthy individual and in a Type 1 DM patient.