Propostas Submetidas

DEI - FCTUC
Gerado a 2024-03-28 17:47:23 (Europe/Lisbon).
Voltar

Titulo Estágio

Management of Diabetes and Hypertension Patients by means of remote monitoring solutions

Áreas de especialidade

Sistemas Inteligentes

Local do Estágio

Laboratório de Computação Adaptativa do CISUC

Enquadramento

This thesis is part of the on-going funded project POWER with Altice Labs. The main goal of this thesis is to research and implement a set of algorithms and models, based on computational intelligence methodologies, to support the remote management of patients suffering from chronic diseases, in particular hypertension and diabetes.

Hypertension, known as the “silent killer”, is a leading cause of disability and death. The development of digital solutions, aiming the promotion and enhancement of life style behaviours addressing the blood pressure (BP) self-management, in order to manage the disease, improve compliance and achieve healthy living are of major importance.
Type 1 and type 2 diabetes mellitus are a serious chronic disorder with important short and long-term consequences. Management of diabetes includes insulin administration, blood glucose monitoring, meal and exercise planning, and screening for comorbid conditions and diabetes-related complications. The goal of the patients’ self-management via blood glucose monitoring is to reach glycemic control, i.e., increase amount of time-in-range and to reduce the number of hyperglycemic events.

Objetivo

The main goal of this thesis is to research and implement a solution for remote chronic patients management (in particular diabetic and hypertension patients), integrating models able to incorporate the clinical evolution history of the patient with his/her context information and to provide clinical useful feedback and recommendations to the user.
Two major applications are required two perform an adequate management of remote monitoring patients: i) algorithms for assess the current status and to predict the evolution of blood pressure (glycaemia) values, enabling the early detection of hypertension episodes (hyperglycemic events); ii) recommendation solutions, addressing interventions in life habits (e.g., meal intake, exercise, etc.) based on the current status and the future evolution of patient's condition.
Recently, a case based reasoning approach for predicting the evolution of physiological signals has been developed by the research team. This approach has the inherent advantage of being interpretable while avoiding the need of an explicit model building of the process underlying the physiological signal (usually a complex and potentially unstable approach). In parallel, as result of previous work carried out in POWER projects, some machine learning models have been implemented for blood pressure and glycaemia prediction. However, these prediction models are uni-variable, and do not take into account the effect of other variables/context.

The first goals is to extend the developed models to consider several sources of information, in particular context information such as meal intake and exercise.
The second research goal is the development of a recommendation application, related to interventions in life habits (e.g., meal intake, exercise, etc.). To address this issue, knowledge-based approach in cooperation with experts from the Centro Hospitalar e Universitário of Coimbra (CHUC) will be researched and implemented.
The third task, comprises the development of a robust, reliable and modular package, integrated the models developed in the two previous tasks. In particular the final package should be adaptable to the particular conditions, such as the number of variables that each patient collects (or not) every day (blood pressure, glycaemia, weight, adherence to prescription, physical exercise, meals, ...) .

The first goals is to extend the developed models to consider several sources of information, in particular context information such as meal intake and exercise.
The second research goal is the development of a recommendation application, related to interventions in life habits (e.g., meal intake, exercise, etc.). To address this issue, knowledge-based approach in cooperation with experts from the Centro Hospitalar e Universitário of Coimbra (CHUC) will be researched and implemented.
The third task, comprises the development of a robust and reliable package, integrated the models developed in the two previous tasks.
The research work will be supported on private databases currently available to the team as well as a new data collection study that will be performed in collaboration with the Cardiology Department of CHUC in the framework of the POWER project.

Plano de Trabalhos - Semestre 1

Work plan, first semester
-State-of-the-art on hypertension and diabetes
-State-of-the-art on multivariable prediction models
-State-of-the-art related to knowledge-based recommendation solutions
-State-of-the-art related to achieve modular solutions
-Adapt and evolve the CBR-based and machine learning models prediction framework to the blood pressure and glycaemia problems
-Use case specification

Plano de Trabalhos - Semestre 2

Work plan, second semester (500)
-Research and develop of a multivariable predictions models, in particular to be able to integrate discrete information (e.g meal, exercise)
-Research and development of the life habits intervention recommendation module
-Development of the modular package to the managment of chronic disease
-Writing the thesis

Condições

Coordinators
Paulo de Carvalho
Jorge Henriques

Observações

In case of more than one application, two students can be enrolled
There is a possibility of having one/two scholarships.

Orientador

Paulo de Carvalho
carvalho@dei.uc.pt 📩