Go to the content (press return)

Eines algorítmiques per a l'anàlisi de l'evolució temporal de pacients en els sistemes de salut

Type of activity
Competitive project
Funding entity
AGAUR. Agència de Gestió d'Ajuts Universitaris i de Recerca
Funding entity code
2018 DI 069
33.960,00 €
Start date
End date
Modern healthcare systems face the high complexity of efficient management because of the high prevalence of complex, chronic disease related to aging. 5% of the patients use up to 50% of the resources, and the high costs they create are threatening the very survival of the systems.

A key step in planning and organizing healthcare system is to understand how these high-cost patients use the system its resources and how they are going to evolve in the future. The goal of the thesis is to create algorithms for discovering, analyzing, visualizing and interpreting patient "trajectories". These come in several flavors, such as:

- temporal disease trajectories. How do patients evolve over time, which complications and risks they are going to develop progressively, which resources are they going to use or need.

- patient mobility: how are patients distributed in territories, have we provisioned the resources necessary to assist them, and if some go elsewhere than we imagined, which ones and why.

- system mobility. Which facilities do patient use, how do they move across levels (primary care, emergencies, hospital, nursing centers, mental health). Are we treating each patient at the level that best balances complexity, costs and required attention? How can we redirect the flows?

The research will likely involve techniques such as frequent sequence mining, latent variable models (HMM-like), predictive models, clustering, and process mining. Visualization and interpretation are necessary for bringing the method to real, practical use. Associated to the algorithms there has to be a methodology for their optimal use.

The algorithms and methodologies developed will be incororated into the interactive Big Data tools that Amalfi Analytics is developing. Implementability issues and time/memory efficiency are therefore of prime importance. Also, the candidate will have to participate in the frequent meetings with clinicians and healthcare managers of the institutions with which Amalfi co-designs their tools.
Adm. Generalitat
Estratègia de recerca i innovació per a l'especialització intel·ligent de Catalunya (RIS3CAT)
Resoluton year
Funcding program
Funding call
Doctorats Industrials
Grant institution
Agència De Gestió D'ajuts Universitaris I De Recerca (agaur)