On August 25, 2023, PhD student Mahbub Ul Alam will present his ongoing work on “Improving COVID-19 & Early Sepsis Detection with Machine Learning & Internet of Medical Things”. The seminar takes place at the Department of Computer and Systems Sciences (DSV), Stockholm University.

Respondent: Mahbub Ul Alam, DSV
Opponent: Saguna Saguna, Luleå University of Technology
Main supervisor: Rahim Rahmani, DSV
Supervisor: Jaakko Hollmén, DSV
Professor closest to the subject: Panagiotis Papapetrou, DSV

Contact Mahbub Ul Alam

Abstract

This thesis critically examines the transformative implications of Machine Learning (ML), the Internet of Medical Things (IoMT), and Clinical Decision Support Systems (CDSSs) in contemporary healthcare landscapes. The shift toward patient-centric models has precipitated a need for personalized, participatory care, which these technologies are primed to provide. With the advent of IoMT, a potent platform for data aggregation, analysis, and transmission has been constructed, thereby empowering healthcare practitioners to render more efficacious care. The utility of IoMT has been accentuated amid the COVID-19 pandemic, notably in remote patient surveillance and controlling disease proliferation.

The fusion of ML-driven CDSSs and IoMT harbors the potential to restructure healthcare by providing real-time decision-making assistance, thereby augmenting patient health outcomes. The capability of ML to scrutinize intricate medical datasets, discern patterns and correlations, and accommodate evolving conditions bolster its predictive competencies incrementally. This thesis elaborates on the development of IoMT-grounded CDSS applications targeting afflictions such as COVID-19 and early sepsis, leveraging medical data and state-of-the-art ML methods.

The research underscores the criticality of predictive capacities, addressing pertinent issues such as data annotation scarcity, data sparsity, and data heterogeneity, as well as the preservation of security and privacy, and ensuring widespread accessibility. By concentrating on these pivotal areas and enhancing the usability and interpretability of ML models, a refined healthcare paradigm can be realized, contributing to the evolution of global healthcare. The thesis prioritizes ethical considerations, ensuring that the research adheres to the highest ethical standards. The potential repercussions of these technologies in clinical environments, specifically the deployment of the CDSS, are examined, underlining future trajectories for research and progress in healthcare technology.

Essentially, this thesis aims to enhance stakeholder comprehension in this crucial field while acknowledging the need for ongoing efforts to maintain progress.