Universitat Internacional de Catalunya

Health IA

Health IA
3
16516
4
First semester
op
Main language of instruction: Catalan

Other languages of instruction: English, Spanish,

Teaching staff


Student consultations will take place by appointment, arranged via the professor's institutional email: jgonzalezs@uic.es.

Introduction

This course introduces the fundamentals, applications, and evaluation criteria of artificial intelligence in the healthcare sector from a bioengineering perspective. It covers examples related to medical imaging, biomedical signals, sensors, devices, biomedical research, and clinical decision support. Students will learn to identify a need, assess the quality of data and scientific evidence, and formulate a proposal that takes into account users, safety, ethics, and implementation conditions. The course is structured around an individual project involving the design and evaluation of an artificial intelligence application for healthcare. The course is taught in English.

Pre-course requirements

It is recommended to have a basic understanding of biostatistics, physiology, biomedical instrumentation, signals and systems, and ethics, as acquired in previous courses of the degree program. Students must be able to follow classes and consult scientific literature in English. No prior knowledge of programming or artificial intelligence model training is required.

Objectives

 To understand the main families of artificial intelligence systems and their applications in bioengineering and health.
 Define health problems, intended uses, and technological alternatives based on specific needs.
 Acquire criteria for analyzing data, interpreting results, and critically evaluating scientific evidence.
 Integrate human factors, safety, ethics, data protection, and regulatory requirements into the design and evaluation of a proposal.
 Develop and defend a well-founded individual proposal, identifying its feasibility, limitations, and validation requirements.

Learning outcomes of the subject

Upon completion of the course, students will be able to:
1. Explain the differences between rule-based systems, machine learning, deep learning, generative artificial intelligence, and agents, and recognize their applications and limitations.
2. Define a health need, the users, the population, the context, the decision, and the usage boundaries of a proposal, comparing it with a non-AI alternative.
3. Analyze data provenance, quality, and representativeness, and identify biases, labeling issues, information leakage, and distribution shifts.
4. Interpret metrics, thresholds, calibration, and validation results, and assess whether the evidence justifies clinical or care-related utility.
5. Design a workflow specifying user roles, human oversight, and actions to be taken in the event of errors or exceptional situations.
6. Evaluate the ethical, equity, accessibility, privacy, and sustainability implications of an application.
7. Identify the risks and the regulatory, quality, and safety issues that must be reviewed prior to potential implementation in a healthcare setting.
8. Recognize basic cybersecurity threats and propose measures for prevention, response, and secure continuity.
9. Formulate verifiable requirements and a plan for the validation, monitoring, updating, and decommissioning of a system.
10. Prepare and defend—in English—a well-founded recommendation, citing identifiable sources, defined responsibilities, and explicit uncertainties.

Syllabus

Block 1 Fundamentals and problem definition
Main families of artificial intelligence and applications in bioengineering. Relationship between model, system, and medical device. Health need, population, users, context, intended use, and non-AI alternatives.
Block 2 Biomedical data and applications
Images, signals, sensors, text, and clinical data. Quality, representativeness, labeling, privacy, and traceability. Biases, confounding factors, data leakage, and data drift. Conceptual selection of the technological approach.
Block 3 Performance and scientific evidence
Confusion matrix, prevalence, metrics, thresholds, calibration, and uncertainty. Internal and external validation. Comparators, generalization, and clinical utility. Literature search and critical appraisal, implementation experiences, and commercial claims.
Block 4 Users and healthcare organization
Workflows, responsibilities, capacity, and resources. Human factors, professional oversight, accessibility, and communication. Errors, disagreements between people and systems, workload, and alternative procedures.
Block 5 Ethics and safety
Autonomy, beneficence, non-maleficence, equity, transparency, and sustainability. Data protection, regulation of artificial intelligence and medical devices, quality, and risk management. Cybersecurity and safe use of generative systems and agents.
Block 6 Proposal development and evaluation
Verifiable requirements, interoperability, comparison of solutions, and feasibility. Design of a potential pilot,
outcome measures, and stopping criteria. Monitoring, incidents, updates, and withdrawal. Preparation of the evidence
dossier and defense of the final recommendation.
The content is covered over twelve learning and project-work sessions. The thirteenth session
is dedicated exclusively to final presentations and defenses.

Teaching and learning activities

In person



Teaching combines participatory lectures, case analyses, critical reading of scientific articles, data interpretation exercises, and guided individual work. Sessions last two hours and integrate an initial presentation of concepts with activities focused on application and project development. Each student progressively compiles a single individual report; session activities contribute to its various sections and allow for the revision of earlier decisions. Three formative checkpoints are scheduled to discuss the evidence, feasibility, and final recommendation. Instructor reviews and feedback are incorporated directly into the work.
Independent work includes consulting the literature, analyzing sources and cases, drafting the report, and preparing for the defense. The project is a design and evaluation proposal; it does not require programming, model training, or obtaining patient data. Consultations with professionals, where possible and authorized, will complement documentary sources.

Evaluation systems and criteria

In person



Assessment is based on the individual project developed throughout the course and on the combined component of oral defense and attendance, with the following weighting:
Activity | Weighting
Individual project | 70%
Oral defense and attendance | 30%
The project assessment will consider the definition of the problem and intended use, the quality and traceability of sources, the critical analysis of evidence, the justification of decisions, the feasibility of the proposal, and the consideration of risks and limitations.
The defense assessment will consider understanding of the work, clarity of presentation, and the ability to justify decisions and answer questions. Attendance is included in the same 30% component. Project progress checkpoints serve a formative purpose and do not constitute activities with additional weighting.
Students must identify the sources used and take responsibility for the decisions and conclusions reached. When artificial intelligence tools are used for support, their use must be declared, and the results critically verified. The handling of academic issues will be in accordance with UIC regulations.

Bibliography and resources

Basic materials: textbook, presentations, worksheets, and a selection of scientific articles provided by the lecturer on the virtual campus.
Reference bibliography:
World Health Organization. Ethics and governance of artificial intelligence for health. 2021.
World Health Organization. Regulatory considerations on artificial intelligence for health. 2023.
Regulatory resources for consultation:
European Union. Regulation (EU) 2024/1689 on artificial intelligence.
European Union. Regulation (EU) 2017/745 on medical devices.
European Union. Regulation (EU) 2016/679 General Data Protection Regulation.
Topic-specific bibliography and reference excerpts will be indicated on the virtual campus.