PhD Programme in “Health Sciences”

Curriculum: “Biostatistics and Methods for Big Data Analysis”

The Curriculum includes the following topics and research areas, primarily applied to chronic neurodegenerative diseases:

1. Models for the Study of Prognostic Factors, Effect Modifiers, and Surrogate Endpoints

• Effect modifiers: Concept of interaction and its interpretation in statistical models (e.g., biological vs. statistical interaction).
• Basic statistical models: Study and application of linear regression models, logistic regression, and Cox models for survival analysis.
• Model validation: Study of internal and external validation techniques.
• Surrogate endpoints: Definition, validity criteria, and use in clinical practice and clinical trials.

2. Meta‑analysis

3. Models for the Study of Disability Trajectories

• Difference between static and dynamic models in the study of disability.
• Latent growth models: Use of structural equation models to study trajectories.
• Mixed models (multilevel models): Application of random‑effects models to analyze intra‑ and inter‑individual variation.
• Latent trajectories: Use of latent class models to identify subgroups of individuals with similar disability trajectories.
• Longitudinal disability measurement: Validation of disability scales over time and their use in modeling.

4. Methodological Issues in the Quantitative Study of Disease Progression in Chronic Diseases

• Measurement of progression: Selection and validation of quantitative indicators of progression (e.g., disease scores).
• Censoring and truncation issues: Statistical methods for handling censoring in survival data.
• Nonlinear models: Proportional hazards models, accelerated failure time models, and flexible models for disease progression.
• Survival bias and immortal time bias: Methods to address biases in progression data.
• Missing data and dropout: Study of multiple imputation techniques and models for handling missing data in longitudinal studies.

5. Early Prognostic Markers and Treatment Response Markers

• Identification of markers: Machine learning techniques for discovering prognostic and predictive markers.
• Validation studies: Design of studies for biomarker validation.
• Biomarker‑based prognosis: Study and validation of models incorporating biomarkers.
• Dynamic prognosis: Study of statistical models that dynamically update prognostic estimates over time based on markers.
• Treatment response markers: Statistical approaches to evaluating treatment response using markers (e.g., pharmacogenetics).

6. Methods for Big Data Analysis

• Tools for Big Data management and analysis
• Machine learning and artificial intelligence: Application of ML techniques for large‑scale dataset analysis.
• Predictive analytics: Regression and classification models for forecasting using Big Data.
• Privacy and security: Ethical and legal aspects of Big Data use in healthcare.
• Data visualization: Advanced data visualization techniques for exploring large datasets.