Course Syllabus
Obiettivi
L'insegnamento vuole far apprendere a studenti e studentesse come analizzare dati medici (specialmente quelli di cartelle cliniche elettroniche) attraverso tecniche di statistica computazionale, di apprendimento automatico, e di analisi di sopravvivenza per scoprire nuova conoscenza sulle condizioni dei pazienti.
Questo insegnamento si propone di fornire i concetti di base dell'epidemiologia che sono alla base di un corretto approccio metodologico a un progetto di ricerca in sanità pubblica. Lo studente sarà in grado di trattare i dati in sanità pubblica in particolare concentrandosi su diversi aspetti tra cui il disegno dello studio, la gestione e l'analisi dei dati. Lo studente sarà in grado di implementare e di calcolare indicatori di qualità/performance.
Contenuti sintetici
Dataset search and retrieval
Data preparation and data cleaning
Exploratory data analysis
Unsupervised machine learning
Supervised machine learning
Feature ranking
Result understanding and validation
R and Python programming languages
Survival analysis
Epidemiologia della popolazione
Disegni di studio
Metodi statistici con applicazione ai registri e ai dati sanitari amministrativi
Programma esteso
Dataset search and retrieval
Data preparation and data cleaning
Exploratory data analysis
Unsupervised machine learning
Supervised machine learning
Feature ranking
Result understanding and validation
R and Python programming languages
Basics in population epidemiology.
Study designs: advanced designs to combine data from different sources
(registry data, biomarkers, biobanks, surveys).
Survival analysis: survival estimate and Cox model regression.
Record linkage approaches and statistical methods with application to registries and administrative health data.
Examples of Quality/performance indicators, outcome research with administrative data, system of indicators to evaluate the appropriateness of clinical pathways in chronic diseases.
Prerequisiti
Statistica di base e basi dell'apprendimento automatico
Conoscenza di base di R o Python
Modalità didattica
Lezioni in presenza ed esercitazioni in presenza
3 lezioni di 2 ore condotte in modalità remota (asincrona)
Materiale didattico
Slides presentate a lezione ed articoli scientifici segnalati a lezione
Articoli scientifici:
Davide Chicco, Vasco Coelho (2025) "A teaching proposal for a short course on biomedical data science", PLOS Computational Biology 21(4): e1012946. https://doi.org/10.1371/journal.pcbi.1012946
Libri di testo:
Kenneth J. Rothman Sander Greenland, Timothy L. Lash . Modern Epidemiology. Lippincott Williams & Wilkins; 3rd ed.
Eric Vittinghoff, David V. Glidden, Stephen C. Shiboski, Charles E. McCulloch. Regression Methods in Biostatistics
Linear, Logistic, Survival, and Repeated Measures Models. Statistics for Biology and Health book series. Springer; 2nd edition (March 6, 2012)
Marie Reilly "Beyond classic epidemiological designs" https://www.routledge.com/Controlled-Epidemiological- Studies/Reilly/p/book/9780367186784 Chapman & Hall/CRC Biostatistics Series 2023
Periodo di erogazione dell'insegnamento
Secondo semestre
Modalità di verifica del profitto e valutazione
Lavoro personale su un progetto scientifico comprendente entrambe le unità didattiche per verificare le capacità dello studente nell'applicazione della metodologia di ricerca in sanità pubblica. Consegna di una relazione e presentazione orale del lavoro svolto per l'unità didattica Big Data in Public and Social Services.
Questionario a risposta chiusa per valutare la preparazione sul programma dell'unità didattica Big Data in Public Health.
Orario di ricevimento
Da concordare via email scrivendo a davide.chicco(AT)unimib.it o paola.rebora(AT)unimib.it
Sustainable Development Goals
Aims
This module aims at teaching students how to analyze medical data (especially, data of electronic health records) through computational statistics and machine learning techniques to infer new knowledge about the conditions of patients.
This course aims to provide the basic concepts of epidemiology that are at the basis of a proper methodological
approach to a research project in public health. The student will be able to deal with data in public health
particularly focusing on several aspects including study design, data managment and analysis. The student will be
able to implement design strategies on registries and administrative health data and able to
calculate quality/performance indicators
Contents
Dataset search and retrieval
Data preparation and data cleaning
Exploratory data analysis
Unsupervised machine learning
Supervised machine learning
Feature ranking
Result understanding and validation
R and Python programming languages
Survival analysis
Population epidemiology
Study designs
Statistical methods with application to registries and administrative health data
Detailed program
Dataset search and retrieval
Data preparation and data cleaning
Exploratory data analysis
Unsupervised machine learning
Supervised machine learning
Feature ranking
Result understanding and validation
R and Python programming languages
Basics in population epidemiology.
Study designs: advanced designs to combine data from different sources
(registry data, biomarkers, biobanks, surveys).
Survival analysis: survival estimate and Cox model regression.
Record linkage approaches and statistical methods with application to registries and administrative health data.
Examples of Quality/performance indicators, outcome research with administrative data, system of indicators to evaluate the appropriateness of clinical pathways in chronic diseases.
Prerequisites
Basic statistics and basic machine learning
Basic knowledge of R o Python
Teaching form
In-person theory classes and practice exercise classes
3 2-hour lectures conducted in a remote (asynchronous) delivery mode
Textbook and teaching resource
Classes slides and scientific papers mentioned during classes
Articles:
Davide Chicco, Vasco Coelho (2025) "A teaching proposal for a short course on biomedical data science", PLOS Computational Biology 21(4): e1012946. https://doi.org/10.1371/journal.pcbi.1012946
Textbooks:
Kenneth J. Rothman Sander Greenland, Timothy L. Lash . Modern Epidemiology. Lippincott Williams & Wilkins; 3rd ed.
Eric Vittinghoff, David V. Glidden, Stephen C. Shiboski, Charles E. McCulloch. Regression Methods in Biostatistics
Linear, Logistic, Survival, and Repeated Measures Models. Statistics for Biology and Health book series. Springer; 2nd edition (March 6, 2012)
Marie Reilly "Beyond classic epidemiological designs" https://www.routledge.com/Controlled-Epidemiological- Studies/Reilly/p/book/9780367186784 Chapman & Hall/CRC Biostatistics Series 2023
Semester
Second semester
Assessment method
Personal work on a scientific project including both teaching units to test the ability of the student in the application of research methodology in public health. Delivery of a report, and oral presentation of the work done, for the Data in Public and Social Services unit.
Questionnaire with closed answer to evaluate the preparation on the program of the teaching unit Big Data in Public Health.
Office hours
To define via email by writing to davide.chicco(AT)unimib.it or paola.rebora(AT)unimib.it
Sustainable Development Goals
Staff
-
Giulia Capitoli
-
Davide Chicco
-
Paola Rebora
