Teaching code

4S011591

Credits

6

Coordinator

Giuseppe Verlato

Language

English

Scientific Disciplinary Sector (SSD)

MED/01 - MEDICAL STATISTICS

Courses Single

Not Authorized

Moodle Seminars 0

The teaching is organized as follows:

Learning objectives

The course aims to provide students the methodological tools to properly apply statistics to biology and medicine, and to develop the skills to use computer tools to analyze biological data in quantitative form. The course integrates the acquisition of computer skills and statistical knowledge, enabling the student to analyze and reach quantitative conclusions on analyses of experimental data in the biological and medical fields. The use of computers allows the student to directly test his or her degree of understanding of the topics covered, and to bring him or her closer to modern Machine Learning and Artificial Intelligence techniques applied to the analysis of biomedical data. The student will be able to apply the main models of multivariable analysis (multiple linear regression for quantitative outcomes, logistic model for dichotomous outcomes, Cox proportional hazards model for survival analysis) and multivariate analysis (discriminant analysis, principal component analysis, correspondence analysis) to problems in health care. He/she will know how to use Artificial Intelligence and Machine Learning techniques in the "omics" sciences (genomics, transcriptomics, proteomics, metabolomics) and diagnostics (radiodiagnostics, reading histopathological images), and will consciously distinguish between supervised and unsupervised learning techniques.
The student will understand the differences between statistical techniques (multivariable and multivariate analysis) on the one hand and bioinformatics techniques (Machine Learning and Artificial Intelligence) on the other, and will know how to use these methods in an integrated way.
Teaching will involve the integration of theory lectures, laboratory exercises and group work. Analysis of scientific articles related to the topics covered in the course will also be offered.
Upon completion of the course, the student will have acquired:
a) thorough understanding of the different theoretical foundations of generalized linear models and multivariate symmetric models on the one hand and Machine Learning methods on the other hand;
b) ability to discover patterns in data and make predictions based on these patterns and complex models to answer scientific questions in health care;
c) ability to apply the knowledge gained for critical literature review;
d) ability to choose the most appropriate analysis techniques based on the knowledge gained;
e) ability to perform analyses using simple statistical software (e.g., R, MATLAB®,...) and taking advantage of data visualization;
f) ability to work as part of a team, interpret the results of experimental analyses, and communicate them according to the standards of the scientific community.

Prerequisites and basic notions

Students should have basic knowledge of Mathematics and Statistics.

Program

------------------------
UL: Teoria
------------------------
Refreshment of knowledge on inferential methods: confidence intervals to estimate parameters, and hypothesis testing to base decisions on scientific evidence. Simple hypothesis testing: test t for unpaired and paired data, analysis of variance, correlation and regression, chi-square test.
Experimental design: parallel-group, cross-over, and factorial designs. Control and experimental groups, randomization, blinding. Computing an adequate sample size in order to achieve a given power for statistical tests, or a given precision of the estimates. Test-retest to evaluate measurement precision: repeatability versus reproducibility.
Likelihood, maximum likelihood estimate. Generalized linear models.
Models for multivariable statistical analysis: multiple linear regression to study quantitative outcomes, logistic model to study dichotomous outcomes, and Cox model to study survival.
Evidence-Based Medicine; systematic reviews and meta-analyses.
Multivariate analysis: discriminant analysis, principal component analysis, correspondence analysis.
Basic concepts of artificial intelligence and machine learning: supervised and unsupervised learning. Comparison between and integration of traditional statistical methods (multivariable and multivariate analyses) and bioinformatic methods (machine learning).
------------------------
UL: Esercitazioni
------------------------
Practical exercises will be carried out on the theoretical program carried out. Students will have to use a spreadsheet (excel) and R software to perform descriptive statistics and perform simple statistical tests (t-test for unpaired and paired data, one-way ANOVA, correlation and regression, chi-square test).

Bibliography

Didactic methods

------------------------
UL: Teoria
------------------------
The course will consist of frontal lessons in classroom.
------------------------
UL: Esercitazioni
------------------------
Lessons will be held in the computer room. Each student will use a computer equipped with statistical software (R) and will be guided to solve simple statistical problems.

Learning assessment procedures

------------------------
UL: Teoria
------------------------
The exam will be written. In the first part (theory), students will have to answer about 30 multiple choice questions with 5-8 possible answers.
------------------------
UL: Esercitazioni
------------------------
In the second part of the exam students will have to solve problems of inferential statistics, which will require to compute confidence intervals and/or to perform simple statistical tests. For this purpose, students will be enabled to use a computer equipped with a statistical software.

Evaluation criteria

------------------------
UL: Teoria
------------------------
The exam will evaluate not only the acquisition of statistical knowledge, but also the ability to adopt it critically to solve health problems. One point will be awarded for each correct answer, and zero points for incorrect or missing answers. The first part of the exam will weigh 3/5 on the final grade of the module.
------------------------
UL: Esercitazioni
------------------------
The exam will evaluate the ability to solve biological-health problems by combining theoretical statistical knowledge with the computer skills in using a statistical software. The second part of the exam will weigh for 2/5 of the final grade of the module.

Criteria for the composition of the final grade

The final mark will be expressed on a scale ranging from 0 to 30. A minimum mark of 18 will be required in order to pass the exam.

Exam language

------------------------ UL: Teoria ------------------------ Inglese ------------------------ UL: Esercitazioni ------------------------ inglese