Introduction to Biostatistics and Epidemiology

The program is composed of 4 interconnected modules.
TOPIC 1: Module of Epidemiology (3 CFU) (Risk and prognostic factors)
Most diseases are not the product of a single cause, but a complex concatenation multiple causes. In vitro and in animals' experiments are not always adequate to identify all the factors that may be responsible for a disease or accelerate its adverse outcome. On the other hand, the identification of determinants of diseases cannot, for obvious ethical reasons, be based on studies involving the experimental exposure of humans to suspected risk factors. Hence, we recognize the need for planned observational studies that do not require the intervention on the subject, but due to their nature, make it complex to interpret associations between exposure to risk factors and disease occurrence in terms of cause effect relationship.
The purpose of this module is to develop the student ability to carefully evaluate the role and importance of risk and prognostic factors, by learning the principles of planning observational studies. With this aim the students will develop skills to evaluate the measures of disease occurrence and association between risk factors and occurrence of the disease, or between prognostic factors and outcome of the disease, with particular attention to the problem of confounding.
- Introductory Lectures (10 hours):
· Measure of disease occurrence: prevalence and incidence. (2h)
· Types of epidemiologic studies and measures of association between risk factors and disease: relative risks, odds ratios, risks differences. (6h)
· Type of bias, including confounding and effect modification. (2h)
- Practical session (6 hours plus 2 hours of discussion of the final projects): Group work to disentangle and discuss the essential elements of a study, including design, inclusion/exclusion criteria, exposure and outcome assessment, statistical analysis, main results of the analysis, strengths and limitations of the study. Each group (up to four participants) may select one article from those provided by the instructor, covering cross-sectional, case-control, and cohort study designs.
- Take home points: revising theory after the practice (2 hours): Critical evaluation of study designs and risk of bias in the epidemiological practice.

TOPIC 2: Biomedical test evaluation: metrics and applications (1 CFU, 6 hours)
Screening and diagnostic tests are important tools for identifying individuals at risk of developing a disease or confirming a diagnosis based on clinical suspicion. These tests play a crucial role in the early detection of medical conditions, where timely intervention can greatly improve health outcomes.
However, these tests are not error-free, and enhancing diagnostic accuracy remains a central goal of biomedical research. As new technologies emerge, it is increasingly important to assess their performance and compare them to established standards. A comprehensive understanding of both the strengths and limitations of diagnostic tools is essential for evidence-based decision-making in clinical and public health settings.
This module provides students with the theoretical foundations and practical skills necessary to evaluate the accuracy and clinical utility of screening and diagnostic tests. Students will work with real data and use the statistical software R to calculate the diagnostic metrics introduced in lectures. Guided assignments will offer hands-on experience in test evaluation and data-driven decision-making.
Topics covered:
· Introduction to screening and diagnostic tests
· Reference intervals and test thresholds
· Measures of diagnostic accuracy: sensitivity and specificity
· Predictive values, pre- post-test probabilities: application of Bayes' theorem , the Fagan's Nomogram as a clinical tool. (2h)
· ROC curves: assessing diagnostic accuracy across different cut-off points

TOPIC 3: Drawing conclusions from samples (2 CFU, 14 hours)
Health science research relies on data collected from samples, even though the ultimate goal is to make informed decisions that impact the entire population of interest. This is possible through statistical inference, a process that allows researchers to draw conclusions about a population based on sample data. However, relying on samples instead of the entire population introduces uncertainty in the sample estimates, known as sampling error. Understanding and managing this uncertainty is essential for drawing valid conclusions on disease burden, the efficacy of a potential treatments, or public health interventions.
This module provides students with the knowledge and skills to manage uncertainty in biomedical studies through lectures and guided practical sessions with the statistical software R.
Topics covered:
· Estimation techniques: point and interval estimates
· The logic of hypothesis testing: the power and the sample size. Statistical significance and clinical relevance.
· Comparing group means: hypothesis testing with continuous data
· Comparing proportions: hypothesis testing with categorical data
· Nonparametric tests

TOPIC 4: Introduction to the use of regression models in medicine (3 CFU)
Regression models for different outcomes: the analysis of change and the linear regression model; the case-control study and logistic regression; the cohort study and the study of incidence using Cox regression. (10h)
Regression in context: descriptive, predictive or causal purpose? (4h)
Variable selection and the true model myth. (6h)