BARC Advanced biostatistics for medical research

The educational activities of the Master’s programme in Advanced biostatistics for medical research are organised into five modules, covering topics such as:

• Module 0: Preliminary Course of R
• Module 1: Linear and non-linear models for randomized and non-randomized clinical trials
• Module 2: Estimands and Missing Data
• Module 3: Bayesian phase 2–3 designs for pharmaceutical and biomedical device trials
• Module 4: Network Meta-Analysis, basket and umbrella trials
• Module 5: Propensity scores for non-randomized clinical trials

The training programme is aimed at providing an in-depth understanding of advanced statistical methods for clinical research.

The Master’s Degree in Advanced Biostatistics for Medical Research provides technical and scientific training for professional statistics figures with high skills in identifying, designing and analysing clinical studies with innovative, non-standard methods. Moreover, one of the specific figures is that of the biostatistician who works or wants to work in the public or private pharmacological research field.

The topics covered are relevant for the pharmaceutical industry and CRO (Contract Research Organizations), which need to develop internal competencies or acquire personnel trained on them. The Master’s Degree bridges this gap, by promoting a statistical preparation closer to the needs of companies and to new emerging methodologies.

Module 1: Linear and non-linear models for randomized and non-randomized clinical trials
Clustered data are common in clinical studies, such as longitudinal studies, repeated measures, or block experiments, where subjects within the same group share similar characteristics and outcomes. In these situations, traditional statistical methods assuming independence among observations may produce biased results. Mixed linear models allow the correct analysis of correlated data by modelling the dependence between units belonging to the same cluster and improving the reliability of estimates.

Module 2: Estimands and Missing Data
Preserving randomization is essential to ensure validity and reduce bias in clinical studies. Protocol deviations occurring after treatment allocation may affect results, particularly when they are related to the treatment itself; for this reason, the intention-to-treat principle is often adopted, keeping subjects in the analyses according to their original assignment. Despite the central role of clinical trials in regulatory decision-making, there is still no universally accepted regulatory definition of the primary estimand in confirmatory studies, leading to methodological heterogeneity across study designs.

Module 3: Bayesian phase 2–3 designs for pharmaceutical and biomedical device trials
Clinical studies on drugs and medical devices increasingly require the management of complex, costly, and correlated data. Bayesian experimental designs make it possible to incorporate information from previous studies or the scientific literature, reducing time, costs, and patient exposure to less effective treatments. They also provide great flexibility, allowing updates to the experimental design and predictive evaluations of possible outcomes. The FDA has recently promoted the use of hierarchical Bayesian models to incorporate historical data into medical device studies. The application of these approaches has been facilitated by computational advances and MCMC methods implemented in dedicated software. This module provides a systematic overview of Bayesian models, including both methodological and regulatory aspects.

Module 4: Network Meta-Analysis, basket and umbrella trials
Systematic reviews and meta-analyses are essential tools for evaluating treatment effectiveness. When multiple therapeutic interventions must be compared, Network Meta-Analysis allows the integration of direct and indirect evidence into a single analysis, even in the absence of head-to-head comparisons among all treatments. The module explores the methodological assumptions of the model and the graphical representation of results through SUCRA plots and rankograms. In the Bayesian framework, tools for assessing the convergence of MCMC algorithms are also presented, with particular attention to Gelman–Rubin diagnostic statistics.

Module 5: Propensity scores for non-randomized clinical trials
This module introduces the principles of causal inference and the main quantities of interest, such as ATE, ATT, and ATC. The conditions required for valid causal inference are examined, including randomization and adequate sample size. The Propensity Score is presented as a key tool for reducing confounding bias in observational studies. The main propensity score–based techniques are illustrated, including matching, stratification, covariate adjustment, and inverse probability weighting. Particular attention is devoted to Propensity Score Matching, from confounder selection and score estimation to checking common support and assessing matching quality. The module concludes with a comparison between propensity score approaches and traditional multivariate regression methods, highlighting their advantages and practical limitations.

The second level Master’s Degree in Advanced biostatistics for medical research is issued online, so that it can be followed also by full-time workers; it is issued on demand through UniPD multimedia Moodle platform, and video-lessons are available 24/7. It has been designed for students and professionals who want to combine other professions and activities with the need to qualify or further specialize.

Lessons will start in November 2025, and the course will last one year.

There will be a frequent and easy interaction between students and professors, through Moodle’s Forum.

The Master is divided into several 4-week modules, between November and May. At the end of each module, attendees will have the time to rewatch the video-lessons, followed by on-demand homework to test the competencies acquired. The project work to discuss the final exam will be drafted between June and July, and it may include cases that specifically interest the student, as agreed with the tutor. The work – prepared during the summer – will be the basis for the diploma-awarding discussion in September, on the Zoom platform.

For more information on the Directors and Professors, and for useful insights on the Master’s Degree in Advanced biostatistics for medical research, here’s the presentation video:  Advanced Biostatistics for Medical Research | Department of Cardiothoracic-Vascular Sciences and Public Healthcare | University of Padua (unipd.it)

The general ranking of merit for the academic year 2026/27 will be published on the Italian page of this Master according to the timing provided in the Call.