Disegno e analisi delle prove cliniche

Course objectives

The primary objective is to provide students with the main statistical techniques for the design and the analysis of clinical trials based on frequentist and Bayesian approaches. Knowledge and learning Upon completion of the course, students know the characteristics of the main experimental designs used in clinical studies. They are able to implement the fundamental procedures for the analysis of randomized clinical trials using different data (normal data, binary data, count responses and survival data). They know how to critically interpret the results, comparing the potentialities of the frequentist and the Bayesian approaches. Furthermore, they are able to apply suitable procedures for the determination of the optimal sample size based on different criteria. Applying knowledge and learning Upon completion of the course, students are able to apply the knowledge acquired, also through the use of the software R, in order to design clinical studies and to interpret the bresults of the statistical analyses Judgement skills Through their study and the various practical applications shown during the lessons, students develop the critical thinking that derives from the comparison between methodologies of analysis and design based on different approaches. Moreover, they acquire autonomous judgment skills that allow to identify the most appropriate methods to analyze clinical data, appropriately select the number of patients to enroll and choose the most suitable monitoring strategies. Communication skills Students, through the study and the performance of practical exercises, acquire the technical-scientific language of the discipline, that must be properly used both in the written and in the oral examinations. Learning skills Students who pass the exam have investigated issues concerning the application of inferential methods for the analysis and design of the clinical studies by exploiting both the frequentist and the Bayesian approach. This allows the development of autonomous skills to compare different procedures. Moreover, the group activities, which involve the use of the R software, develop communication abilities and enhance programming skills that will surely be useful to the students during the preparation of their final dissertation.

Channel 1
VALERIA SAMBUCINI Lecturers' profile

Program - Frequency - Exams

Course program
This course will introduce the main statistical techniques for the design and the analysis of clinical trials based on frequentist and Bayesian approaches. More specifically the course will focus on the following topics. * Introduction to experimental studies (experimental and sub-experimental factors, protocol, primary and secondary endpoints). Experimental designs and biases. Randomization, blinding and placebos. Protocol violations: Intention-To-Treat e Per-Protol analyses. Phases of clinical trials. * Analysis of randomized clinical trials with (i) normal data; (ii) binary data (inferential procedures for risk differences, relative risks and odds-ratios); (iii) count responses (inferential procedures for rate ratios). Survival analysis: censored data, survival function, hazard function and cumulative hazard function. Inferential procedures for censored survival data. * Bayesian analysis with normal distributions. Conjugate analysis for binary data. Exact posterior distribution for odds-ratios.. The "credibilty" of significant trial results. Predictions for normal and binary data. Choice of the prior distribution: methods (i) to elicit experts' opinions, (ii) to obtain non-informative, skeptical and enthusiastic priors and (iii) to summarize historical evidence (Power Priors). * Sample size determination (SSD). SSD methods based on precision analysis: frequentist and Bayesian criteria (Average Coverage Criterion, Average Length Criterion, Worst Outcome Criterion, Length Probability Criterion). SSD methods based on power analysis: conditional approach. Applications when testing for equality, superiority, non-inferiority and equivalence (Two-One Sided Test). SSD methods based on Bayesian power functions (two-priors approach). Two-stage designs for phase II clinical trials: Simon's designs (Optimal and Minimax), Bayesian Single Threshold Design and predictive versions. Monitoring of sequential trials (frequentist and Bayesian designs).
Prerequisites
Knowledge of the classical procedures of statistical inference. Basic knowledge of inferential methods based on a Bayesian approach. Basic knowledge of the use of the software R .
Books
* D.J. Spiegelhalter, K.R. Abrams, J.P. Myles (2004). Bayesian Approaches to Clinical Trials and Health-Care Evaluation. Wiley. Teaching material provided by the teacher : theoretical supplementary notes, slides of the lessons and script with practical applications with R software (https://elearning.uniroma1.it/).
Teaching mode
Lectures will focus on theoretical aspects and practical applications, that will be also presented through the use of a statistical software.
Frequency
It is strongly advised to attend all lectures of this course. Students who are not able to attend the lectures, are invited to contact the lecturer in advance.
Exam mode
The exam consists of two compulsory parts: (a) a final written examination, that contains theoretical questions and practical exercises; (b) an oral examination, that can only be taken once the student has passed the written test. Written and oral examinations are aimed at ascertaining both the acquisition of theoretical concepts and the ability to solve concrete problems.
Bibliography
* P. Armitage, G. Berry, J.N.S Matthews (2002). Statistical Methods in Medical Researh, fourth edn. Blackwell Scientific Publications, Oxford. * A. Bacchieri, G. Della Cioppa (2004). Fondamenti di ricerca clinica. Springer-Verlag Italia, Milano. * J.N.S. Matthews (2006). Introduction to Randomized Controlled Clinical Trials. 2ed., Chapman and Hall/CRC, Boca Raton, FL. * J.M. Lachin (2009). Biostatistical Methods: The Assessment of Relative Risks, 2nd Edition, Wiley
Lesson mode
Lectures will focus on theoretical aspects and practical applications, that will be also presented through the use of a statistical software.
  • Lesson code1044607
  • Academic year2024/2025
  • CourseStatistical Sciences
  • CurriculumBiostatistica
  • Year2nd year
  • Semester1st semester
  • SSDSECS-S/01
  • CFU9
  • Subject areaStatistico