EPIDEMIOLOGICAL AND CLINICAL METHODS IN ONCOLOGY
- Overview
- Assessment methods
- Learning objectives
- Contents
- Bibliography
- Teaching methods
- Contacts/Info
None
Students will select a paper from the relvant scientific literature and prepare a short presentation (20-25 minutes) of the paper's methods, results and clinical implications.
The course aims to provide to the students with advanced methodological instruments for reading and interpreting the epidemiological and clinical research in oncology. The epidemiological study design class (24 hours) will present commonly adopted study designs, from clinical trials to observational studies. The concepts of bias and confounding will be widely discussed. Finally, the students will also learn how to make a systematic literature search. The statistics in clinical research class (24 hours) will focus on advanced methods for statistical analyses of clinical trials and observational studies, including parametric and non-parametric inference, linear and generalized regression models, time to event modelling and basics of meta-analyses.
Epidemiological study design:
1. Introduction: Accumulations of evidences in biomedical studies: from case-reports to metaanalysis; the gold standard: controlled, randomized, blind clinical trials; bias and confounding in biomedical studies; an introduction to observational studies
2. Observational descriptive and analytical studies (1): cross-sectional study; cohort study
3. Observational descriptive and analytical studies (2): case-control study; absolute and relative
measures, attributable fraction; causation criteria
4. Source of error: random error and bias; confounding and effect modification; reverse causation
5. Experimental studies: end-points, randomization, blinding, intention-to-treat, crossover design,
ethics
6. Secondary studies: systematic reviews and meta-analysis; methods for bibliographic search
(exercise)
Statistics in Clinical Research.
Lesson 1. Introduction: fundamentals of statistics in clinical research
- Brief summary of descriptive statistic, casual variable, and population distributions
- Inference in biostatistics: null and alternative hypothesis, type I and II errors, sample distributions
- Basic interference on sample means: Z test, paired and unpaired t test
- Basic interference on sample proportion: Z test, chi square test
Lesson 2. Interval estimation: confidence intervals. One-way ANOVA and multiple comparisons. Correlation. Regression: linear and generalized linear models (logistic and Poisson).
Lesson 3. Non parametric statistics for continuous and categorical variables
- One sample: Sign test and Wilcoxon signed rank test. Binomial distribution for one proportion
- Two samples: Wilcoxon rank-sum test. Fisher exact test
- More than two samples: Kruskal-Wallis analysis of variance
Lesson 4. Statistical analysis and confounding:
- ANCOVA
- Mantel-Haenszel chi-square
- Multi-variate regression: linear and generalized linear models
- Example from epidemiological studies: odds ratios
Lesson 5. Statistical analysis of time to event data
- Intro to survival data: censorship. Survival and hazard functions
- Kaplan-Meier methods and log-rank test
- Regression models for time-to-event data: Cox survival models
Lesson 6. Statistical analysis methods for meta-analyses: an introduction
Statistics in clinical research:Stanton A Glantz. Primer of biostatistics, 6th edition.
Edizione italiana: Statistica per le discipline biomediche. Sesta edizione. McGraw-Hill, 2007
N=48 hours of frontal lessons, divided into an epidemiological study design (24 hours) and statistics in clinical research (24 hours) courses. During each class, after a formal presentation of the methodological contents by the teacher, there will be time for reading and discuss concrete exemplifications from the relevant scientific literature.
None