Pär-Ola Bendahl, email@example.com, Associate professor, Clinical sciences Lund, department of oncology and pathology (course 1 autumn 2021 and course 2 spring 2022)
Anton Nilsson, Anton.Nilsson@med.lu.se (course 2 autumn 2021 and course 1 and 3, spring 2022).
Lanugage of instruction
PhD-students at the medical faculty with a research project within clinical research. Passed course Applied statistics I is required together with basic knowledge in the statistical package used in the same course.
Course 1. Weeks 12-13
Course 2: Weeks 17-18
Course 3: Weeks 20-21 (home exam can be turned in one day week 22 - (day not set yet), due to holiday in Sweden one day week 21).
The courses will be given in a classroom at BMC, Lund, if possible. If we have to change the format, we will post that information here and we will also give that information in the welcome letter for the course.
Weeks 46-47 in English - the course will be given in Zoom
Weeks 48-49 in English - the course will be given in a classroom at BMC, Lund, if possible depending on the Covid situation. If we have to change the format, we will post that information here and we will also give that information in the welcome letter for the course.
Note that 3 hp amounts to 2 weeks of fulltime study, both regarding credits and the actual extent, and that attendance is compulsory for all course days.
Please note that the two courses spring 2021 will be given digitally.
Content of the course
The course includes the following four blocks:
1. Design and analysis of studies with a focus on binary outcomes
• Cohort and case-control studies
• Incidence and risk measures
• Measures of association: Odds ratio, absolute and relative associations
• Basic survival problems (time-to-event)
2. Introduction to correlation- and regression analysis
• Spearman’s and Pearson’s correlation coefficients
• Simple linear regression
• Simple logistic regression
3. Regression analysis - advanced
• Introduction to Cox regression
• Multivariable modelling
4. Randomized Controlled Trials (RCT) and diagnostic studies
• Design and analysis of randomized controlled trials
• Diagnostic measures: sensitivity, specificity, predictive values
This course for clinical researchers focuses on regression methods, especially linear regression models for a continuous outcome variable and one or more independent variables: in what situations they can be used and how the results should be interpreted.
The course will also cover how regression methods can be used to assess and handle confounding and interaction. Further, other types of regression models such as for binary outcomes (logistic regression) and survival analysis (Cox regression) will be introduced. Additionally, statistical methods for diagnostic tests will be covered by introducing sensitivity, specificity, positive and negative predictive value, as well as choice of cut off with ROC-analysis. The course also introduces theory and regulations around randomized controlled trials.
Kirkwood B and Sterne J. Essential Medical Statistics. Blackwell Science, 2nd edition, 2003. Chapter: 10-16, 18-22, 26-27, 29, 34-38.
Available as e-book at Lund University Libraries (www.lub.lu.se)
The course will be held in English and requires active participation and access to laptop with installed statistical package that you are familiar with from earlier courses (SPSS, STATA or R). More information will be provided before the course.