The focus in the course is on the understanding of statistical reasoning in the analysis of epidemiological data and research. The measurement process is discussed in the context of statistical model choice and data management. Regression analysis is a statistical technique used for analyzing the relationship between outcome (dependent variable) and explanatory variables (independent variables). Different types of regression analyses are described and discussed. Initially, linear regression where the outcome variable is normally distributed is discussed. Then some other regression models are introduced and discussed; Logistic regression where the outcome is binary, Poisson regression where the outcome is “counts” and Cox regression where the outcome is “time to event”. Basic concepts in survival analysis are described and discussed e.g. censoring, survival- and hazard function. Basic multilevel modeling will also be introduced. Further, different advanced epidemiological methods to analyze causation such as mediation analysis, propensity score matching, and instrumental variable analysis will also be introduced and discussed. Methods to deal with complicated situations in research are discussed, e.g. bias, confounding and effect modification. Epidemiological research articles will be discussed with focus on the statistical and epidemiological methods. During the course, theoretical parts will be illustrated with discussion exercises and computer exercises with epidemiological data.
Teaching: Web lectures, Lectures, Lessons, Seminars, Computer exercises