GIS (Geographical Information Systems) in Environmental Epidemiology (4 days course)
8 – 11 July 2024
Dr. Danielle Vienneau and Dr. Kees de Hoogh, Department of Epidemiology and Public Health, SwissTPH, University of Basel, Switzerland
The physical and social environment that surrounds us plays an important part in our health and wellbeing. The geography concept of ‘place’ thus cannot be ignored in environmental epidemiology and public health. Whether investigating the level of environmental pollution, access to recreation or health services, or patterns of disease, Geographic Information Systems (GIS) provide the standard platform for exploring spatial attributes and relationships between our environment and health.
This course offers an introduction to GIS and how it is used in environmental epidemiological research. It will introduce students to the basics including: working with and integrating spatial and non-spatial data; geographic scale and spatial precision; geocoding; visualisation; thematic mapping; and understanding spatial relationships. Specific skills and tools will also be introduced in relation to methods for spatial linkage of exposure, contextual and confounder information for epidemiological or health risk assessment studies.
This course will be a mix of lectures, demonstrations and practical time for hands-on data analysis in ArcGIS and QGIS.
No prior knowledge of GIS is required for this intensive course.
Students will gain knowledge in the fundamentals of GIS for spatial data handling and analysis. By the end of the course, students will
- Understand how GIS can be used to enhance public health and epidemiological research;
- Be able to acquire, add, manipulate, visualise and map spatial data in a GIS; and
- Be able to perform basic spatial analyses in a GIS.
Modern time series methods for public health and epidemiology (5 days course)
8 – 12 July 2024
Dr. Antonio Gasparrini, London School of Hygiene & Tropical Medicine, London, UK, Dr. Ana Maria Vicedo-Cabrera, University of Bern, Bern, Switzerland, and Dr. Francesco Sera, University of Florence, Florence, Italy
Time series analysis is a key but underused tool for epidemiological and public health research. In the last two decades, there has been an intense activity to develop more sophisticated study designs and statistical models for using time series data in health studies, with applications spanning various research areas. For instance, time series methods can now be applied for evaluating public health interventions, for assessing health effects associated with environmental stressors and climate change, and for quantifying beneficial or side effects of drugs or clinical practices.
This course will offer a thorough overview of established approaches and recent advancements in methods using time series data for health research, including a theoretical introduction as well as practical examples in public health, environmental, clinical, cancer, and pharmaco-epidemiology. The sessions will cover standard time series designs for aggregated data, including multi-location studies and recent extensions for small-area and individual-level analysis. Case studies will illustrate the use of novel data resources such as remote sensing satellite measurements, electronic health records, real-time smartphone data, and climate models for health impact projection studies
The course will involve short lectures followed by practical sessions using the statistical software R, to illustrate the use of time series analysis in various settings using real-data examples. While no previous knowledge of time series methods is expected, having basic experience on the use of R for epidemiological analysis is an advantage
Genetic and Epigenetic Epidemiology (5 days course)
8 – 12 July 2024
Dr. David Evans, University of Queensland, Australia, Dr. Gibran Hemani, University of Bristol, Bristol, UK, Dr. Matthew Suderman, University of Bristol, Bristol, UK, and Dr Paul Yousefi, University of Bristol, Bristol, UK
Genetic epidemiology refers to the study of the role of genetic factors in determining health and disease in families and in populations. Genetic epidemiological studies have made substantial contributions to understanding the aetiology of complex traits and diseases, and hold great promise for personalised healthcare in the future. This course provides an introduction to the design, analysis and interpretation of genetic and epigenetic epidemiological studies of disease, with a focus on genome-wide and epigenome-wide association studies (GWAS and EWAS). Topics that will be covered include design and analysis of GWAS, imputation, meta-analysis, bioinformatic follow-up, whole genome and polygenic approaches including G-REML and LD score regression, epigenetics, EWAS, and Mendelian randomization (MR). As well as lectures, participants will gain practical experience in analysing genetic and epigenetic datasets. We will use the R statistical software package for the majority of analyses and participants will get plenty of hands on training in this package. By the end of the course participants should have a good working knowledge of concepts in genetic and epigenetic epidemiology, and will be able to perform analyses of genetic and epigenetic datasets
Perinatal and Early Life Epidemiology (5 days course)
8 – 12 July 2024
Dr. Anne-Marie Nybo Andersen, University of Copenhagen, Denmark, Dr. David Taylor-Robinson, University of Liverpool, UK, and Dr. Katrine Strandberg-Larsen, University of Copenhagen, Demark, Stine Kjær Urhøj, University of Copenhagen, Demark
The Developmental Origin of Health and Disease (DoHAD) approach to human health has put early life into focus. Determinants for health in early life are also determinants for health later in life, and optimal health in early life is the first step in prevention of disease later in life. This is why social disparities in early life health are particularly detrimental and this is why life-course epidemiology begins with perinatal epidemiology.
There are a number of methodological challenges in perinatal and early life epidemiology. These
include often ‘ignored’ epidemiological principles such as defining numerators and denominators, problems with timing of exposures, challenges with repeated yet independent outcomes, and much more.
In this course we will cover theories and methods in epidemiologic DoHAD research, issues related to fertility research and public health, theories and methods in studies of congenital anomalies, fetal and infant mortality, challenges when addressing gestational age at birth and birthweight as outcome and exposure measures. Methods in life-course studies of important diseases in the first decade of life, and childhood adversity and life-course health will be covered.
We will also cover opportunities and challenges in data sources for Perinatal and Early Life Epidemiology, such as birth cohort and register data.
The course will include lectures, practicals, and discussion sessions, but no data analysis elements. There will also be opportunities to discuss the students’ research projects.
Computational Epidemiology (4 days course)
8 – 11 July 2024
Dr. Claus Thorn Ekstrøm, University of Copenhagen, Denmark, Dr. Mikkel Andersen, University of Aalborg, Denmark
Many modern epidemiological research projects employ designs and collect data on a scale and complexity that requires and benefits from advanced, flexible computational methods. This course introduces and explains some of the more general computational methods suitable for analyzing large-scale complex data and experimental designs encountered in epidemiological research. We will cover generic techniques and concepts from machine learning such as data-splitting, bootstrapping, and cross-validation, but will also discuss flexible data-driven parametric and non-parametric modelling approaches for high-dimensional data such as penalized regression, random forests, neural networks and deep learning.
Furthermore, the course will enable the participants to evaluate when these data-driven and computational approaches apply to a given epidemiological data analysis problem, discuss potential pitfalls and understand the assumptions underlying these techniques.
This course will offer a thorough overview of machine learning approaches applicable to epidemiological data analysis and equips the participants with new practical skills to expand their data analysis toolbox. Topics cover bootstrapping, cross-validation, false discovery rates, penalized regression, random forests, boosting, neural networks, deep learning, scoring rules, accuracy metrics and how these methods play together with causal inference. Case studies will be from public health.
The course will involve a mixture of short lectures followed by hands-on sessions using the statistical software R in order to apply the methods introduced. While no previous knowledge of machine learning methods or R is expected, having basic experience using R for epidemiological analysis is an advantage.
Quantitative Bias Analysis for Epidemiologic Research (5 days course)
8 – 12 July 2024
Dr. Matthew Fox, Boston University, Boston, USA, Dr. Rich MacLehose, University of Minnesota, Minnesota, USA
Students of epidemiology are well versed in ways to reduce systematic error (bias) in the design of their studies and to describe random error in the analysis of their studies using confidence intervals. However, other than with measured confounders, students are rarely taught methodologies for quantifying systematic error in their studies. Quantitative bias analysis (QBA) provides a methodology for assessing the impact of bias on study results by making assumptions about the structure of the bias and bias parameters. QBA allows for assessment of both the direction and magnitude of systematic error and gives an estimate of effect (or a series of estimates of effect) that would have occurred had the bias been absent, assuming the bias parameters are correct. Such analyses allow investigators to go beyond qualitative speculation about the bias in the discussion section of manuscripts and can be a powerful tool for quantifying the impact of such biases.
The sessions on QBA will cover simple and multidimensional bias analysis methods that can be used to gain a better understanding of the impact of uncontrolled confounding, selection bias, and misclassification (measurement error) on study results. These methods can be applied to nearly any dataset, even summary data (such as contingency tables) presented in the literature. Such approaches lay the foundation for more complicated methods, but by themselves, they act as if the bias parameters are known with certainty. We will then continue with probabilistic bias analysis, which requires specification of probability distributions about the bias parameters and then uses Monte Carlo simulations methods to create intervals accounting for the uncertainty in the systematic error. Finally, we will finish with methods for combining the systematic error to create simulation intervals that account for the total error (systematic and random) in the study results.