BEMC Talks are generally held on the first Wednesday of each month at 4pm Berlin time (GMT+01:00) with some adjustments for holidays. An hour-long lecture will be followed by an interactive group discussion with the speaker. Some BEMC Talks are held in a hybrid format (in-person plus live-streaming) and others are held in a digital format only.
Make sure to register in advance to secure your spot. We encourage you to join our BEMC Talks in person on campus whenever possible to engage in post-talk discussions with the speaker and the rest of the audience.
Select BEMC Talks are recorded and posted on our YouTube channel.
2025 BEMC Talks
JANUARY 8, 2025
Statistical Methods for Causal Inference with Time-to-Event Data in Epidemiology
Vanessa Didelez
Universität Bremen, Bremen, Germany
“In this presentation I will address various challenges that arise for typical time-to-event data in epidemiology, including the fact that most treatments or exposures (not just the outcomes) are also time-dependent. First, an overview of methods for time-dependent confounding (g-formula and IPTW) will be given. Then, I will illustrate the importance of the principles of target trial emulation with analyses of health claims data. Specifically, we aimed to investigate the effect of screening colonoscopy on CRC incidence, where conventional observational analyses can easily violate “time-zero alignment” and produce misleading results. Further, competing events occur in these settings, and typical causal estimands are those of total and controlled direct effects. Before presenting an alternative to these estimands, I will give a general idea of causal mediation analyses in longitudinal / time-to-event settings and introduce estimands based on separable treatments. I will then present the novel concept of separable treatment effects for competing events and conclude with a further illustration with a study on prostate cancer.”
On YouTube – watch recording
FEBRUARY-APRIL 2025
No BEMC Talks
MAY 7, 2025
Estimating effects on all-cause mortality in the presence of COVID-19 deaths
Martin Lajous
National Institute of Public Health, Cuernavaca, Mexico; Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland
“In this presentation, I will articulate the challenge that arises for population-based prospective epidemiologic cohorts designed to evaluate the impact of risk factors on non-communicable disease mortality after the SARS-CoV-2 pandemic. Ignoring the presence of COVID-19 deaths could generate uncertainties for the interpretation of treatment effects on all-cause mortality. First, I will introduce the Mexican Teachers’ Cohort and the Polish-Norwegian Study and explore the mortality patterns observed during and after the SARS-CoV-2 pandemic. Second, I will present different analytical approaches in the presence of COVID-19 deaths and discuss the appropriateness of four causal estimands: the total effect, the principal stratum effect, the (controlled) direct effect, and the separable direct effect. Finally, using data from the prospective cohorts introduced in the talk, I will illustrate the estimation of effects to respond to public health questions.”
JUNE 2025
No BEMC Talk
JULY-AUGUST 2025
Summer Break
NEW DATE: OCTOBER 15, 2025
Harnessing the Power of Population-Based Data for Global Health Research
Maja Marcus
Charité – Universitätsmedizin Berlin, Berlin, Germany; Brigham and Women’s Hospital, Boston, US
“Understanding health at a global scale is essential for informing equitable and effective health policies, especially in the face of aging populations, shifting disease burdens, and persistent health inequities. Across the globe, a wealth of population-level health datasets — ranging from national surveys to multinational cohort studies — offers tremendous potential for answering pressing research questions in global health. However, leveraging this potential is often complicated by challenges related to data access, inconsistency in variable definitions, and methodological differences across sources. This talk explores strategies to identify and harmonize various population-level data sources into single, ready-to-use datasets. In this, it will feature the “Global Health and Population Project on Access to Care for Cardiometabolic Diseases (HPACC)” data resource to illustrate not only the kinds of policy-relevant questions such data can help answer, but also the broader impact of harmonized datasets in advancing evidence-based global health.”
In-person registration
Online Zoom registration
NOVEMBER 19, 2025
Interpretational errors in causal inference and how to avoid them
Aaron Sarvet
University of Massachusetts Amherst, Amherst, US
“Pioneering works in causal inference were explicitly grounded in practical disciplines, aiming at formalizing real questions with mathematical definitions. Now, causal inference methods provide an architecture that profoundly regulates what practical questions get asked and how they get answered. Here I consider subtly different approaches for causal inference research, and their implications for theory development and practice. In this process, I formalize an interpretational error that is increasingly apparent in the causal literature, termed “identity slippage”. This formalization can be used for error detection whenever public health and clinical decisions depend on the accurate interpretation of statistical results, which is nearly always the case. Therefore, broad awareness of identity slippage will aid in the successful translation of data into public good. As an illustration, I present case studies of these errors in causal mediation analysis and instrumental variable analysis.”
In-person registration
Online Zoom registration
Talk titles/descriptions are subject to change. Use the links to register in advance. Space is limited!
Berlin Epidemiological Methods Colloquium (BEMC) 