Innovative Approaches to Teaching Survival Analysis for Learners of All Levels

Effective teaching of applied survival analysis requires ongoing adaptation to meet changing student needs and leverage evolving tools. Drawing on over a dozen years teaching epidemiology doctoral students, I will discuss how my approach has evolved in software tools, teaching methods, and the integration of theory with practice, while maintaining focus on core principles of rigorous applied analysis. This talk will cover key aspects of this evolution, including the transition from Stata to R/Quarto and strategies for making complex concepts accessible to epidemiologists, such as reframing competing risks methods through the familiar lens of sensitivity analysis. Most recently, generative AI has emerged as both an opportunity and challenge, causing me to rethink what I need to teach and how I need to teach it. I have maintained strong integration of didactic principles with hands-on application, a balance that remains critical as students gain access to AI-generated code. 

The most common questions in cancer research relate to disease survival. As cancer biostatisticians, we regularly conduct survival analyses as part of applied projects and have to explain the results to our clinical collaborators. While most oncologists have years of experience reading the results of research studies, often they have not received extensive formal training in biostatistics, and survival analysis in particular. As a result it is common to uncover common misunderstandings about the motivation behind survival analyses, and the interpretation of survival analysis results. In this talk, I will discuss some key considerations for teaching survival analyses to clinical collaborators and will show concrete ways to describe the need for survival analysis and explain and interpret the results.

 

Survival analysis is a commonly used statistical method across all medical fields, since improving survival is the ultimate goal for patients with chronic disease or potentially terminal illnesses, such as cancer or heart disease. However, teaching survival analysis can be challenging and innovative approaches are needed to reach learners of all levels. Often, survival analysis is taught to medical researchers with limited statistics knowledge. Teaching survival analysis to these learners need to focus on the basics of survival analysis, while also helping them learn the practicalities of how to do survival analysis for their own research studies. This entails basics on how to properly write-up the statistical methods, and how to properly interpret the results from these studies and other studies they may review. To do this, a traditional lecture-based approach is not the most efficient. For these students, a better approach is to provide real-world examples, hands-on experiences using statistical software, and assignments or projects that provide students the opportunity to run survival analyses, write up the statistical methods and results, and interpret the results. In this talk, I'll discuss the teaching methods and tools I've been using over the last 12 years to help medical researchers learn and interpret survival analysis so they can lead their own high-quality analyses.