Lectures and seminars Workshop on competing risk, multi-state, and multiple event survival analysis

Main teacher

Professor Terry Therneau, Mayo Clinic, Rochester, Minnesota

Course fee

None (free of charge)

Dates

The workshop will be held on five afternoons, from 13:00-16:00, in lecture room Wargentin at MEB (Nobels väg 12A, KI Solna).

Monday, September 18
Friday, September 22
Monday, September 25
Thursday, September 28
Tuesday, October 3

About the workshop

The workshop will discuss theory and methods for medical research studies that apply statistical methods in the field of survival analysis that extend beyond the Kaplan-Meier method and the Cox model; specifically competing risks, multi-state models, and survival analysis with recurrent events. The focus will be on applications rather than statistical theory. Mayo Clinic is a tertiary care center and as such a large portion of the questions I've worked on are "how long until ..", e.g., time on the liver transplant waiting list, time until recurrence of malignancy or death for cancer patients, time to recovery of function after stroke, progression of brain amyloid and tau, etc; hence the focus on survival analysis methods.  In the last ten years, however, the underlying questions have become increasingly multifactorial.  Patients with fatty liver disease, for instance, are at long term risk for cirrhosis and liver failure, but in the interim metabolic comorbidities associated with obesity such as diabetes, hypertension, and dyslipidemia play a major role in their management; we want to understand how these factors co-evolve over time. This class will primarily focus on a series of extended examples and data sets that have informed my own understanding, with side discussions about statistical methods and software. (Sort of like 'how to build a desk', with asides about linden vs. oak and the merits of dovetail joints.) Questions, interruptions and discussion are strongly encouraged.  

Teaching format

Face-to-face lectures and discussion. There will not be any computing sessions. There will be a heavy focus on presentation and discussion of case studies. All participants are invited, in the application for the workshop, to describe a study of their own (past, ongoing, or planned) that they would be interested in discussing. Professor Therneau has committed to discussing these studies with every participant. Some participants will be invited to present and discuss their study with the class. There will also be some guest lectures.

Target audience and prerequisites

The workshop is open to everyone with a strong interest in medical research along with experience in time-to-event analyses. Experience could be via a course such as the KI doctoral course "Biostatistics 3" or through having experience analysing data using a Cox model. We welcome participants from all sectors (university/government authority/industry) and with any background. We are hoping to get a mix of participants with backgrounds in, for example, biology and medicine (with interest in methods) and those with a background in statistics (with an interest in applications in biomedicine). The R survival package will be used for analysis and illustration, but the focus will be on the results and not the process. The code for all slides and examples will be made available for those who want it. 

About the instructor

Terry Therneau grew up on a dairy farm in southeastern Minnesota. (Milking cows is a great inducement to higher education, i.e., finding a different career.)  By great good fortune, his first job after college was working as a programmer for a biostatistics group, which set him on a lifelong path. A few years later he began graduate studies at Stanford, and joined Mayo Clinic in 1985.  The purpose of the Mayo division of biostatistics is to advance medicine --- no graduate students, classes, or "theory for theory's sake" goals.  Over the years he has provided primary support for lung and genitourinary malignancies, hepatology and transplant, dysproteinemias, physical medicine and rehabilitation, and most recently, aging and dementia research.  Statistically, he is perhaps most well known as the author of the R survival package; however, every feature of that package was added in direct response to a specific data set and analysis question.  

Examination and certificate

There is no examination, no homework, no grades, and no certificate. 

Selection criteria

The workshop will be limited to 20 participants. We will accept the first 20 applicants who fulfill the pre-requisites and provide a suitable motivation as to why the course content is relevant for their work/research. The motivation should include a brief description of a project in which you have been involved, on a topic covered by the course, along with questions or discussion topics arising from the project.  

Applications

The workshop is now full (we have received 40 applications for the 20 places) and are no longer accepting applications. 

Funding

This workshop is organised by the Department of Medical Epidemiology and Biostatistics at Karolinska Institutet and is part-funded by a Research Environment Grant from Vetenskapsrådet (VR-REG 19-00227: Statistical Methods for Register-Based Research, node Karolinska Institutet).