Introduction

This course provides an advanced (second level) examination of the methods of the analysis of data where the outcome is the time to an event. The course will be oriented towards observational studies in the teaching examples and course assignment. However, these methods have strong applications to RCTs. The focus of the course is on the application of the techniques to solve epidemiological problems rather than on the statistical methods and underlying mathematical principles. The course will examine non-regression methods (e.g. Kaplan-Meier and log-rank test) as well as regression modeling (COX models). Students will be expected to undertake computer analysis of data provided as SAS data sets.

This course is an elective course for all students. All students need permission of the instructor prior to registering. The class will normally be limited to a maximum enrolment of about 20 students.

TEACHING PHILOSOPHY

This course will be taught using interactive lectures (i.e. I will present material to the class but will encourage interaction and discussion during the lecture process). The focus of the course will be on practical applications rather than theory. If enrollment is low, consideration will given to modifying some of the lecture time for seminar discussions.

TIMING

Tuesday, 0900-1200 (or 1230). There will be no final examination. For a detailed schedule, click here.

LOCATION

All classes will be taught in room 3001 at the Health Sciences Building (Roger Guindon Hall) at the University of Ottawa (451 Smyth Rd).

INSTRUCTORS

The instructor is Dr. Birkett. Contact information can found here.

PREREQUISTES

Students have to demonstrate prior mastery of basic biostatistics and epidemiology. None of the prerequisite material will be taught during the class. Rather, the class will build on the prerequisite material to demonstrate the application of advanced statistical methods to address core epidemiological issues. The specific prerequisites are:

The official prerequisite is the completion of EPI 5340 (Epidemiological Methods). That course requires prior completion of EPI 5240 and 5242.
Students will be expected to be familiar with the fundamentals of hypothesis testing, confidence intervals, and multiple linear regression analysis modeling. Students will also be expected to be familiar with confounding and effect modification. Prior exposure to the application of the Mantel-Hanzel test for adjusting for confounding would be useful.
Students are encouraged to take EPI 5345 (Applied Logistic Regression) in the first half of the winter semester.
Familiarity with the use of computers for statistical analysis is strongly recommended. The course assignments will require the student to perform computer-based analysis. Data sets will be provided in SAS format.

CORE TEXTBOOK(S)

I am suggesting that students consider buying the following book for this year's class. This book is very practical, focusing on the use of multiple examples with SAS to illustrate various survival analysis methods and tasks. It has just been re-written (2010) and contains up-to-date information on using current SAS procedures. It is a little light on the theory so I will complement it by readings from other sources. The lectures will often provide background to the material in the book. Allison provides digital versions of his data sets and code; I will make these available through the course website. Copies have been ordered at the Health Sciences bookstore.

Allison PD . Survival Analysis Using the SAS System: A Practical Guide (2nd edition). SAS Institute Inc., Cary, 2010.

A second SAS published book provides complementary information to the Allison book. I considered basing the course around this book (as I did a couple of years ago). But, the SAS code is now dated and it makes extensive use of SAS macros to generate results. I can still strongly recommend it, especially since it covers many topics that complement the Allison book. I can make available copies of all of the macros, SAS programmes, etc. used in this book.

Cantor AB. SAS Survival Analysis Techniques for Medical Research. Springer Sciences and SAS Institute, Cary, USA, 2003. ISBN: 978-1-59047-135-7

A third book (by Machin et al) gives a somewhat more advanced presentation of the core material with a more mathematical bent. It includes some topics that are not covered by Allison and Cantor. This would be a good book for people with a deeper interest in theory behind the methods we will be covering.

Machin D, Cheung YB, Parmar MKB. Survival Analysis: A Practical Approach. 2nd edition. John Wiley and Sons, Ltd. 2006. ISBN: 0-470-87040-0

To supplement this book and my suggested readings, additional resources and reading material can always be useful. These are described in the section on books.

SAS (Computer Software)

This course is instructed using SAS. For ease of participation, students are encouraged to use SAS for their own work. However, that is not a requirement - students can use any statistical software they want. Students should be aware that the data set for the second assignment will be provided as a SAS data set, to reduce your coding time. Conversion to non-SAS format will require a special request.

If you want to develop a stronger foundation with SAS, the following two books are very useful. The Delwiche and Slaughter book is not a great reference guide (since it limits discussion to 2 pages on every topic). But, it gives an excellent over-view of most topics in data management, etc. In particular, make sure to look at the ODS material and the discussion on data set manipulations. The Cody book is more detailed and provides a slightly more advanced coverage of material. You can explore these books at SAS (http://support.sas.com/publishing/). In fact, large chunks of these books can be read on-line at no charge.

Delwiche LD, Slaughter SJ. The Little SAS Book: A Primer, 3rd edition. SAS Institute, Cary, NC, 2003.
Cody R. Learning SAS by Example: A Programmer’s Guide. 2nd edition. SAS Institute, Cary, NC, 2007.

Survival Analysis in Epidemiology: Dr. N. Birkett

Survival Analysis in Epidemiology (EPI 5344)