Course: Survival Analysis in S-PLUS 20-21 April, UK by Terry Therneau

Newsman (newsman@statsci.co.uk)
Wed, 21 Jan 1998 12:12:22 -0000


SURVIVAL ANALYSIS IN S-PLUS

Presented by=20

Dr. Terry Therneau

Author of both SAS and S-PLUS=20
Survival Analysis Routines

Hosted by StatSci Europe

Oxford 20-21 April 1998

In the last 2-3 years the survival analysis software available has =
experienced a major increment in functionality, and is no longer limited =
to the triad of Kaplan-Meier curves, logrank tests and simple Cox =
models. This course will give an overview of modern survival analysis =
methods using S-PLUS. The written materials will go into considerable =
detail on several of the topics, and also include SAS code for many of =
the examples, particularly those dealing with extending the Cox model.

Dr. Terry Therneau has worked in medical research statistics for over 10 =
years, largely in the areas where survival analysis plays a fundamental =
role. He is currently head of the Section of Biostatistics at the Mayo =
Clinic. He has written many papers on the use of residuals with the Cox =
model. Dr Therneau has given numerous presentations at ASA, ENAR and =
Interface meetings. He is also the author of the S survival routines =
incorporated into S-PLUS, and prior to that several SAS macros =
procedures for survival (COXREG & SURTEST).

Course Format:

The course will consist of both lecture and hands-on instruction. =
Desktop PC's will be provided to all students. Training materials will =
also be provided during the course. These materials will be yours to =
keep.

Cost:

Commercial Delegates (not under maint. with StatSci Europe) 795 =
GBP+VAT=20
Commercial Delegates (under maint. with StatSci Europe) 695 =
GBP+VAT=20
Academic Delegates (not under maintenance with StatSci Europe) 595 =
GBP+VAT=20
Academic Delegates (under maintenance with StatSci Europe) 525 =
GBP+VAT=20

Please find course details and a registration form attached. For further =
details of this and other courses run by StatSci Europe, please contact =
our Training Department: Tel +44 (0)1865 200 952.

Provisional Course Outline

1. INTRODUCTION

Course materials
Course goals
Notes on SAS and S
Overall strengths and weaknesses
Survival routines found in SAS
Survival routines found in S-PLUS

2. USING S-PLUS

Datasets
Arithmetic
Printing
Graphics

3. INTRODUCTION TO SURVIVAL

Right censoring
Survival methods vs. traditional methods
cumulative density - survival curves
t tests long - rank test
regression - Cox model
More general censoring
left and interval censored
counting process data
=09

4. ESTIMATING THE SURVIVAL CURVE

Types of survival estimates
Kaplan-Meier estimates; the survfit() function
Variance estimates
Modified lover limits
Comparing one or more curves; the survdiff() function
Left-censored data
Counting process data

5. REGRESSION WITH SURVIVAL DATA, PARAMETRIC MODELS

Accelerated failure time (AFT) models; the survreg() function
Supported distributions
Residuals and predicted values - generalized linear models

6. REGRESSION WITH SURVIVAL DATA,
THE COX MODEL

Simple Cox models; the coxph() function
stratification variables
log likelihood, score, and Wald tests
handling of ties
infinite coefficients

=20
7. THE COUNTING PROCESS FORMULATION=20
OF A COX MODEL

Mathematics -History
Particular cases
multiple events
time dependent covariates
discontinuous intervals of risk
multiple time scales
Implementation in SAS and S
Stanford heart transplant
Parkinson's disease

8. RESIDUALS FOR COX MODELS

Notation
Martingale and deviance residuals
Score residuals
Schoenfeld residuals
General residuals
What can se test in a Cox model?

9. ESTIMATION OF FUNCTIONAL FORM

Simple martingale residuals
using a null model
using a fitted model
deficiencies -extensions
Improved martingale residuals
relationship to Poisson regression
Examples
primary biliary cirrhosis
Veterans Administration data
Time dependent covariates

10. TIME DEPENDENT COEFFICIENT MODELS

As a test for proportional hazards
Estimation using scaled Schoenfeld residuals
Choice of time scale, relationship to other tests
Examples -simulation
Veterans Administration data

11. LEVERAGE

Influence of a single data point on the estimates
Approximating the influence
Jackknife (sandwich) estimates of variance - relation to other methods
Example
PBC study

12. WEIGHTED DESIGNS AND CASE-COHORT STUDIES

Case weights
frequency vs. Sampling weights
Survey sampling designs
Case-cohort designs

13. MULTIPLE EVENTS

Particular cases
different types of event: PBC UDCA study
multiple occurrence: rhDNase study, bladder cancer study
paired events: diabetic retinopathy study
Methods
time to first event
special time-dependent covariates
random effects models
marginal models
Motivation for the marginal model
Honest variance estimates
relationship to the jackknife and GEE
Case influence for grouped data
doubled data
delayed second event
Model choices - time scales
Strata - time-dependent covariates
counting process form

diabetic retinopathy
survival and progression
bladder cancer study
UDCA study - rhDNase study

14. EXPECTED SURVIVAL

Population rate tables
Individual expected survival vs. cohort survival
Three estimates of cohort survival=20
The survexp function

Registration Form: Survival Analysis using S-PLUS, April 21st - 22nd, =
1998, Oxford.

Registration by Feb 28th is recommended to ensure a place on this =
course.

Commercial Delegates (no maintenance with StatSci Europe) 795 GBP+VAT [ =
]
Commercial Delegates (maintenance with StatSci Europe) 695 GBP+VAT [ =
]
Academic Delegates (no maintenance with StatSci Europe) 595 GBP+VAT [ =
]
Academic Delegates (maintenance with StatSci Europe) 525 GBP+VAT [ =
]

Lunch (optional) =A320 for the two days [ ]
=09
Title........ First Name .................... =
Surname...............................

Organisation.............................................................=

Mailing Address........................................................=20

.......................................................................

.......................................................................

Post Code...................... Country.............................

Tel:........................... Fax: =
.................................

E-mail...................................................................=
....

Note: Payment for this course must be made prior to 1/4/98.

Delete (i) - (iv) leaving one line only:

(i) I enclose a cheque for ................. with this registration =
form.

(ii) Payment will be made by cheque before 1/4/98.

(iii) Please send me an invoice. Payment will be made by 1/4/98.

(iv) Payment will be made by bank transfer by 1/4/98.

Signed......................... Date................................
(if returned by fax/post)=20

Please make cheques payable to Statistical Sciences UK Ltd.

Bank transfers should be made to the following account and include
the name of your institution and the reference No. SA200498

Bank: Lloyds Bank
Branch: Headington Branch
Address: London Road, Headington, Oxford, UK.
Account Name: Statistical Sciences UK Ltd.
Account Number: 0256764
Sorting Code: 30-94-04

Mail Registration To: StatSci Europe, Osney House, Mill Street, =
Oxford OX2 0JX, United Kingdom
Fax Registration To: +44 (0)1865 200 953 =20

Note: Delegates that have not paid course fees in full prior to the =
start of the course will be refused entry to the course.=20