PubH 7450 Updates
PubH 7450 Survival Analysis
Updates
- Final exam: 8:00--10:00 on Dec 17 (Tuesday), in the
same classroom we have been using for this class in the semester.
Note: you can bring 2-page (8 by 11 in) notes and a calculator.
- News on 12/13/13:
I've put your graded HWK8 in the Biostat office (A460 Mayo)
for you to pick up during the work hours.
- Notes: Other topics.
- Week 15 plan: Other topics (notes): recurrent events; penalized Cox regression; No class on Thursday (W is the last day of class).
- Notes:
Chapter 13.
- HWK8
Due by 2pm on Dec 12 (Thursday), turn in at A460 Mayo.
- Notes:
semi-parametric AFT.
- Notes:
Chapter 12.
- Week 14: parametric AFT model (Chapter 12);
semi-parametric AFT (notes);
multivariate survival analysis (Chapter 13).
- Week 13:
Score and Schoenfeld residuals (p.376);
time-dependent coefficient model and Therneau-Grambsch tests
for PH (Therneau and Grambsch 2000, Section 6.2);
influence analysis (11.6);
No class on Thursday: Thanksgiving.
- HWK7
due on Dec 5 (Th) .
- Notes:
Chapter 11.
- Week 12: Cox-Snell residuals (11.2); martingale residals (11.3);
No class on Thursday: read i) the parts on cumulative sum of martingale
residual plots and standardized score process plots
in SAS Proc Phreg (and Proc Genmod); ii) the following paper:
Prentice RL et al. (2007). Low-Fat Dietary Pattern and Cancer Incidence
in the Women's Health Initiative Dietary Modification Randomized Controlled
Trial. Journal of the National Cancer Institute, 99:1534-1543.
download
.
- HWK6
due on Nov 19.
The article for Problem 1 is given below:
DJ Sargent et al. (2005).
Disease-free survival versus overall survival as a primary end point
for adjuvant colon cancer studies: individual patient data from 20,898
patients on 18 randomized trials.
J Clin Oncol. 2005 Dec 1;23(34):8664-70.
download.
It should be free if you download from a computer on campus.
Problem 4:
data.
- Week 11: time-dependent covariates (9.2); stratified PHM (9.3);
left-truncated data (9.4).
- Notes:
Chapter 9.
- Notes:
CLR.
- HWK5
due on Nov 12.
- Week 10:
Estimating survival functions (8.8);
model building (8.7);
conditional logistic reg (notes);
- Week 9: No class on Tuesday
(Topic: sample size calculations; Read Shih 1995);
partial likelihood (8.3, 8.4);
local tests (8.5);
Shih JH (1995). Sample size calculation for complex clinical trials
with survival endpoints.
Controlled Clinical Trials, 16:395-407. Download
paper and
SAS macro %size.
See also SAS Proc Power.
- Notes:
Chapter 8.
- Week 8: exam review;
Regression models (8.1; 2.6);
interpretation of PHM (8.2);
partial likelihood (8.3);
- Week 7:
One-sample tests (7.2);
Mid-term exam on Th (covering Chapters 1--7).
- HWK4
due on Oct 22.
Bland JM and Altman DG (2004).
The logrank test.
BMJ, 328:1073.
download.
- Chocolate consumption and winning a Nobel Prize: a paper published in
NEJM;
download.
- Week 6: K-sample tests (7.1, 7.3);
test for trend (7.4); Stratified tests (7.5);
Renyi-type tests (7.6); other 2-sample tests (7.7);
testing survival differences at a given time point (7.8)
- Notes:
Chapter 7.
- HWK3
due on Oct 10.
Bellis et al. (2007).
Elvis to Eminem: quantifying the price of fame through early mortality of
European and North American rock and pop stars.
Journal of Epidemiology and Community Health, 61:896-901.
download.
- Week 5: Life-table methods (5.4);
estimating hazard function (6.2); estimating excess mortality (6.3);
K-sample tests (7.1, 7.3).
- SAS macro
%ICE
to compute NPMLE for interval censored data.
- Notes:
Chapters 5 and 6.
- HWK2
due on Oct 1.
- Week 4:
estimators of the mean and median times (4.5); estimates for
left-truncated and right-censored data (4.6);
arbitrarily censored or/and
truncated data (5.2-5.3, Turnbull 1976);
- Week 3:
K-M estimator of a survival function
(4.1-4.2): its form; four theoretical justifications: reduced-sample
approach, redistribution-to-the-right algorithm, self-consistency,
and NPMLE; Greenwood's formula.
Other estimators of a survival function and cumulative hazard function
(4.2).
Point-wise confidence intervals (4.3);
confidence bands (4.4).
- Notes:
Chapter 4.
- Agresti's slides on the topic.
- Two other related articles:
1. Agresti A and Coull B (1998),
Approximate Is Better than "Exact" for Interval Estimation of Binomial
Proportions.
The American Statistician, Vol. 2, No. 2, pp. 119-126.
download.
2. Pan W (2002),
Approximate confidence intervals for one proportion and difference of
two proportions.
Computational Statistics and Data Analysis, Vol 40, Issue 1, pp. 143-157. Download its
reprint or
preprint.
- The article on coverage probabilty of the Wald and score CI
is by A Agresti and B Caffo (2000),
Simple and Effective Confidence Intervals for Proportions and Differences
of Proportions Result from Adding Two Successes and Two Failures.
The American Statistician, Vol. 54, No. 4, pp. 280-288.
It should be downloadable from a computer on campus.
- HWK1
due data: Sept 19, 2013.
- Notes:
Chapter 2.
- Week 2:
mrl and median life (2.4); parametric models (2.5);
likelihood construction (3.5); review of likelihood theory (Appendix B).
inference on a binomial proportion (Agresti and Coull 1998);
- If you need a Biostatistics computer account to run SAS, email me.
If you have never used UNIX before,
read computer help
with UNIX, prepared by Dr Lynn Eberly.
- R web site.
- Notes:
Chapters 1 and 3.
- Week 1: course intro; types of censorings and truncations (3.1-3.4;
Type II censoring is not required);
example data (chapter 1); survival function (2.2); hazard function (2.3).