BIOSTATISTICS
Fall 2021
|
Instructor: |
Guan-Hua Huang, Ph.D. |
|
|
Office: 423 Joint Education Hall |
|
|
Phone: 03-513-1334 |
|
|
Email: ghuang@stat.nctu.edu.tw |
|
Class meetings: |
Monday 9:00 – 12:00 at 406
Joint Education
Hall |
|
Office hours: |
By
appointment |
|
Class website: |
|
|
Credit: |
Three (3) credits |
COURSE
SUMMARY
This course will cover statistical methods for
design, conduct, and analysis of epidemiologic studies such as case-control and
cohort studies. Topics include overview of designs for research studies;
epidemiologic study design; rates and risk; measures of association; classical
contingency table methods; logistic, polytomous logistic, proportional odds and
Poisson regression; and proportional hazard regression. Examples from the
literature and data from a number of well-known epidemiologic studies will be
used for illustration of the concepts involved.
HANDOUTS
AND TEXTBOOKS
Handouts corresponding to each lecture
will be available on the class website before each class. There is no required textbook for this course.
Following books are recommended for further reading:
Nicholas P.
Jewell (2003). Statistics
for Epidemiology.
CRC Press.
Breslow NE, Day NE (1980). Statistical Methods in
Cancer Research, Vol. I, The Analysis of Case-Control Studies. International
Agency for Research on Cancer, Lyon.
Breslow NE, Day NE (1987). Statistical Methods in
Cancer Research, Vol. II, The Design and Analysis of Cohort Studies.
International Agency for Research on Cancer, Lyon.
Hosmer DW, Lemeshow S (1989). Applied Logistic Regression, John Wiley & Sons.
McCullagh P,
Nelder JA (1989). Generalized Linear Models, 2nd edition, Chapman
and Hall.
Klein, Moeschberger (2003). Survival Analysis:
Techniques for Censored and Truncated Data, 2nd edition.
PREREQUISITES
Students are
expected to be familiar with basic epidemiologic concepts.
METHOD
OF STUDENT EVALUATION
The course grade will be based on four homework assignments
(40%),
one midterm exam (25%),
and one final exam (35%).
COURSE OUTLINE
Nicholas P. Jewell
(2003): Statistics for Epidemiology (Jewell),
Breslow NE, Day NE (1980): Statistical Methods in
Cancer Research, Vol. I (Breslow & Day I),
Breslow NE, Day NE (1987): Statistical Methods in
Cancer Research, Vol. II (Breslow & Day II),
Hosmer DW,
Lemeshow S (1989): Applied Logistic Regression (Hosmer & Lemeshow),
McCullagh P, Nelder
JA (1989): Generalized Linear Models, 2nd edition (McCullagh &
Nelder),
Klein, Moeschberger (2003). Survival Analysis:
Techniques for Censored and Truncated Data, 2nd edition. (Klein &
Moeschberger)
|
Module |
Topic |
|
|
1 |
Introduction and overview of
goals of epidemiologic research and epidemiologic study designs |
Jewell Chapter 1, 5 Breslow & Day I Chapter 1 Breslow & Day II Chapter 1 |
|
2 |
Measures of
disease occurrence and association: a.
Different rates b.
Inferences for rates c.
Comparing two rates d.
2×2 table e.
c2 test and Fisher’s exact test f.
R×C table g.
Relative risk and odds ratio h.
Inferences for relative risk and odds ratio |
Jewell Chapters 2,
4, 6, 7 Breslow & Day I Chapter 2, 4 |
|
3 |
Control of
extraneous factors: a.
Confounding and interaction b.
Mantel-Haenszel method c.
Analysis of matched data- McNemar’s test d.
Measure of agreement- Kappa statistic |
Jewell Chapter 8,
9, 10, 11 |
|
4 |
Nonparametric
methods a. Sign test b. Wilcoxon signed-rank test c. Wilcoxon rank-sum test
(Mann-Whitney test) d. Kruskal-Wallis test e. Spearman rank correlation |
|
|
5 |
Multiple
comparison procedures a. Family-wise error rate, false
discovery rate b. Adjusted p-value c. Bonferroni procedure d. Sidak procedure e. Holm procedure f. Hochberg procedure g. Benjamini-Hochberg procedure |
|
|
6 |
Logistic
regression: a.
Concepts b.
Model and assumption c.
Interpretation of regression coefficients d.
Inferences e.
Likelihood ratio test f.
Goodness-of-fit test g.
Residual plot |
Jewell Chapter 12, 13 Hosmer & Lemeshow Chapter 1, 2 |
|
7 |
Confounding
and interaction in logistic regression: a.
Modeling b.
Assumption c.
Interpretation |
Jewel Chapter 14 |
|
8 |
Logistic
regression for contingency tables: a.
Modeling of 2×2 table b.
Modeling of 2×2×2 table c.
Modeling of K×2 table d.
Modeling of M×2×2 table |
No reading |
|
9 |
Logistic regression for
case-control data and conditional logistic regression: a. Case-control studies b. Risk and odds ratio in case-control studies c. Logistic regression for case-control data d. When there are many strata with few observations in each e. Conditional logistic regression for matched data |
Jewell Chapter 16 Hosmer & Lemeshow Chapter 7 |
|
10 |
Analysis of
polytomous data: a. Polytomous data: nominal
versus ordinal scale b. Model for nominal scales
(polytomous logistic regression) c. Model for ordinal scales
(proportional odds model) |
McCullagh
& Nelder Chapter 5 |
|
11 |
Poisson
regression and log-linear model: a. Introduction to generalized
linear model b. Poisson regression c. Application to cohort study d. Log linear model for
contingency table |
Breslow & Day II Chapter 4 |
|
12 |
Analysis of
survival data: a.
Examples of survival data b.
Survival, hazard and cumulative hazard functions c.
Censoring and truncation |
Klein & Moeschberger Chapters 1, 2, 3 |
|
13 |
One-sample
estimation for survival data a. Parametric models for complete
and censored data b. Nonparametric models for
complete and censored data c. Kaplan-Meier estimator |
Klein & Moeschberger Chapters 4, 7 |
|
14 |
Proportional
hazards model: a.
Assumption of the model b.
Testing and estimation c.
Computing programs and interpreting results d.
Time dependent covariates |
Klein & Moeschberger Chapter 8 |
|
15 |
Two-sample
testing for survival data a. Complete failure times b. Censored failure times c. Log-rank test |
Klein & Moeschberger Chapter 7 |