Fall 2008
TA: Siyoen Kil
Office hours: Tuesday Thursday 11:00 to 1:00
Office: CH 320
Phone: 614-292-7538

e-mail: siyoen@stat.osu.edu


Current Index to Statistics (CIS)

Public computer labs

Clicking on a data hypertext will bring up the data set, if available.

Week of

Homework

Links


9/21/08

  1. Review the Law of Large Numbers and the concepts of probability and expectation.
  2. Read the ICH Guidance E8.
  3. Read the Celebrex "label".
  4. Read the definition of ACR20.


9/28/08

  1. Read the ICH Guidance E9.
  2. Read the EMEA 2001 guidance on AVERAGE bioequivalence, as well as the FDA's January 2001 guidance on population and indiviaul bioequivalence.
  3. Pretend that the data in Dong et al (Journal of the American Medical Association, Volume 277(15), 16 April 1997, pp 1205-1213) came from a balanced one-way design with a sample size of six for each of the four products. Assume that the errors are independently, identically normally distributed. Use the summary statistics in Table 3 for 0.1 mg preparation to compute the 95% bioequivalence confidence interval for Cmax of Total Thyroxine for each of six pairs of products.
    • Note 1: PRETEND that the data in Dong et al (1997) have errors that are independently, identically normally distributed (without having to take the log), so that it is reasonable to pool the variance estimates.
    • Note 2: Do not adjust for multiplicity in computing bioequivalence confidence intervals.
    • Note 3: Table 4 in Dong et al (1997) for ratios was obtained by taking log of original data, computing confidence intervals, and then exponentiating. We cannot reproduce confidence intervals as in Table 4 of Dong et al (1997), because we do not have original data. So simply report confidence intervals for differences.
  4. For those interested: the Chinese ICH E9.


10/05/08

  1. October is breast cancer awareness month.
    • Read about MammaPrint.
    • Read about Herceptin.
  2. Download and read Ebbutt and Frith (1998, Statistics in Medicine 17, 1691-1701). Download and read Weda et al (2000 European Journal of Pharmaceutics and Biopharmaceutics 49, 295-302).


10/12/08

  1. Find and read the amended Helsinki Declaration.
  2. Read ICH Guidance E10.
  3. Do Exercise 2 at the end of Chapter 1. (Note the answer involves only parameters and the standard normal distritution. It does not involve any data or random variable or statistical analysis method.)


10/19/08

  1. Watch the DNA from the Beginning animations (at least module 36).
  2. Download and read the Herceptin label.
  3. Download and read Frank Bretz, Jason Hsu, Jose Pinheiro, and Yi Liu (2008). Dose Finding - A Challenge in Statistics Biometrical Journal: 50, 480-504 DOI: 10.1002/bimj.200810438.
  4. Prove directly, using the definition of a confidence interval, that the bioequivalence interval given in class notes is a 1-alpha confidence interval for AUC. (due Tuesday)
  5. Read ICH Guidance E4.


10/26/08

  1. No class on 4 November, 2009 (election day).
  2. Midterm on Thursday is open-notes.


11/02/08

  1. Read section 1.6 of lecture notes.


11/09/08

  1. Apply 1-sided Dunnett's method to the Ruberg data, comparing each dosed-group with the placebo (dose 0), at the 95% confidence level.

    SAS codes could include:
    proc glm;
    class dose;
    model response = dose;
    LSmeans dose/pdiff=controlU("0") adjust = dunnett CL alpha=.05;

    MINITAB commands could include:
    ANOVA
    General Linear Model
    Comparisons
    Dunnett
    Greater than
  2. You can download some useful articles here.


11/16/08

  1. Apply 1-sided Dunnett's method to the Seroquel data, comparing each dosed-group with the placebo (dose 0), at the 99% confidence level. Suppose the Statistical Analysis Plan (SAP) specifies the doses are to be tested in the order of dose 4, dose 3, dose 5, dose 2, dose 6. (There are typos in Steps 2-5 on p. 490 of Bretz, Hsu, Pinheiro, Liu 2008!)


11/23/08

  1. Do exercise 1 at the end of Chapter 5.


11/30/08

  1. Download Huang, Yifan, and Jason C. Hsu (2007). Hochberg's step-up method: cutting corners off Holm's step-down method. Biometrika 94, doi: 10.1093/biomet/asm067.
  2. Final Exam is Monday 12/08/08 9:30 - 11:18 AM (note: date posted earlier was incorrect)
  3. Attend seminar by Yi Liu, 3:30 p.m. Thursday 2/5/09 in EA 170, to learn method of analyzing primary-secondary endpoints.