1 Mohamed Alosh, Ph.D. Kathleen Fritsch, Ph.D. Shiowjen Lee, Ph.D. DBIII, OB, CDER, FDA Efficacy Evaluation in Acne Clinical Trials.

Slides:



Advertisements
Similar presentations
A Spreadsheet for Analysis of Straightforward Controlled Trials
Advertisements

Repeated Measures/Mixed-Model ANOVA:
Fall 2013Biostat 5110 (Biostatistics 511) Discussion Section Week 4 Sandrine Moutou Medical Biometry I.
Four levels of measurement Nominal Ordinal Interval Ratio 4 Nominal: the lowest level 4 Ordinal 4 Interval 4 Ratio: the highest level.
Statistical Analysis and Data Interpretation What is significant for the athlete, the statistician and team doctor? important Will Hopkins
Copyright © Allyn & Bacon (2007) Factorial Designs Graziano and Raulin Research Methods: Chapter 12 This multimedia product and its contents are protected.
STA305 week 31 Assessing Model Adequacy A number of assumptions were made about the model, and these need to be verified in order to use the model for.
Regression Analysis Once a linear relationship is defined, the independent variable can be used to forecast the dependent variable. Y ^ = bo + bX bo is.
HSRP 734: Advanced Statistical Methods July 24, 2008.
Measures of Dispersion
Factorial and Mixed Factor ANOVA and ANCOVA
PSY 307 – Statistics for the Behavioral Sciences Chapter 20 – Tests for Ranked Data, Choosing Statistical Tests.
Copyright © Allyn & Bacon (2010) Factorial Designs Graziano and Raulin Research Methods: Chapter 12 This multimedia product and its contents are protected.
Measures of Dispersion or Measures of Variability
January 7, afternoon session 1 Multi-factor ANOVA and Multiple Regression January 5-9, 2008 Beth Ayers.
Chapter18 Determining and Interpreting Associations Among Variables.
Statistics for the Social Sciences Psychology 340 Fall 2006 Review For Exam 1.
PSYC512: Research Methods PSYC512: Research Methods Lecture 19 Brian P. Dyre University of Idaho.
Lecture 24: Thurs., April 8th
Analysis of Variance & Multivariate Analysis of Variance
Marshall University School of Medicine Department of Biochemistry and Microbiology BMS 617 Lecture 14: Non-parametric tests Marshall University Genomics.
Basic Analysis of Variance and the General Linear Model Psy 420 Andrew Ainsworth.
8/10/2015Slide 1 The relationship between two quantitative variables is pictured with a scatterplot. The dependent variable is plotted on the vertical.
Repeated measures: Approaches to Analysis Peter T. Donnan Professor of Epidemiology and Biostatistics.
ANCOVA Lecture 9 Andrew Ainsworth. What is ANCOVA?
Chapter 1 Exploring Data
Simple Linear Regression
Statistics in psychology Describing and analyzing the data.
S519: Evaluation of Information Systems Week 14: April 7, 2008.
STA291 Statistical Methods Lecture 31. Analyzing a Design in One Factor – The One-Way Analysis of Variance Consider an experiment with a single factor.
230 Jeopardy Unit 4 Chi-Square Repeated- Measures ANOVA Factorial Design Factorial ANOVA Correlation $100 $200$200 $300 $500 $400 $300 $400 $300 $400 $500.
Measures of Central Tendency or Measures of Location or Measures of Averages.
Choosing and using statistics to test ecological hypotheses
Statistical Power 1. First: Effect Size The size of the distance between two means in standardized units (not inferential). A measure of the impact of.
Educational Research: Competencies for Analysis and Application, 9 th edition. Gay, Mills, & Airasian © 2009 Pearson Education, Inc. All rights reserved.
Biostatistics Case Studies 2007 Peter D. Christenson Biostatistician Session 3: Incomplete Data in Longitudinal Studies.
RESULTS & DATA ANALYSIS. Descriptive Statistics  Descriptive (describe)  Frequencies  Percents  Measures of Central Tendency mean median mode.
Biostatistics Case Studies 2008 Peter D. Christenson Biostatistician Session 5: Choices for Longitudinal Data Analysis.
Linear correlation and linear regression + summary of tests
1 OTC-TFM Monograph: Statistical Issues of Study Design and Analyses Thamban Valappil, Ph.D. Mathematical Statistician OPSS/OB/DBIII Nonprescription Drugs.
© 2014 by Pearson Higher Education, Inc Upper Saddle River, New Jersey All Rights Reserved HLTH 300 Biostatistics for Public Health Practice, Raul.
Examination of Analysis Methods for Positive Continuous Dependent Variables: Model Fit and Cost Saving Implications Brian P SmithMaria De Yoreo Biostatistics.
Descriptive Statistics The goal of descriptive statistics is to summarize a collection of data in a clear and understandable way.
Analysis of Covariance vs. Analysis of Difference Scores
Unit 4: Describing Data After 8 long weeks, we have finally finished Unit 3: Linear & Exponential Functions. Now on to Unit 4 which will last 3 weeks.
1 Chapter 4 Numerical Methods for Describing Data.
BPS - 5th Ed. Chapter 251 Nonparametric Tests. BPS - 5th Ed. Chapter 252 Inference Methods So Far u Variables have had Normal distributions. u In practice,
Chapter Eight: Quantitative Methods
 An exposure-response (E-R) analysis in oncology aims at describing the relationship between drug exposure and survival and in addition aims at comparing.
Numerical descriptions of distributions
Session 6: Other Analysis Issues In this session, we consider various analysis issues that occur in practice: Incomplete Data: –Subjects drop-out, do not.
Chapter 21prepared by Elizabeth Bauer, Ph.D. 1 Ranking Data –Sometimes your data is ordinal level –We can put people in order and assign them ranks Common.
Biostatistics Case Studies Peter D. Christenson Biostatistician Session 3: Missing Data in Longitudinal Studies.
Marshall University School of Medicine Department of Biochemistry and Microbiology BMS 617 Lecture 7 – Non-normality and outliers.
Approaches to quantitative data analysis Lara Traeger, PhD Methods in Supportive Oncology Research.
PSY 325 AID Education Expert/psy325aid.com FOR MORE CLASSES VISIT
1 Clinical Studies Section of Labeling Joseph Porres, M.D., Ph.D. Medical Officer Division of Dermatologic and Dental Drug Products FDA.
Simulation setup Model parameters for simulations were tuned using repeated measurement data from multiple in-house completed studies and baseline data.
Repeated measures: Approaches to Analysis
Statistics in psychology
12 Inferential Analysis.
Numerical Descriptive Measures
Joanna Romaniuk Quanticate, Warsaw, Poland
Quartile Measures DCOVA
An Introductory Tutorial
12 Inferential Analysis.
Graziano and Raulin Research Methods: Chapter 12
Two randomized, double-blind, controlled trials of 2219 subjects to compare the combination clindamycin/tretinoin hydrogel with each agent alone and vehicle.
Presentation transcript:

1 Mohamed Alosh, Ph.D. Kathleen Fritsch, Ph.D. Shiowjen Lee, Ph.D. DBIII, OB, CDER, FDA Efficacy Evaluation in Acne Clinical Trials

2 An Outline 1.Re-visit choice of primary endpoints (statistical viewpoint) 2.Statistical analysis methods and data transformations 3.Repeated measurements vs. final assessment 4. Effect of baseline severity 5. Final Comments

3 1. Primary Endpoints in Acne Trials (Statistical Viewpoint) §Inflammatory, non-inflammatory, and Total Lesion Counts can be analyzed as l final lesion counts l change from baseline l percent change §Investigator Global Evaluation (IGE)

4 Analysis of Change Scores Pros l Easy to interpret & analyze l Attempts to remove influence of baseline counts Cons l Baseline may still have influence since change is negatively correlated with final counts l Change and percent change scores may have highly skewed distributions (violates parametric tests)

5 Figure 1. Mean Lesion Counts by Type over Time (Drug X, Study 1)

6 Figure 2. Mean Lesion Counts by Type over Time (Drug Y, Study 2)

7 Figure 3. Inflammatory Counts at Week 12 vs. Baseline Count (Drug X, Vehicle Arm)

8 Figure 4. Inflammatory Counts at Week 12 vs. Baseline Count (Drug X, Active Arm)

9 Figure 5. Mean % Change in Infl. Lesions over Time (Drug X, Study 1)

10 Figure 6. Subject’s Total Lesion Count over Time -- Center A (Drug X, Study 1)

11 2. Statistical Analysis Issues (a) Analysis units l original l transformed data (ranks, log, etc.) (Pros & Cons) (b) Analysis methods for final study endpoint l Simple comparisons l ANOVA (treatment, center, interaction) l ANCOVA (include baseline count as covariate) (Comparison of results presented in Tables 1a-c & Tables 2a-c)

12 3. Repeated Measures vs. Final Assessment §Repeated measures may increase power for detecting treatment effect §Must select number of timepoints to be included in repeated measurements model §Repeated Measures Models: MANOVA, GLM, MIXED (Comparison of results presented in Tables 1a-c & Tables 2a-c))

13 Table 1a: Treatment effect p-values for various statistical methods (Drug X, Study 1) *(active, vehicle)(R) = Repeated Measures

14 Table 1b: Treatment effect p-values for various statistical methods (Drug X, Study 1) *(active, vehicle)(R) = Repeated Measures

15 Table 1c: Treatment effect p-values for various statistical methods (Drug X, Study 1) *(active, vehicle)(R) = Repeated Measures

16 Comments on Results for Drug X §results for total are similar to non-infl. (strong corr., highly signif.) §no general pattern for ranks vs. original data §for infl. lesions, % change has smaller p-values than counts or change §for change and % change, ANCOVA has similar results to Week 12 analysis §p-values for rep. meas. in general are larger than final study endpoint analysis (prev. weeks not signif.)

17 Table 2a: Treatment effect p-values for various statistical methods (Drug Y, Study 2) *(active, vehicle)(R) = Repeated Measures

18 Table 2b: Treatment effect p-values for various statistical methods (Drug Y, Study 2) *(active, vehicle)(R) = Repeated Measures

19 Table 2c: Treatment effect p-values for various statistical methods (Drug Y, Study 2) *(active, vehicle)(R) = Repeated Measures

20 Comments on Results for Drug Y §results for total & non-infl. are similar, as for X (but here less signif.) §no general pattern for ranks vs. original data §for infl. lesions, % change has larger p-values than counts or change §for change and % change, ANCOVA has similar results to Cycle 6 analysis §p-values for GLM(R)/ANCOVA(R) in general are smaller than final study endpoint

21 4. Effect of Baseline Severity §Divide subjects into equal sized groups (e.g., quartiles) based on baseline lesion count Plots of lesion counts by baseline category §Compare efficacy results by baseline category l Tables 3a-b present results for lesion counts l Tables 4a-b present results for IGE

22 Figure 7. Mean Week 12 Lesion Counts by Type over Baseline Category (Drug X, Study 1)

23 Figure 8. Mean Cycle 6 Lesion Counts by Type over Baseline Category (Drug Y, Study 2)

24 Table 3.a. Treatment Effect at Week 12 by Baseline Category (Drug X, Study 1)

25 Table 3.b. Comparison of IGE Success* Rates at Week 12 by Baseline Category (Drug X, Study 1) *Success = None or Minimal Acne

26 Table 4.a. Treatment Effect at Cycle 6 by Baseline Category (Drug Y, Study 2)

27 Table 4.b. Comparison of IGE Success* Rates by Baseline Category (Drug Y, Study 2) *Success = ‘Clear’ or ‘Almost Clear/Mild’

28 Comments about efficacy results by baseline category §For the two drugs considered, no general pattern for results for lesion counts by type, their change, or percent change §Similarly, for the two drugs considered, there is also no general pattern for the IGE §For the range of lesion counts in these studies, efficacy results do not appear to vary by baseline severity

29 5. Final Comments (a) Analysis of change from baseline, percent change, and final counts with baseline as covariate attempt to account for variability at baseline (b) Percent change data could have extreme outliers when the baseline count is relatively small (e.g., inflammatory lesions)

30 Final Comments (cont’d) (c) Repeated measurements approach attempts to reduce the influence of outliers (flares) by ‘averaging’ over time. The impact of the repeated measures on the p- value depends on whether efficacy reached a plateau at previous time points. (d) For the data sets considered, treatment efficacy did not vary by baseline severity whether one considers analysis of lesion counts or IGE