Clinical Trials - Missing Data

Resize the browser window to see the effect.




Missing Data Analysis in Clinical Trials

taught by Geert Molenberghs

Aim of Course:

Conventional methods for handling missing data in a controlled clinical trial, like complete case analysis, single imputation, and last observation carried forward, waste data, sacrifice power, and can yield biased estimation and unreliable inferences. Much better results can be obtained with the newer but still established methods of direct maximum likelihood, direct Bayesian analysis, inverse probability weighting, and/or multiple imputation, which have become practical in the last few years with the introduction of widely available and user-friendly software. They are broadly valid under the so-called assumption of 'missing at random' (MAR). They apply to continuous data, binary data, categorical data, count data, etc. Furthermore, they are applicable throughout all areas of application, whether in biomedical sciences, economy, psychology, social and behavioral sciences, agriculture, biology, etc. This online course, "Missing Data Analysis in Clinical Trials" will address the issues arising with the conventional methods, and provide a basis for the more promising methods, with focus on maximum likelihood, inverse probability weighting, and multiple imputation. A formal basis will be provided without being overly mathematical. Furthermore, case studies will be discussed and software implementation will be discussed. The issues arising when the MAR assumption is not met are sketched, together with the need for sensitivity analysis.

Course Program:

WEEK 1: Setting the Scene

  • Review of models for continuous hierarchical data
  • Missing-data patterns (monotone, non-monotone)
  • Modeling frameworks (selection models, pattern-mixture models, shared-parameter models)
  • Missing-data mechanisms (missing completely at random, missing at random, missing not at random)
  • The failure of simple methods

WEEK 2: Direct Likelihood Methods

  • Inferential paradigms (likelihood, Bayesian, frequentist)
  • Ignorability
  • The principle for direct likelihood
  • Case studies
  • Software implementation

WEEK 3: Multiple Imputation

  • Rationale for multiple imputation
  • Principles underlying multiple imputation
  • Proper imputation
  • Case studies
  • Software implementation

WEEK 4: Inverse Probability Weighting

  • Review of models for non-continuous hierarchical data
  • Rationale for inverse probability weighting
  • Weighted generalized estimating equations
  • Case studies
  • Software implementation
  • Comments on methods for missing not at random
  • Comments on sensitivity analysis

In addition to assigned readings, this course also has supplemental readings available online, and an end of course data modeling project.


Missing Data Analysis in Clinical Trials

Who Should Take This Course:

Any statistical analyst who works with data from controlled trials is likely to encounter missing observations and will benefit from this course.

Level:

intermediate/advanced

Prerequisite:

You should be familiar with introductory statistics.  Try these self tests to check your knowledge.

To take this course, you should have a good working knowledge of the principles and practice of multiple regression, as well as elementary statistical inference. But you do not need to know matrix algebra, calculus, or likelihood theory.


Organization of the Course:

This course takes place online at the Institute for 4 weeks. During each course week, you participate at times of your own choosing - there are no set times when you must be online. Course participants will be given access to a private discussion board. In class discussions led by the instructor, you can post questions, seek clarification, and interact with your fellow students and the instructor.

At the beginning of each week, you receive the relevant material, in addition to answers to exercises from the previous session. During the week, you are expected to go over the course materials, work through exercises, and submit answers. Discussion among participants is encouraged. The instructor will provide answers and comments, and at the end of the week, you will receive individual feedback on your homework answers.

Time Requirement:
About 15 hours per week, at times of  your choosing.

Credit:
Students come to the Institute for a variety of reasons. As you begin the course, you will be asked to specify your category:

  1. You may be interested only in learning the material presented, and not be concerned with grades or a record of completion.
  2. You may be enrolled in PASS (Programs in Analytics and Statistical Studies) that requires demonstration of proficiency in the subject, in which case your work will be assessed for a grade.
  3. You may require a "Record of Course Completion," along with professional development credit in the form of Continuing Education Units (CEU's).  For those successfully completing the course,  CEU's and a record of course completion will be issued by The Institute, upon request.


Course Text:

The required text for this course is Missing Data in Clinical Studies by Geert Molenberghs and Michael Kenward.

PLEASE ORDER YOUR COPY IN TIME FOR THE COURSE STARTING DATE.

Software:

Hands-on computer assignments are a part of the course. SAS, Stata and R are suitable programs for doing these assignments; the instructor is familiar with SAS and can offer advice; more limited help is available from the TA's for Stata and R.


Missing Data Analysis in Clinical Trials

Instructor(s):
Dates:
To be scheduled.

Course Fee: $589