Docstoc

Applied Biostatistics in Clinical Research

Document Sample
Applied Biostatistics in Clinical Research Powered By Docstoc
					Applied Biostatistics in Clinical
          Research
             Required test

• Regression Methods in Biostatistics
  Linear, Logistic, Survival, and Repeated
  Measured Models

• By Eric Vittinghoff, David V. Glidden,
  Stephen C. Chiboski, Charles E.
  McCulloch
       Longitudinal methods

• All of the text’s regression methods work
  with longitudinal data
  (the causative exposures precede the
  outcomes)
         Major course topics
• Linear regression
  Predictive modeling of continuous outcomes

• Examples: height, blood pressure

• Explanatory variables
  Binary, nominal, ordinal, continuous
        Major course topics
• Logistic regression
  Predictive modeling of binary outcomes

• Examples: death, developing lung cancer

• Explanatory variables
  Binary, nominal, ordinal, continuous
        Major course topics
• Survival analysis
  Predictive modeling of time to event data

• Examples: time to death, time to
  developing lung cancer

• Explanatory variables
  Binary, nominal, ordinal, continuous
     Additional Course Topics
• Poisson Regression
• Predictive modeling of count data

• Examples: Number of emergency room
  visits

• Explanatory variables
  Binary, nominal, ordinal, continuous
     Additional Course Topics


• Clinical diagnosis with likelihood ratios
          Omitted text topic
• Regression with repeated measures



• Regression using repeated measures is
  the primary topic of the optional third
  statistics course
  Examples of repeated measures

• Monthly blood pressure measurements

• Single measurements of members of a
  family (repeated measures of the family)
    Types of repeated measures
              analyses
• Patterns of change over time (change
  trajectories)

• Binary, ordinal, or continuous outcomes
  allowed

• Example: Physiological measures by days
  post-intervention
    Types of repeated measures
              analyses
• Multi-level analyses

• Binary, ordinal, or continuous outcomes
  allowed

• Predictors from multiple levels
• Example: Patient, physician, and hospital
  characteristics
     Types of repeated measures
               analyses
• Multi-process analyses

• Multiple outcomes allowed
  (Binary, ordinal, continuous, or time to event)

• Predictors from single or multiple levels
• Examples:
  Changes in height, weight, and blood pressure
  during growth
  Birth in a hospital and infant survival
             Optional text

• The Little SAS Book for Enterprise Guide
  3.0

• By Susan Slaughter and Lora Delwich
          Structure of classes
• About an hour of lecture
  A clinical example, coverage of the text, and text
  examples using SAS Enterprise Guide

• Student presentations of exercises related to the
  lectures

• Student presentations of analyses of their class
  projects

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:2
posted:2/27/2012
language:
pages:15