Applied Missing Data Analysis in the Health Sciences

Index

• Algorithm
  • computational
  • expectation-maximization (See Expectation-maximization (EM) algorithm)
  • Gibbs sampling
  • numerical optimization
  • quasi-Newton
• As-treated methods
• Auxiliary data
• Average sample estimate (efficiency)

• Bahadur model
• Bayesian methods
  • analysis for generalized linear models
    • with nonignorably missing covariates
  • approximate Bayesian bootstrap (ABB) hot deck imputation
  • computation for missing data
    • code used in
    • IMPACT study
    • implemented for Windows/Linux systems/JAGS
    • imputation
    • NHANES study
    • posterior
    • simulation example
  • ignorable missingness
  • inference
  • joint model
  • parametric Bayesian simulation methods
  • proper multiple imputation by
• bias
  • coefficient
  • complete-case data
  • estimator
  • MI
  • ordinary GEE
  • parametric model
  • pseudo-unbiased sample generated from
  • results of CC
  • for small selection probability

• CACE. See Complier average causal effect (CACE)
• Clinical trials
  • with a crossover design
  • factors influencing
  • limiting missing data in designing, strategies for
    • improve data collection
    • improve intervention
    • improve study population selection
  • limiting missing data, strategies in conducting
    • data tracking
    • patient experience
    • training
  • minimize impact of missing data
  • randomized (See Randomized clinical trial)
  • setting, factors, influential
• Complete-case method
• Complier average causal effect (CACE)
• Covariate-dependent missingness (CDM)
• Cox proportional hazards model
• Cox regression
  • with completely observed covariates
  • missing data
  • regression coefficients ...