Multiple imputation for missing data in a longitudinal cohort study: a tutorial based on a detailed case study involving imputation of missing outcome data

Katherine J. Lee, Gehan Roberts, Lex W. Doyle, Peter J. Anderson, John B. Carlin

Research output: Contribution to journalArticleOtherpeer-review

Abstract

Multiple imputation (MI), a two-stage process whereby missing data are imputed multiple times and the resulting estimates of the parameter(s) of interest are combined across the completed datasets, is becoming increasingly popular for handling missing data. However, MI can result in biased inference if not carried out appropriately or if the underlying assumptions are not justifiable. Despite this, there remains a scarcity of guidelines for carrying out MI. In this paper we provide a tutorial on the main issues involved in employing MI, as well as highlighting some common pitfalls and misconceptions, and areas requiring further development. When contemplating using MI we must first consider whether it is likely to offer gains (reduced bias or increased precision) over alternative methods of analysis. Once it has been decided to use MI, there are a number of decisions that must be made during the imputation process; we discuss the extent to which these decisions can be guided by the current literature. Finally we highlight the importance of checking the fit of the imputation model. This process is illustrated using a case study in which we impute missing outcome data in a five-wave longitudinal study that compared extremely preterm individuals with term-born controls.

Original languageEnglish
Pages (from-to)575-591
Number of pages17
JournalInternational Journal of Social Research Methodology
Volume19
Issue number5
DOIs
Publication statusPublished - 2 Sep 2016
Externally publishedYes

Keywords

  • cohort study
  • longitudinal analysis
  • longitudinal study
  • missing at random
  • Missing data
  • multiple imputation

Cite this

@article{05232e6f16344018b7e7dd59465116c0,
title = "Multiple imputation for missing data in a longitudinal cohort study: a tutorial based on a detailed case study involving imputation of missing outcome data",
abstract = "Multiple imputation (MI), a two-stage process whereby missing data are imputed multiple times and the resulting estimates of the parameter(s) of interest are combined across the completed datasets, is becoming increasingly popular for handling missing data. However, MI can result in biased inference if not carried out appropriately or if the underlying assumptions are not justifiable. Despite this, there remains a scarcity of guidelines for carrying out MI. In this paper we provide a tutorial on the main issues involved in employing MI, as well as highlighting some common pitfalls and misconceptions, and areas requiring further development. When contemplating using MI we must first consider whether it is likely to offer gains (reduced bias or increased precision) over alternative methods of analysis. Once it has been decided to use MI, there are a number of decisions that must be made during the imputation process; we discuss the extent to which these decisions can be guided by the current literature. Finally we highlight the importance of checking the fit of the imputation model. This process is illustrated using a case study in which we impute missing outcome data in a five-wave longitudinal study that compared extremely preterm individuals with term-born controls.",
keywords = "cohort study, longitudinal analysis, longitudinal study, missing at random, Missing data, multiple imputation",
author = "Lee, {Katherine J.} and Gehan Roberts and Doyle, {Lex W.} and Anderson, {Peter J.} and Carlin, {John B.}",
year = "2016",
month = "9",
day = "2",
doi = "10.1080/13645579.2015.1126486",
language = "English",
volume = "19",
pages = "575--591",
journal = "International Journal of Social Research Methodology",
issn = "1364-5579",
publisher = "Taylor & Francis",
number = "5",

}

Multiple imputation for missing data in a longitudinal cohort study : a tutorial based on a detailed case study involving imputation of missing outcome data. / Lee, Katherine J.; Roberts, Gehan; Doyle, Lex W.; Anderson, Peter J.; Carlin, John B.

In: International Journal of Social Research Methodology, Vol. 19, No. 5, 02.09.2016, p. 575-591.

Research output: Contribution to journalArticleOtherpeer-review

TY - JOUR

T1 - Multiple imputation for missing data in a longitudinal cohort study

T2 - a tutorial based on a detailed case study involving imputation of missing outcome data

AU - Lee, Katherine J.

AU - Roberts, Gehan

AU - Doyle, Lex W.

AU - Anderson, Peter J.

AU - Carlin, John B.

PY - 2016/9/2

Y1 - 2016/9/2

N2 - Multiple imputation (MI), a two-stage process whereby missing data are imputed multiple times and the resulting estimates of the parameter(s) of interest are combined across the completed datasets, is becoming increasingly popular for handling missing data. However, MI can result in biased inference if not carried out appropriately or if the underlying assumptions are not justifiable. Despite this, there remains a scarcity of guidelines for carrying out MI. In this paper we provide a tutorial on the main issues involved in employing MI, as well as highlighting some common pitfalls and misconceptions, and areas requiring further development. When contemplating using MI we must first consider whether it is likely to offer gains (reduced bias or increased precision) over alternative methods of analysis. Once it has been decided to use MI, there are a number of decisions that must be made during the imputation process; we discuss the extent to which these decisions can be guided by the current literature. Finally we highlight the importance of checking the fit of the imputation model. This process is illustrated using a case study in which we impute missing outcome data in a five-wave longitudinal study that compared extremely preterm individuals with term-born controls.

AB - Multiple imputation (MI), a two-stage process whereby missing data are imputed multiple times and the resulting estimates of the parameter(s) of interest are combined across the completed datasets, is becoming increasingly popular for handling missing data. However, MI can result in biased inference if not carried out appropriately or if the underlying assumptions are not justifiable. Despite this, there remains a scarcity of guidelines for carrying out MI. In this paper we provide a tutorial on the main issues involved in employing MI, as well as highlighting some common pitfalls and misconceptions, and areas requiring further development. When contemplating using MI we must first consider whether it is likely to offer gains (reduced bias or increased precision) over alternative methods of analysis. Once it has been decided to use MI, there are a number of decisions that must be made during the imputation process; we discuss the extent to which these decisions can be guided by the current literature. Finally we highlight the importance of checking the fit of the imputation model. This process is illustrated using a case study in which we impute missing outcome data in a five-wave longitudinal study that compared extremely preterm individuals with term-born controls.

KW - cohort study

KW - longitudinal analysis

KW - longitudinal study

KW - missing at random

KW - Missing data

KW - multiple imputation

UR - http://www.scopus.com/inward/record.url?scp=84958063275&partnerID=8YFLogxK

U2 - 10.1080/13645579.2015.1126486

DO - 10.1080/13645579.2015.1126486

M3 - Article

VL - 19

SP - 575

EP - 591

JO - International Journal of Social Research Methodology

JF - International Journal of Social Research Methodology

SN - 1364-5579

IS - 5

ER -