Show simple item record

dc.contributor.authorDe Silva, AP
dc.contributor.authorMoreno-Betancur, M
dc.contributor.authorDe Livera, AM
dc.contributor.authorLee, KJ
dc.contributor.authorSimpson, JA
dc.date.accessioned2020-12-10T00:19:25Z
dc.date.available2020-12-10T00:19:25Z
dc.date.issued2019-01-10
dc.identifierpii: 10.1186/s12874-018-0653-0
dc.identifier.citationDe Silva, A. P., Moreno-Betancur, M., De Livera, A. M., Lee, K. J. & Simpson, J. A. (2019). Multiple imputation methods for handling missing values in a longitudinal categorical variable with restrictions on transitions over time: a simulation study. BMC MEDICAL RESEARCH METHODOLOGY, 19 (1), https://doi.org/10.1186/s12874-018-0653-0.
dc.identifier.issn1471-2288
dc.identifier.urihttp://hdl.handle.net/11343/253459
dc.description.abstractBACKGROUND: Longitudinal categorical variables are sometimes restricted in terms of how individuals transition between categories over time. For example, with a time-dependent measure of smoking categorised as never-smoker, ex-smoker, and current-smoker, current-smokers or ex-smokers cannot transition to a never-smoker at a subsequent wave. These longitudinal variables often contain missing values, however, there is little guidance on whether these restrictions need to be accommodated when using multiple imputation methods. Multiply imputing such missing values, ignoring the restrictions, could lead to implausible transitions. METHODS: We designed a simulation study based on the Longitudinal Study of Australian Children, where the target analysis was the association between (incomplete) maternal smoking and childhood obesity. We set varying proportions of data on maternal smoking to missing completely at random or missing at random. We compared the performance of fully conditional specification with multinomial and ordinal logistic imputation, and predictive mean matching, two-fold fully conditional specification, indicator based imputation under multivariate normal imputation with projected distance-based rounding, and continuous imputation under multivariate normal imputation with calibration, where each of these multiple imputation methods were applied, accounting for the restrictions using a semi-deterministic imputation procedure. RESULTS: Overall, we observed reduced bias when applying multiple imputation methods with restrictions, and fully conditional specification with predictive mean matching performed the best. Applying fully conditional specification and two-fold fully conditional specification for imputing nominal variables based on multinomial logistic regression had severe convergence issues. Both imputation methods under multivariate normal imputation produced biased estimates when restrictions were not accommodated, however, we observed substantial reductions in bias when restrictions were applied with continuous imputation under multivariate normal imputation with calibration. CONCLUSION: In a similar longitudinal setting we recommend the use of fully conditional specification with predictive mean matching, with restrictions applied during the imputation stage.
dc.languageEnglish
dc.publisherBMC
dc.rights.urihttps://creativecommons.org/licenses/by/4.0
dc.titleMultiple imputation methods for handling missing values in a longitudinal categorical variable with restrictions on transitions over time: a simulation study
dc.typeJournal Article
dc.identifier.doi10.1186/s12874-018-0653-0
melbourne.affiliation.departmentMelbourne School of Population and Global Health
melbourne.affiliation.departmentPaediatrics (RCH)
melbourne.source.titleBMC Medical Research Methodology
melbourne.source.volume19
melbourne.source.issue1
melbourne.identifier.nhmrc1104975
dc.rights.licenseCC BY
melbourne.elementsid1365430
melbourne.contributor.authorDe Silva, Anurika
melbourne.contributor.authorMoreno-Betancur, Margarita
melbourne.contributor.authorde Livera, Alysha
melbourne.contributor.authorLee, Katherine
melbourne.contributor.authorSimpson, Julie
dc.identifier.eissn1471-2288
melbourne.identifier.fundernameidNHMRC, 1104975
melbourne.accessrightsOpen Access


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record