The impact of missing data on analyses of a time-dependent exposure in a longitudinal cohort: a simulation study

Abstract

BackgroundMissing data often cause problems in longitudinal cohort studies with repeated follow-up waves. Research in this area has focussed on analyses with missing data in repeated measures of the outcome, from which participants with missing exposure data are typically excluded. We performed a simulation study to compare complete-case analysis with Multiple imputation (MI) for dealing with missing data in an analysis of the association of waist circumference, measured at two waves, and the risk of colorectal cancer (a completely observed outcome).MethodsWe generated 1,000 datasets of 41,476 individuals with values of waist circumference at waves 1 and 2 and times to the events of colorectal cancer and death to resemble the distributions of the data from the Melbourne Collaborative Cohort Study. Three proportions of missing data (15, 30 and 50%) were imposed on waist circumference at wave 2 using three missing data mechanisms: Missing Completely at Random (MCAR), and a realistic and a more extreme covariate-dependent Missing at Random (MAR) scenarios. We assessed the impact of missing data on two epidemiological analyses: 1) the association between change in waist circumference between waves 1 and 2 and the risk of colorectal cancer, adjusted for waist circumference at wave 1; and 2) the association between waist circumference at wave 2 and the risk of colorectal cancer, not adjusted for waist circumference at wave 1.ResultsWe observed very little bias for complete-case analysis or MI under all missing data scenarios, and the resulting coverage of interval estimates was near the nominal 95% level. MI showed gains in precision when waist circumference was included as a strong auxiliary variable in the imputation model.ConclusionsThis simulation study, based on data from a longitudinal cohort study, demonstrates that there is little gain in performing MI compared to a complete-case analysis in the presence of up to 50% missing data for the exposure of interest when the data are MCAR, or missing dependent on covariates. MI will result in some gain in precision if a strong auxiliary variable that is not in the analysis model is included in the imputation model.

Effects of Changes in Number of Medications and Drug Burden Index Exposure on Transitions Between Frailty States and Death: The Concord Health and Ageing in Men Project Cohort Study

“…As such, the model should not be used as an individual prediction or simulation tool. Last, methods of addressing missing data due to missed clinic visits, such as multiple imputation, were not performed because the benefits of multiple imputation were likely to be minimal because the exposures of interest (total number of medications and DBI) would have required imputation …”

Gestational age and school achievement: a population study

“…Multiple imputation by chained equations was performed using Stata/IC 13.1 ‘mi impute chained’ under the assumption data were missing at random. All analysed variables were used in imputation, along with predictors of missingness34 (Australian-born mother, pregnancy procedures, maternal stillbirths, spontaneous onset of labour and discharge from hospital within 28 days following birth (mother and child)). We generated 20 datasets and undertook 50 cycles of regression switching.…”

Using decision trees to understand structure in missing data

“…Common methods of handling missing data, such as complete case analysis, missing indicator method, and last case carried forward have been shown to be acceptable when data is MCAR. 12 13 That being said, most recommendations now are to use multiple imputation, but subject to some care as it only reduces bias from analysis when data are MAR or MCAR; multiple imputation also requires variables that influence missingness to be included in the imputation model. 1–4 14 When data are MNAR, multiple imputation can be used but requires the MNAR mechanism to be known, which is not often undertaken in practice.…”