Choice Of Imputation Method Research Articles

Missing Outcome Data in Recent Perinatal and Neonatal Clinical Trials.

Missing outcome data in clinical trials may jeopardize the validity of the trial results and inferences for clinical practice. Although sick and preterm newborns are treated as a captive patient population during their stay in the NICUs, their long-term outcomes are often ascertained after discharge. This greatly increases the risk of attrition. We surveyed recently published perinatal and neonatal randomized trials in 7 high-impact general medical and pediatric journals to review the handling of missing primary outcome data and any choice of imputation methods. Of 87 eligible trials in this survey, 77 (89%) had incomplete primary outcome data. The missing outcome data were not discussed at all in 9 reports (12%). Most study teams restricted their main analysis to participants with complete information for the primary outcome (61 trials; 79%). Only 38 of the 77 teams (49%) performed sensitivity analyses using a variety of imputation methods. We conclude that the handling of missing primary outcome data was frequently inadequate in recent randomized perinatal and neonatal trials. To improve future approaches to missing outcome data, we discuss the strengths and limitations of different imputation methods, the appropriate estimation of sample size, and how to deal with data withdrawal. However, the best strategy to reduce bias from missing outcome data in perinatal and neonatal trials remains prevention. Investigators should anticipate and preempt missing data through careful study design, and closely monitor all incoming primary outcome data for completeness during the conduct of the trial.

How handling missing data may impact conclusions: A comparison of six different imputation methods for categorical questionnaire data.

Objectives:Missing data is a recurrent issue in many fields of medical research, particularly in questionnaires. The aim of this article is to describe and compare six conceptually different multiple imputation methods, alongside the commonly used complete case analysis, and to explore whether the choice of methodology for handling missing data might impact clinical conclusions drawn from a regression model when data are categorical.Methods:In addition to the commonly used complete case analysis, we tested the following six imputation methods: multiple imputation using expectation–maximization with bootstrapping, multiple imputation using multiple correspondence analysis, multiple imputation using latent class analysis, multiple hot deck imputation and multivariate imputation by chained equations with two different model specifications: logistic regression and random forests. The methods are tested on real data from a questionnaire-based study in the Norwegian opioid maintenance treatment programme.Results:All methods performed relatively well when the sample size was large (n = 1000). For a smaller sample size (n = 200), the regression estimates depend heavily on the level of missing. When the amount of missing was ⩾20%, in particular, complete case analysis, hot deck and random forests had biased estimates with too low coverage. Multiple imputation using multiple correspondence analysis had the best performance all over.Conclusion:The choice of missing handling methodology has a significant impact on the clinical interpretation of the accompanying statistical analyses. With missing data, the choice of whether to impute or not, and choice of imputation method, can influence clinical conclusion drawn from a regression model and should therefore be given sufficient consideration.

On imputation for planned missing data in context questionnaires using plausible values: a comparison of three designs

BackgroundThis paper extends a recent study by Kaplan and Su (J Educ Behav Stat 41: 51–80, 2016) examining the problem of matrix sampling of context questionnaire scales with respect to the generation of plausible values of cognitive outcomes in large-scale assessments.MethodsFollowing Weirich et al. (Nested multiple imputation in large-scale assessments. In: Large-scale assessments in education, 2. http://www.largescaleassessmentsineducation.com/content/2/1/9, 2014) we examine single + multiple imputation and multiple + multiple imputation methods using predictive mean matching imputation under three different context questionnaire matrix sampling designs: a two-form design studied by Adams et al. (On the use of rotated context questionnaires in conjunction with multilevel item response models. In: Large-scale assessments in education. http://www.largescaleassessmentsineducation.com/content/1/1/5, 2013), a three-form design implemented in PISA 2012, and a partially-balanced incomplete design studied by Kaplan and Su (J Educ Behav Stat 41: 51–80, 2016).ResultsOur results show that the choice of design has a larger impact on the reduction of bias than the choice of imputation method. Specifically, the three-form design used in PISA 2012 yields considerably less bias compared to the two-form design and the partially balanced incomplete design. We further show that the partially balanced incomplete block design produces less bias than the two-form design despite having the same amount of missing data.ConclusionsWe discuss the results in terms of implications for the design of context questionnaires in large-scale assessments.

OP96 Applying Missing Data Methods to Routine Data: A Prospective Population-Based Register of People with Diabetes

BackgroundResearch using data from large population-based datasets is often hindered by the presence of non-trivial proportions of missing data. Numerous approaches for handling missing data are available, each of which...

Tracking poverty with coarse data: evidence from South Africa

Household surveys often contain coarse data, which consist of a mixture of missing values, interval-censored values and point (fully-observed) values, making it difficult to construct a continuous money-metric measure of wellbeing. This paper assesses the sensitivity of poverty and inequality estimates to the multiple imputation of coarse earnings data and reported zero values using the 2001–2006 South African Labour Force Surveys. Estimates of poverty amongst the employed are shown not to be sensitive to multiple imputation of missing and interval-censored data, but are sensitive to the treatment of workers reporting zero earnings. Poverty trends are generally robust to the choice of method, and a significant decline in poverty is evident. Inequality estimates, on the other hand, appear more sensitive to the treatment of zero values and the choice of imputation methods, and, overall, no particular trends in inequality could be discerned.

Farm Commodity Payments and Imputation in the Agricultural Resource Management Survey

Item non response is a pervasive problem in survey data, and a large literature in statistics has developed to address missing data due to item non response. Government statistical agencies regularly impute for missing values in the microdata that they make available to researchers. Unfortunately, while the imputation methods typically used by government statistical agencies may be suitable for the main objectives of the agency—usually, publishing aggregated descriptive statistics—these imputations are usually unsuitable for multivariate analysis of microdata. Using USDA’s Agricultural Resource Management Survey (ARMS), we apply a new imputation method to missing 2008ARMS data on federal government farm commodity payments and off-farm household income items. We find that the choice of imputation method can have a noticeable effect on important economic indicators, even when the rate of missingness is small.

Methods for handling missing data in palliative care research

Missing data is a common problem in palliative care research due to the special characteristics (deteriorating condition, fatigue and cachexia) of the population. Using data from a palliative study, we illustrate the problems that missing data can cause and show some approaches for dealing with it. Reasons for missing data and ways to deal with missing data (including complete case analysis, imputation and modelling procedures) are explored. Possible mechanisms behind the missing data are: missing completely at random, missing at random or missing not at random. In the example study, data are shown to be missing at random. Imputation of missing data is commonly used (including last value carried forward, regression procedures and simple mean). Imputation affects subsequent summary statistics and analyses, and can have a substantial impact on estimated group means and standard deviations. The choice of imputation method should be carried out with caution and the effects reported.