Dataset Information

Comparison of methods for imputing limited-range variables: a simulation study.

ABSTRACT:

Background

Multiple imputation (MI) was developed as a method to enable valid inferences to be obtained in the presence of missing data rather than to re-create the missing values. Within the applied setting, it remains unclear how important it is that imputed values should be plausible for individual observations. One variable type for which MI may lead to implausible values is a limited-range variable, where imputed values may fall outside the observable range. The aim of this work was to compare methods for imputing limited-range variables, with a focus on those that restrict the range of the imputed values.

Methods

Using data from a study of adolescent health, we consider three variables based on responses to the General Health Questionnaire (GHQ), a tool for detecting minor psychiatric illness. These variables, based on different scoring methods for the GHQ, resulted in three continuous distributions with mild, moderate and severe positive skewness. In an otherwise complete dataset, we set 33% of the GHQ observations to missing completely at random or missing at random; repeating this process to create 1000 datasets with incomplete data for each scenario.For each dataset, we imputed values on the raw scale and following a zero-skewness log transformation using: univariate regression with no rounding; post-imputation rounding; truncated normal regression; and predictive mean matching. We estimated the marginal mean of the GHQ and the association between the GHQ and a fully observed binary outcome, comparing the results with complete data statistics.

Results

Imputation with no rounding performed well when applied to data on the raw scale. Post-imputation rounding and imputation using truncated normal regression produced higher marginal means than the complete data estimate when data had a moderate or severe skew, and this was associated with under-coverage of the complete data estimate. Predictive mean matching also produced under-coverage of the complete data estimate. For the estimate of association, all methods produced similar estimates to the complete data.

Conclusions

For data with a limited range, multiple imputation using techniques that restrict the range of imputed values can result in biased estimates for the marginal mean when data are highly skewed.

SUBMITTER: Rodwell L

PROVIDER: S-EPMC4021274 | biostudies-literature | 2014 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Comparison of methods for imputing limited-range variables: a simulation study.

Rodwell Laura L Lee Katherine J KJ Romaniuk Helena H Carlin John B JB

BMC medical research methodology 20140426

<h4>Background</h4>Multiple imputation (MI) was developed as a method to enable valid inferences to be obtained in the presence of missing data rather than to re-create the missing values. Within the applied setting, it remains unclear how important it is that imputed values should be plausible for individual observations. One variable type for which MI may lead to implausible values is a limited-range variable, where imputed values may fall outside the observable range. The aim of this work was ...[more]

PMID: 24766825

Similar Datasets

Project description:BackgroundInterprofessional collaborative practice is essential for meeting patients' needs and improving their health outcomes; thus, the effectiveness of interprofessional education (IPE) should be clearly identified. There is insufficient evidence in the literature to determine the outcomes of IPE compared to traditional single-profession education (SPE). This study aimed to compare the outcomes of IPE and SPE during a simulation training course.MethodsThe study design was a mixed-methods, incorporated cross-over design and a qualitative survey. A total of 54 students including 18 medical students and 36 nursing students were recruited from March to April 2019. The 4-week simulation course was designed based on Kolb's experimental learning theory and Bandura's social learning theory. Participants were evenly divided into group 1 (received IPE-learning followed by SPE-learning), and group 2 (received SPE-learning followed by IPE-learning). Students' medical task performance, team behavior performance, teamwork attitude, and patient safety attitude were collected at pretest, mid-test, and posttest. Descriptive statistics and repeated measures analysis of variance were used. End-of-study qualitative feedback was collected, and content analysis was performed.ResultsBoth groups demonstrated moderate-to-large within-group improvements for multiple learning outcomes at mid-test. Group 1 students' medical task performance (F = 97.25; P < 0.001) and team behavior performance (F = 31.17; P < 0.001) improved significantly. Group 2 students' medical task performance (F = 77.77; P < 0.001), team behavior performance (F = 40.14; P < 0.001), and patient safety attitude (F = 6.82; P < 0.01) improved significantly. Outcome differences between groups were nonsignificant. Qualitative themes identified included: personal factor, professional factor, interprofessional relationship, and learning. The IPE program provided students with exposure to other professions and revealed differences in expertise and responsibilities.ConclusionIPE-simulation and SPE-simulation were effective interventions that enabled medical and nursing students to develop critical medical management and team behavior performance. IPE-simulation provided more opportunities for improving competencies in interprofessional collaborative practice. In circumstances with limited teaching resources, SPE-simulation can be an acceptable alternative to IPE-simulation.

Dataset Information

Comparison of methods for imputing limited-range variables: a simulation study.

Background

Methods

Results

Conclusions

Publications

Comparison of methods for imputing limited-range variables: a simulation study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets