Multiple imputation: dealing with missing data

Goeij, Moniek C.M. de; Diepen, Merel van; Jager, Kitty J.; Tripepi, Giovanni; Zoccali, Carmine; Dekker, Friedo W.

doi:10.1093/ndt/gft221

Cited by 107 publications

(90 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To keep the full sample size and thus the same number of comparisons, the mean value substitution method was used to replace the missing values with the average value calculated over all the values available from the same group. This method was described to be adequate when the missing occurred completely at random [29]. However, the mean substitution biases the standard error downwards due to a higher number of values centered on the mean.…”

Section: Methodsmentioning

confidence: 99%

Effect of Topical Iloprost and Nitroglycerin on Gastric Microcirculation and Barrier Function during Hemorrhagic Shock in Dogs

et al. 2017

View full text Add to dashboard Cite

Background: Topical drug application is used to avoid systemic side effects. The aim of this study was to analyze whether locally applied iloprost or nitroglycerin influence gastric mucosal perfusion, oxygenation, and barrier function during physiological and hemorrhagic conditions. Methods: In repeated experiments, 5 anesthetized dogs received iloprost, nitroglycerin, or normal saline during physiological and hemorrhagic (-20% blood volume) conditions. Macro- and microcirculatory variables were recorded continuously. Gastric barrier function was assessed via translocation of sucrose into the blood. Results: During hemorrhage, gastric mucosal oxygenation decreased from 77 ± 4 to 37 ± 7%. This effect was attenuated by nitroglycerin (78 ± 6 to 47 ± 13%) and iloprost (82 ± 4 to 54 ± 9%). Sucrose plasma levels increased during hemorrhage from 7 ± 4 to 55 ± 15 relative amounts. This was alleviated by nitroglycerin (5 ± 8 to 29 ± 38 relative amounts). These effects were independent of systemic hemodynamic variables. Conclusions: During hemorrhage, topical nitroglycerin and iloprost improve regional gastric oxygenation without affecting perfusion. Nitroglycerin attenuated the shock-induced impairment of the mucosal barrier integrity. Thus, local drug application improves gastric microcirculation without compromising systemic hemodynamic variables, and it may also protect mucosal barrier function.

show abstract

Section: Methodsmentioning

confidence: 99%

Effect of Topical Iloprost and Nitroglycerin on Gastric Microcirculation and Barrier Function during Hemorrhagic Shock in Dogs

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Because the exclusion of subjects with missing values could have resulted in biased prospective results, multiple imputation (fully conditional specification according to the Markov Chain Monte Carlo method) was used to obtain 5 imputed data sets (24,25). Rubin's rules were followed to obtain pooled estimates of the regression coefficients and their SEs across the imputed data sets (26).…”

Section: Statistical Analysesmentioning

confidence: 99%

Functional vitamin B-6 status and long-term mortality in renal transplant recipients

Minovíc

Veen

Faassen

et al. 2017

The American Journal of Clinical Nutrition

View full text Add to dashboard Cite

Background: Low plasma concentrations of pyridoxal 5#-phosphate (PLP) are common in renal transplant recipients (RTRs) and confer increased risk of long-term mortality. To our knowledge, it is not known whether low plasma PLP concentrations have functional (i.e., intracellular) consequences and, if so, whether such consequences are associated with increased risk of mortality. Objectives: We assessed the association of plasma PLP with functional vitamin B-6 status and explored the potential association of functional vitamin B-6 status with long-term mortality in RTRs.

show abstract

“…As there were missing values on the additional candidate variables (table 1), we applied multiple imputation [19,20] before developing and validating the new models. Non-normally distributed variables were transformed to variables with a more normal distribution either logarithmically (cholesterol, phosphate, calcium and Kt/V) or by taking the square root (GFR and Karnofsky score) before imputing.…”

Section: Methodsmentioning

confidence: 99%

Improved Mortality Prediction in Dialysis Patients Using Specific Clinical and Laboratory Data

Hemke¹,

Heemskerk²,

Diepen

et al. 2015

Am J Nephrol

View full text Add to dashboard Cite

Background: Risk prediction models can be used to inform patients undergoing renal replacement therapy about their survival chances. Easily available predictors such as registry data are most convenient, but their predictive value may be limited. We aimed to improve a simple prediction model based on registry data by incrementally adding sets of clinical and laboratory variables. Methods: Our data set includes 1,835 Dutch patients from the Netherlands Cooperative Study on the Adequacy of Dialysis. The potential survival predictors were categorized on availability. The first category includes easily available clinical data. The second set includes laboratory values like albumin. The most laborious category contains glomerular filtration rate (GFR) and Kt/V. Missing values were substituted using multiple imputation. Within 1,225 patients, we recalibrated the registry model and subsequently added parameter sets using multivariate Cox regression analyses with backward selection. On the other 610 patients, calibration and discrimination (C-index, integrated discrimination improvement (IDI) index and net reclassification improvement (NRI) index) were assessed for all models. Results: The recalibrated registry model showed adequate calibration and discrimination (C-index = 0.724). Adding easily available parameters resulted in a model with 10 predictors, with similar calibration and improved discrimination (C-index = 0.784). The IDI and NRI indices confirmed this, especially for short-term survival. Adding laboratory values resulted in an alternative model with similar discrimination (C-index = 0.788), and only the NRI index showed minor improvement. Adding GFR and Kt/V as candidate predictors did not result in a different model. Conclusion: A simple model based on registry data was enhanced by adding easily available clinical parameters.

show abstract

Multiple imputation: dealing with missing data

Cited by 107 publications

References 13 publications

Effect of Topical Iloprost and Nitroglycerin on Gastric Microcirculation and Barrier Function during Hemorrhagic Shock in Dogs

Effect of Topical Iloprost and Nitroglycerin on Gastric Microcirculation and Barrier Function during Hemorrhagic Shock in Dogs

Functional vitamin B-6 status and long-term mortality in renal transplant recipients

Improved Mortality Prediction in Dialysis Patients Using Specific Clinical and Laboratory Data

Contact Info

Product

Resources

About