Missing covariate data in medical research: to impute is better than to ignore.

By K. Janssen, A. R. T. Donders, Frank E. Harrell, Y. Vergouwe, Qingxia Chen, D. Grobbee, Karel G.M. Moons. Journal of Clinical Epidemiology, 2010

599 citations recorded by the source corpus.

Subfields: Statistics and Probability, Surgery

View Paper: https://www.semanticscholar.org/paper/73b806b9afee4db0d559be116a812fc73bdbb170

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

In Surgery

Citations in the corpus

A new variable importance measure for random forests with missing data (2012, background): “As the latter is known to produce biased results in many situations (cf., e.g., Janssen et al. 2009, 2010) this strongly indicates that the omission of observations with missing values has induced artifacts because the values were not missing at random.”
Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS Checklist (2014, background): “However, reporting on the frequency and type of missing data is often poor [2,15,56,62,78,103–105] despite the adverse effects of missing data on development, validation, and updating of a prediction model [34,103,105–112].”
Missing covariate data in clinical research: when and when not to use the missing-indicator method for analysis (2012, result): “When the proportion of missingness increased, the difference between the methods became larger (results not shown), as shown by others.(19) The observation that the association between the variable with missing values (D-dimer level) and the outcome (deep venous thrombosis) is apparently unbiased (F”
Feature Robustness in Non-stationary Health Records: Caveats to Deployable Model Performance in Common Clinical Machine Learning Tasks (2019, methodology): “To account for the high rate of missingness within MIMIC-III, we impute our data Janssen et al. (2010) via a strategy known as “simple imputation,” which was introduced and validated by Che et al. (2018) for MIMIC time-series prediction tasks.”
External validation of multivariable prediction models: a systematic review of methodological conduct and reporting (2014, background): “Of particular concern is if those omitted are not representative of the whole population, that is the reason for the missingness is not completely at random [30].”
Risk prediction models: I. Development, internal validation, and assessing the incremental value of a new (bio)marker (2012, background): “Missing data Missing values are common in medical research, including prediction research.(49) 50 The potential influence of missing values on study results increases with the percentage of data that are missing.”