On the Relation between Prediction and Imputation Accuracy under Missing Covariates.

Missing covariates in regression or classification problems can prohibit the direct use of advanced tools for further analysis. Recent research has realized an increasing trend towards the use of modern Machine-Learning algorithms for imputation. This originates from their capability of showing favorable prediction accuracy in different learning problems. In this work, we analyze through simulation the interaction between imputation accuracy and prediction accuracy in regression learning problems with missing covariates when Machine-Learning-based methods for both imputation and prediction are used. We see that even a slight decrease in imputation accuracy can seriously affect the prediction accuracy. In addition, we explore imputation performance when using statistical inference procedures in prediction settings, such as the coverage rates of (valid) prediction intervals. Our analysis is based on empirical datasets provided by the UCI Machine Learning repository and an extensive simulation study.

High-sensitivity cardiac troponin I after coronary artery bypass grafting for post-operative decision-making.

AIMS: Current troponin cut-offs suggested for the post-operative workup of patients following coronary artery bypass graft (CABG) surgery are based on studies using non-high-sensitive troponin assays or are arbitrarily chosen. We aimed to identify an optimal cut-off and timing for a proprietary high-sensitivity cardiac troponin I (hs-cTnI) assay to facilitate post-operative clinical decision-making. METHODS AND RESULTS: We performed a retrospective analysis of all patients undergoing elective isolated CABG at our centre between January 2013 and May 2019. Of 4684 consecutive patients, 161 patients (3.48%) underwent invasive coronary angiography after surgery, of whom 86 patients (53.4%) underwent repeat revascularization. We found an optimal cut-off value for peak hs-cTnI of >13 000 ng/L [>500× the upper reference limit (URL)] to be significantly associated with repeat revascularization within 48 h after surgery, which was internally validated through random repeated sampling with 1000 iterations. The same cut-off also predicted 30-day major adverse cardiovascular events and all-cause mortality after a median follow-up of 3.1 years, which was validated in an external cohort. A decision tree analysis of serial hs-cTnI measurements showed no added benefit of hs-cTnI measurements in patients with electrocardiographic or echocardiographic abnormalities or haemodynamic instability. Likewise, early post-operative hs-cTnI elevations had a low yield for clinical decision-making and only later elevations (at 12-16 h post-operatively) using a threshold of 8000 ng/L (307× URL) were significantly associated with repeat revascularization with an area under the curve of 0.92 (95% confidence interval 0.88-0.95). CONCLUSION: Our data suggest that for hs-cTnI, higher cut-offs than currently recommended should be used in the post-operative management of patients following CABG.

Asymptotic-based bootstrap approach for matched pairs with missingness in a single arm.

The issue of missing values is an arising difficulty when dealing with paired data. Several test procedures are developed in the literature to tackle this problem. Some of them are even robust under deviations and control type-I error quite accurately. However, most of these methods are not applicable when missing values are present only in a single arm. For this case, we provide asymptotic correct resampling tests that are robust under heteroskedasticity and skewed distributions. The tests are based on a meaningful restructuring of all observed information in quadratic form-type test statistics. An extensive simulation study is conducted exemplifying the tests for finite sample sizes under different missingness mechanisms. In addition, illustrative data examples based on real life studies are analyzed.

A cautionary tale on using imputation methods for inference in matched-pairs design.

MOTIVATION: Imputation procedures in biomedical fields have turned into statistical practice, since further analyses can be conducted ignoring the former presence of missing values. In particular, non-parametric imputation schemes like the random forest have shown favorable imputation performance compared to the more traditionally used MICE procedure. However, their effect on valid statistical inference has not been analyzed so far. This article closes this gap by investigating their validity for inferring mean differences in incompletely observed pairs while opposing them to a recent approach that only works with the given observations at hand. RESULTS: Our findings indicate that machine-learning schemes for (multiply) imputing missing values may inflate type I error or result in comparably low power in small-to-moderate matched pairs, even after modifying the test statistics using Rubin's multiple imputation rule. In addition to an extensive simulation study, an illustrative data example from a breast cancer gene study has been considered. AVAILABILITY AND IMPLEMENTATION: The corresponding R-code can be accessed through the authors and the gene expression data can be downloaded at www.gdac.broadinstitute.org. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.