Departments of Population Health and Radiology Center for Data Science New Y ork University Langone Medical Center Abstract As machine learning (ML) models, trained on real-world datasets, become common practice, it is critical to measure and quantify their potential biases. In this paper, we focus on renal failure and compare a commonly used traditional risk score, Tangri, with a more powerful machine learning model, which has access to a larger variable set and trained on 1.6 million patients' EHR data. We will compare and discuss the generalization and applicability of these two models, in an attempt to quantify biases of status quo clinical practice, compared to MLdriven models. 1 Introduction Data-driven models have become more common in the U.S. healthcare field as their use in clinical operations and diagnosing procedures have expanded exponentially. The ever-increasing processing power of machine-learning algorithms allows automatic analysis of huge quantities of data, theoretically maximizing the efficiency and accuracy of the medical diagnosing process. Predictions from machine-learning models already drive important healthcare decisions for over 70 million people across the United States[7].

MLP networks provided slightly better risk prediction than conventional logistic regression when used to predict the risk of death, stroke, and renal failure on 1257 patients who underwent coronary artery bypass operations. Bootstrap sampling was required to compare approaches and regularization provided by early stopping was an important component of improved performance. A simplified approach to generating confidence intervals for MLP risk predictions using an auxiliary "confidence MLP" was also developed. The confidence MLP is trained to reproduce the confidence bounds that were generated during training by 50 MLP networks trained using bootstrap samples. Current research is validating these results using larger data sets, exploring approaches to detect outlier patients who are so different from any training patient that accurate risk prediction is suspect, developing approaches to explaining which input features are important for an individual patient, and determining why MLP networks provide improved performance.

MLP networks provided slightly better risk prediction than conventional logistic regression when used to predict the risk of death, stroke, and renal failure on 1257 patients who underwent coronary artery bypass operations. Bootstrap sampling was required to compare approaches and regularization provided by early stopping was an important component of improved performance. A simplified approach to generating confidence intervals for MLP risk predictions using an auxiliary "confidence MLP" was also developed. The confidence MLP is trained to reproduce the confidence bounds that were generated during training by 50 MLP networks trained using bootstrap samples. Current research is validating these results using larger data sets, exploring approaches to detect outlier patients who are so different from any training patient that accurate risk prediction is suspect, developing approaches to explaining which input features are important for an individual patient, and determining why MLP networks provide improved performance.

MLP networks provided slightly better risk prediction than conventional logistic regression when used to predict the risk of death, stroke, and renal failure on 1257 patients who underwent coronaryartery bypass operations. Bootstrap sampling was required to compare approaches and regularization provided by early stopping was an important component of improved performance. A simplified approach to generating confidence intervals for MLP risk predictions using an auxiliary "confidence MLP" was also developed. The confidence MLP is trained to reproduce the confidence bounds that were generated during training by 50 MLP networks trained using bootstrap samples. Current research is validating these results usinglarger data sets, exploring approaches to detect outlier patients who are so different fromany training patient that accurate risk prediction is suspect, developing approaches toexplaining which input features are important for an individual patient, and determining why MLP networks provide improved performance.

In recent years many attempts have been made to use the computer as an aid to diagnosis, but little has been done to exploit the potential of computer technology as a more general aid to decision making. We describe the use of the discipline of decision analysis as the basis for an experimental interactive computer program designed to assist the physician in the clinical management of acute oliguric renal failure. The program deals with alternative courses of action, either tests or treatments, for which the potential risks or benefits may be large, and it balances the anticipated risk of a given strategy against the anticipated benefit that it offers the patient. The appraisals of the different courses of action open to the physician are expressed in quantitative terms as expected value. The program has been evaluated by comparing its recommendations to those of experienced nephrologists in 18 simulated cases of acute oliguric renal failure.