Unmasking bias in artificial intelligence: a systematic review of bias detection and mitigation strategies in electronic health record-based models

Abstract

Objectives: Leveraging artificial intelligence (AI) in conjunction with electronic health records (EHRs) holds transformative potential to improve healthcare. However, addressing bias in AI, which risks worsening healthcare disparities, cannot be overlooked. This study reviews methods to handle various biases in AI models developed using EHR data.

Materials and methods: We conducted a systematic review following the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines, analyzing articles from PubMed, Web of Science, and IEEE published between January 01, 2010 and December 17, 2023. The review identified key biases, outlined strategies for detecting and mitigating bias throughout the AI model development, and analyzed metrics for bias assessment.

Results: Of the 450 articles retrieved, 20 met our criteria, revealing 6 major bias types: algorithmic, confounding, implicit, measurement, selection, and temporal. The AI models were primarily developed for predictive tasks, yet none have been deployed in real-world healthcare settings. Five studies concentrated on the detection of implicit and algorithmic biases employing fairness metrics like statistical parity, equal opportunity, and predictive equity. Fifteen studies proposed strategies for mitigating biases, especially targeting implicit and selection biases. These strategies, evaluated through both performance and fairness metrics, predominantly involved data collection and preprocessing techniques like resampling and reweighting.

Discussion: This review highlights evolving strategies to mitigate bias in EHR-based AI models, emphasizing the urgent need for both standardized and detailed reporting of the methodologies and systematic real-world testing and evaluation. Such measures are essential for gauging models' practical impact and fostering ethical AI that ensures fairness and equity in healthcare.

Keywords: artificial intelligence; bias; deep learning; electronic health record; scoping review.

Figure 1.

PRISMA flow diagram.

Figure 2.

Bias handling workflow in artificial intelligence (AI) model development. The pipeline for the AI applications is shown in the blue box. Possible types of bias that could be introduced in each step are shown in the orange box, while possible handling approaches during each step of the model development are shown in green. Preprocessing handling methods refer to analyzing and adjusting the existing data set to preempt biases resulting from inadequate data during data collection and preparation. In-processing handling methods aim to handle bias during the model training and testing to avoid bias during training and eliminate bias from input data. Postprocessing handling methods account for handling model biases by interpreting or adjusting model outputs and correctly making use of the result. Approaches marked with * were approaches that were not covered by papers in this review.