Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review

By Benjamin Alan Goldstein, Ann Marie Navar, Michael J. Pencina, John P. A. Ioannidis. 2017

840 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Epidemiology, Health Information Management

How researchers see this paper

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

In Artificial Intelligence

In Epidemiology

In Health Information Management

Citations in the corpus

AutoScore: A Machine Learning–Based Automatic Clinical Score Generator and Its Application to Mortality Prediction Using Electronic Health Records (2020, background): “At present, the increasing popularity of electronic health records (EHRs) [9] creates an opportunity to take advantage of its growing quantity and diversity of data for creating novel risk models with both domain expert–curated approaches and advanced machine learning solutions.”
Scalable and accurate deep learning with electronic health records (2018, background): “In spite of the richness and potential of available data, scaling the development of predictive models is difficult because, for traditional predictive modeling techniques, each outcome to be predicted requires the creation of a custom dataset with specific variables [17].”
The role of explainability in creating trustworthy artificial intelligence for health care: a comprehensive survey of the terminology, design choices, and evaluation strategies (2020, background): “Although external validation is recognized as an area of improvement for clinical risk prediction models [85], it is increasingly feasible with the adoption of common data structures in health care (e.”
Clinical Information Extraction Applications: A Literature Review (2017, methodology): “[3] evaluated the state of EHR-based risk prediction modeling through a systematic review of clinical prediction studies using EHR”
Machine learning in cardiovascular medicine: are we there yet? (2018, background): “40 However, machine learning algorithms are not a panacea and also have limitations.”
Artificial intelligence-enabled healthcare delivery (2018, background): “However, at this stage, there are some issues like data and label availability, the explainability of the machine learning model and the ability to easily integrate such models with existing electronic health record systems that have yet to be addressed.(27) New algorithmic approaches using transfer”
FADL: Federated-Autonomous Deep Learning for Distributed Electronic Health Record (2018, background): “Healthcare records are stored in different locations and data silos, including but not limited to hospitals, pharmacies, payors, and personal devices[6, 17, 2, 9].”
GRU-ODE-Bayes: Continuous modeling of sporadically-observed time series (2019, background): “Electronic Health Records (EHR) analysis is crucial to achieve data-driven personalized medicine (Lee et al., 2017; Goldstein et al., 2017; Esteva et al., 2019).”
Role of Artificial Intelligence in Patient Safety Outcomes: Systematic Literature Review (2020, background): “A review [53] conducted in 2017 reported that only about 54% of studies that developed prediction models based on EHRs accounted for missing data.”
Using clinical Natural Language Processing for health outcomes research: Overview and actionable suggestions for future advances (2018, neutral): “Thus when using outputs from NLP approaches in clinical research studies, it is not always clear how best to incorporate and interpret NLP performance metrics.”
Natural Language Processing of Clinical Notes on Chronic Diseases: Systematic Review (2019, methodology): “Methods based in analysis of EHRs [2] are resulting in improved understanding of patient clinical trajectories [3] while enabling better patient stratification and risk prediction [4-6].”
Potential Biases in Machine Learning Algorithms Using Electronic Health Record Data (2018, background): “A 2017 review 13 found that only 54% of studies that generated prediction algorithms based on the electronic health record accountedformissingdata.”