Resolving the Bias in Electronic Medical Records

Zheng, Kaiping; Gao, Jinyang; Ngiam, Kee Yuan; Ooi, Beng Chin; Yip, Wei Luen James

doi:10.1145/3097983.3098149

Cited by 38 publications

(24 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In general, issues of sampling bias are not unique to EHR data, and many authors have explored the impact of sampling on inference. Some works exploring selection/observation biases in the EHR setting include Zheng et al, Phelan et al, Goldstein et al, and Rusanov et al However, additional characterizations of the mechanisms by which we can have sampling bias in biobank and EHR research may help guide study design in the future.…”

Section: Statistical Issues Related To Biobank Researchmentioning

confidence: 99%

The emerging landscape of health research based on biobanks linked to electronic health records: Existing resources, statistical challenges, and potential opportunities

Beesley

Salvatore

Fritsche

et al. 2019

Statistics in Medicine

View full text Add to dashboard Cite

Biobanks linked to electronic health records provide rich resources for health‐related research. With improvements in administrative and informatics infrastructure, the availability and utility of data from biobanks have dramatically increased. In this paper, we first aim to characterize the current landscape of available biobanks and to describe specific biobanks, including their place of origin, size, and data types. The development and accessibility of large‐scale biorepositories provide the opportunity to accelerate agnostic searches, expedite discoveries, and conduct hypothesis‐generating studies of disease‐treatment, disease‐exposure, and disease‐gene associations. Rather than designing and implementing a single study focused on a few targeted hypotheses, researchers can potentially use biobanks' existing resources to answer an expanded selection of exploratory questions as quickly as they can analyze them. However, there are many obvious and subtle challenges with the design and analysis of biobank‐based studies. Our second aim is to discuss statistical issues related to biobank research such as study design, sampling strategy, phenotype identification, and missing data. We focus our discussion on biobanks that are linked to electronic health records. Some of the analytic issues are illustrated using data from the Michigan Genomics Initiative and UK Biobank, two biobanks with two different recruitment mechanisms. We summarize the current body of literature for addressing these challenges and discuss some standing open problems. This work complements and extends recent reviews about biobank‐based research and serves as a resource catalog with analytical and practical guidance for statisticians, epidemiologists, and other medical researchers pursuing research using biobanks.

show abstract

Section: Statistical Issues Related To Biobank Researchmentioning

confidence: 99%

The emerging landscape of health research based on biobanks linked to electronic health records: Existing resources, statistical challenges, and potential opportunities

Beesley

Salvatore

Fritsche

et al. 2019

Statistics in Medicine

View full text Add to dashboard Cite

show abstract

“…Studies on machine learning prediction models using EMR data have inherent problems such as high dimensionality and sparsity, data bias, and few abnormal events [ 16 - 18 ]. Previous studies have tried to resolve the abovementioned problems by using several techniques such as oversampling, undersampling, data handling, and feature selection [ 19 - 22 ]. However, the performance of the model that learned processed data by using data augmentation has not significantly improved compared to that of the previous prediction models, and the EMR-based prediction model is still being challenged [ 17 , 20 , 21 ].…”

Section: Introductionmentioning

confidence: 99%

A Predictive Model Based on Machine Learning for the Early Detection of Late-Onset Neonatal Sepsis: Development and Observational Study

et al. 2020

View full text Add to dashboard Cite

Background Neonatal sepsis is associated with most cases of mortalities and morbidities in the neonatal intensive care unit (NICU). Many studies have developed prediction models for the early diagnosis of bloodstream infections in newborns, but there are limitations to data collection and management because these models are based on high-resolution waveform data. Objective The aim of this study was to examine the feasibility of a prediction model by using noninvasive vital sign data and machine learning technology. Methods We used electronic medical record data in intensive care units published in the Medical Information Mart for Intensive Care III clinical database. The late-onset neonatal sepsis (LONS) prediction algorithm using our proposed forward feature selection technique was based on NICU inpatient data and was designed to detect clinical sepsis 48 hours before occurrence. The performance of this prediction model was evaluated using various feature selection algorithms and machine learning models. Results The performance of the LONS prediction model was found to be comparable to that of the prediction models that use invasive data such as high-resolution vital sign data, blood gas estimations, blood cell counts, and pH levels. The area under the receiver operating characteristic curve of the 48-hour prediction model was 0.861 and that of the onset detection model was 0.868. The main features that could be vital candidate markers for clinical neonatal sepsis were blood pressure, oxygen saturation, and body temperature. Feature generation using kurtosis and skewness of the features showed the highest performance. Conclusions The findings of our study confirmed that the LONS prediction model based on machine learning can be developed using vital sign data that are regularly measured in clinical settings. Future studies should conduct external validation by using different types of data sets and actual clinical verification of the developed model.

show abstract

“…Hidden Markov models extend the latent variable approach by allowing a time-varying latent process. Zheng et al 52 and Alaa et al 53 used hidden Markov models to capture IO, but the way they incorporated the observation process differs. Hidden Markov model–based prediction models incorporate IO by allowing the measurement frequency or rate to infer the clinical state at any given time.…”

Section: Resultsmentioning

confidence: 99%

Informative presence and observation in routine health data: A review of methodology for clinical risk prediction

Sisk

Lin

Sperrin

et al. 2020

Journal of the American Medical Informatics Association

View full text Add to dashboard Cite

Objective Informative presence (IP) is the phenomenon whereby the presence or absence of patient data is potentially informative with respect to their health condition, with informative observation (IO) being the longitudinal equivalent. These phenomena predominantly exist within routinely collected healthcare data, in which data collection is driven by the clinical requirements of patients and clinicians. The extent to which IP and IO are considered when using such data to develop clinical prediction models (CPMs) is unknown, as is the existing methodology aiming at handling these issues. This review aims to synthesize such existing methodology, thereby helping identify an agenda for future methodological work. Materials and Methods A systematic literature search was conducted by 2 independent reviewers using prespecified keywords. Results Thirty-six articles were included. We categorized the methods presented within as derived predictors (including some representation of the measurement process as a predictor in the model), modeling under IP, and latent structures. Including missing indicators or summary measures as predictors is the most commonly presented approach amongst the included studies (24 of 36 articles). Discussion This is the first review to collate the literature in this area under a prediction framework. A considerable body relevant of literature exists, and we present ways in which the described methods could be developed further. Guidance is required for specifying the conditions under which each method should be used to enable applied prediction modelers to use these methods. Conclusions A growing recognition of IP and IO exists within the literature, and methodology is increasingly becoming available to leverage these phenomena for prediction purposes. IP and IO should be approached differently in a prediction context than when the primary goal is explanation. The work included in this review has demonstrated theoretical and empirical benefits of incorporating IP and IO, and therefore we recommend that applied health researchers consider incorporating these methods in their work.

show abstract

Resolving the Bias in Electronic Medical Records

Cited by 38 publications

References 13 publications

The emerging landscape of health research based on biobanks linked to electronic health records: Existing resources, statistical challenges, and potential opportunities

The emerging landscape of health research based on biobanks linked to electronic health records: Existing resources, statistical challenges, and potential opportunities

A Predictive Model Based on Machine Learning for the Early Detection of Late-Onset Neonatal Sepsis: Development and Observational Study

Informative presence and observation in routine health data: A review of methodology for clinical risk prediction

Contact Info

Product

Resources

About