In JMIR medical informatics ; h5-index 23.0
BACKGROUND : Patients acquire pressure injuries (PI) in the hospital due to low mobility, exposure to localized pressure, circulatory conditions, and other predisposing factors. Over 2.5 million Americans develop pressure injuries annually. The Center for Medicare and Medicaid considers hospital acquired pressure injuries (HAPI) as the most frequent preventable event, and they are the 2nd most common claim in lawsuits. With growing utilization of electronic health records (EHRs) in hospitals, an opportunity exists to build machine learning models to identify and predict HAPI, rather than relying on occasional manual assessments by human experts. Yet accurate computational models rely on high-quality data HAPI labels. Unfortunately, the different data sources within EHRs can provide conflicting information on HAPI occurrence in the same patient. Furthermore, existing definitions of HAPI disagree with each other, even within the same patient population. The inconsistent criteria make it impossible to benchmark machine learning methods to predict HAPI.
OBJECTIVE : The objective of this project is threefold: 1) identify discrepancies in HAPI sources within EHRs; 2) develop a comprehensive definition for HAPI classification using data from all EHR sources; and 3) illustrate the importance of an improved HAPI definition.
METHODS : We assessed congruence among HAPI occurrences documented in clinical notes, diagnosis codes, procedure codes, and chart events from the Medical Information Mart for Intensive Care III (MIMIC-III) database. We analyzed the criteria used for three existing HAPI definitions and their adherence with regulatory guidelines. We proposed Emory HAPI (EHAPI), an improved and more comprehensive HAPI definition. We then evaluated the importance of the labels in training a HAPI classification model using tree-based and sequential neural network classifiers.
RESULTS : We illustrate the complexity of defining HAPI, with less than 13% of hospital stays having at least 3 PI-indications documented across 4 data sources. While chart events were the most common indicator, it was the only PI documentation for over 49% of stays. We demonstrate lack of congruence across existing HAPI definitions and EHAPI, with only 219 stays where there was a consensus positive label. Our analysis highlights the importance of our improved HAPI definition, with classifiers trained using our labels outperforming others on a small manually labeled set from nurse annotators and a consensus set in which all definitions agreed on the label.
CONCLUSIONS : Standardized HAPI definitions are important for accurate HAPI nursing quality metric assessment and determining HAPI incidence for preventive measures. We demonstrate the complexity of defining an occurrence of HAPI given conflicting and incomplete EHR data. Our EHAPI definition has favorable properties making it a suitable candidate for HAPI classification tasks.
Sotoodeh Mani, Zhang Wenhui, Simpson Roy, Hertzberg Vicki, Ho Joyce
2023-Jan-14
