Imprimir Ver referencias Citación Disclaimer: These citations have been automatically generated based on the information we have and it may not be 100% accurate. Please consult the latest official manual style if you have any questions regarding the format accuracy. AMA Citation Lennon J, Shah R. Lennon J, & Shah R Lennon, Jack, and Ravi Shah. Predicting Death By Suicide Following An Emergency Department Visit For Parasuicide With Administrative Health Care System Data and Machine Learning. 2 Minute Medicine, 28 febrero 2020. McGraw-Hill, 2020. AccessMedicina. https://accessmedicina.mhmedical.com/updatesContent.aspx?gbosid=533502§ionid=240560543APA Citation Lennon J, Shah R. Lennon J, & Shah R Lennon, Jack, and Ravi Shah. (2020). Predicting death by suicide following an emergency department visit for parasuicide with administrative health care system data and machine learning. (2020). 2 minute medicine. McGraw-Hill. https://accessmedicina.mhmedical.com/updatesContent.aspx?gbosid=533502§ionid=240560543.MLA Citation Lennon J, Shah R. Lennon J, & Shah R Lennon, Jack, and Ravi Shah. "Predicting Death By Suicide Following An Emergency Department Visit For Parasuicide With Administrative Health Care System Data and Machine Learning." 2 Minute Medicine McGraw-Hill, 2020, https://accessmedicina.mhmedical.com/updatesContent.aspx?gbosid=533502§ionid=240560543. Descargar archivo de la citación: RIS (Zotero) EndNote BibTex Medlars ProCite RefWorks Reference Manager Mendeley © Copyright Clip Capítulo completo Sólo figuras Sólo cuadros Solo Videos Supplementary Content Arriba Predicting Death By Suicide Following An Emergency Department Visit For Parasuicide With Administrative Health Care System Data and Machine Learning by Jack Lennon, Ravi Shah Listen +Originally published by 2 Minute Medicine® (view original article). Reused on AccessMedicine with permission. +1. This study is the first to offer strong machine-learning performance to suggest the potential for clinical utility in predicting future suicide deaths within 90 days of a visit to the emergency department. +Evidence Rating Level: 1 (Excellent) +Suicide is a leading global health concern increasing annually in many countries. The use of administrative healthcare system data and machine learning is becoming a popular method for establishing predictive methods based on a large number of medical and socioecological predictors. Five administrative healthcare data systems comprised the dataset, which included virtually all ambulatory care visits, physician visits, hospitalizations, and community pharmacy uses in Alberta, Canada. A total of 101 predictors were selected through literature search of assessment tools and statistical prediction models, which were set for each of the eight quarters prior to the quarter of suicide death. All individuals visiting an emergency department (ED) for parasuicide (ED visit for self-harm not resulting in death) between 2010 and 2017 were extracted, with the most recent ED visit being the data point of use. A total of 268 individuals died by suicide within 90 days and 33,426 did not. Due to this imbalance, models utilized class weights of 124/125 for persons that died by suicide and 1/125 for persons that did not die by suicide. The 10-fold cross-validation area under the curve (AUC) estimates for logistic regression with L2 regularization penalty enabled (to avoid overfitting) were 0.8590, 0.8632, 0.8572, 0.8454, 0.8392 for 1, 2, 4, 6, and 8 quarter(2 years) of data, respectively. With the optimal gradient boosted tree (XGB) model configuration, the 10-fold cross-validation AUC estimate was 0.8786. While similar performances, the XGB model demonstrated higher optimal AUC estimates. Given these findings, which are solely dependent on the models derived from a given administrative healthcare data system, this study suggests a true feasibility for quantifying suicide risk and, more specifically, suicide death within 90 days following an ED visit for parasuicide. Further research is necessary for different settings, with ongoing alterations prior to allowing these models to make clinical judgments. +Click to read the study in EClincalMedicine +©2020 2 Minute Medicine, Inc. All rights reserved. No works may be reproduced without expressed written consent from 2 Minute Medicine, Inc. Inquire about licensing here. No article should be construed as medical advice and is not intended as such by the authors or by 2 Minute Medicine, Inc.