Robert M Cronin1, Jacob P VanHouten2, Edward D Siew3, Svetlana K Eden4, Stephan D Fihn5, Christopher D Nielson6, Josh F Peterson7, Clifton R Baker8, T Alp Ikizler3, Theodore Speroff9, Michael E Matheny10. 1. Geriatric Research Education Clinical Center, Tennessee Valley Health System, Veterans Health Administration, Nashville, TN, USA Department of Biomedical Informatics, Vanderbilt University School of Medicine, Nashville, TN, USA Division of General Internal Medicine and Public Health, Vanderbilt University School of Medicine, Nashville, TN, USA. 2. Department of Biomedical Informatics, Vanderbilt University School of Medicine, Nashville, TN, USA Department of Biostatistics, Vanderbilt University School of Medicine, Nashville, TN, USA. 3. Division of Nephrology, Vanderbilt University School of Medicine, Nashville, TN, USA. 4. Department of Biostatistics, Vanderbilt University School of Medicine, Nashville, TN, USA. 5. Office of Analytics and Business Intelligence, VA Central Office, Veterans Health Administration, Seattle, WA, USA Division of General Internal Medicine, University of Washington, Seattle, WA, USA. 6. Office of Analytics and Business Intelligence, VA Central Office, Veterans Health Administration, Seattle, WA, USA Division of Pulmonary Medicine and Critical Care, University of Nevada, Reno, NV, USA. 7. Department of Biomedical Informatics, Vanderbilt University School of Medicine, Nashville, TN, USA. 8. Office of Analytics and Business Intelligence, VA Central Office, Veterans Health Administration, Seattle, WA, USA. 9. Geriatric Research Education Clinical Center, Tennessee Valley Health System, Veterans Health Administration, Nashville, TN, USA Division of General Internal Medicine and Public Health, Vanderbilt University School of Medicine, Nashville, TN, USA Department of Biostatistics, Vanderbilt University School of Medicine, Nashville, TN, USA. 10. Geriatric Research Education Clinical Center, Tennessee Valley Health System, Veterans Health Administration, Nashville, TN, USA Department of Biomedical Informatics, Vanderbilt University School of Medicine, Nashville, TN, USA Division of General Internal Medicine and Public Health, Vanderbilt University School of Medicine, Nashville, TN, USA Department of Biostatistics, Vanderbilt University School of Medicine, Nashville, TN, USA Michael.matheny@vanderbilt.edu.
Abstract
OBJECTIVE: Hospital-acquired acute kidney injury (HA-AKI) is a potentially preventable cause of morbidity and mortality. Identifying high-risk patients prior to the onset of kidney injury is a key step towards AKI prevention. MATERIALS AND METHODS: A national retrospective cohort of 1,620,898 patient hospitalizations from 116 Veterans Affairs hospitals was assembled from electronic health record (EHR) data collected from 2003 to 2012. HA-AKI was defined at stage 1+, stage 2+, and dialysis. EHR-based predictors were identified through logistic regression, least absolute shrinkage and selection operator (lasso) regression, and random forests, and pair-wise comparisons between each were made. Calibration and discrimination metrics were calculated using 50 bootstrap iterations. In the final models, we report odds ratios, 95% confidence intervals, and importance rankings for predictor variables to evaluate their significance. RESULTS: The area under the receiver operating characteristic curve (AUC) for the different model outcomes ranged from 0.746 to 0.758 in stage 1+, 0.714 to 0.720 in stage 2+, and 0.823 to 0.825 in dialysis. Logistic regression had the best AUC in stage 1+ and dialysis. Random forests had the best AUC in stage 2+ but the least favorable calibration plots. Multiple risk factors were significant in our models, including some nonsteroidal anti-inflammatory drugs, blood pressure medications, antibiotics, and intravenous fluids given during the first 48 h of admission. CONCLUSIONS: This study demonstrated that, although all the models tested had good discrimination, performance characteristics varied between methods, and the random forests models did not calibrate as well as the lasso or logistic regression models. In addition, novel modifiable risk factors were explored and found to be significant. Published by Oxford University Press on behalf of the American Medical Informatics Association 2015. This work is written by US Government employees and is in the public domain in the US.
OBJECTIVE: Hospital-acquired acute kidney injury (HA-AKI) is a potentially preventable cause of morbidity and mortality. Identifying high-risk patients prior to the onset of kidney injury is a key step towards AKI prevention. MATERIALS AND METHODS: A national retrospective cohort of 1,620,898 patient hospitalizations from 116 Veterans Affairs hospitals was assembled from electronic health record (EHR) data collected from 2003 to 2012. HA-AKI was defined at stage 1+, stage 2+, and dialysis. EHR-based predictors were identified through logistic regression, least absolute shrinkage and selection operator (lasso) regression, and random forests, and pair-wise comparisons between each were made. Calibration and discrimination metrics were calculated using 50 bootstrap iterations. In the final models, we report odds ratios, 95% confidence intervals, and importance rankings for predictor variables to evaluate their significance. RESULTS: The area under the receiver operating characteristic curve (AUC) for the different model outcomes ranged from 0.746 to 0.758 in stage 1+, 0.714 to 0.720 in stage 2+, and 0.823 to 0.825 in dialysis. Logistic regression had the best AUC in stage 1+ and dialysis. Random forests had the best AUC in stage 2+ but the least favorable calibration plots. Multiple risk factors were significant in our models, including some nonsteroidal anti-inflammatory drugs, blood pressure medications, antibiotics, and intravenous fluids given during the first 48 h of admission. CONCLUSIONS: This study demonstrated that, although all the models tested had good discrimination, performance characteristics varied between methods, and the random forests models did not calibrate as well as the lasso or logistic regression models. In addition, novel modifiable risk factors were explored and found to be significant. Published by Oxford University Press on behalf of the American Medical Informatics Association 2015. This work is written by US Government employees and is in the public domain in the US.
Entities:
Keywords:
acute kidney injury; logistic regression; random forest; risk models
Authors: R A Greevy; M M Huizinga; C L Roumie; C G Grijalva; H Murff; X Liu; M R Griffin Journal: Clin Pharmacol Ther Date: 2011-11-02 Impact factor: 6.875
Authors: Tobias Breidthardt; Mirjam Christ-Crain; Daiana Stolz; Roland Bingisser; Beatrice Drexler; Theresia Klima; Catharina Balmelli; Philipp Schuetz; Philip Haaf; Michael Schärer; Michael Tamm; Beat Müller; Christian Müller Journal: Am J Med Date: 2012-02 Impact factor: 4.965
Authors: Ewout W Steyerberg; Andrew J Vickers; Nancy R Cook; Thomas Gerds; Mithat Gonen; Nancy Obuchowski; Michael J Pencina; Michael W Kattan Journal: Epidemiology Date: 2010-01 Impact factor: 4.822
Authors: T J Walsh; J W Hiemenz; N L Seibel; J R Perfect; G Horwith; L Lee; J L Silber; M J DiNubile; A Reboli; E Bow; J Lister; E J Anaissie Journal: Clin Infect Dis Date: 1998-06 Impact factor: 9.079
Authors: Jejo D Koola; Sharon E Davis; Omar Al-Nimri; Sharidan K Parr; Daniel Fabbri; Bradley A Malin; Samuel B Ho; Michael E Matheny Journal: J Biomed Inform Date: 2018-03-09 Impact factor: 6.317
Authors: Lijuan Wu; Yong Hu; Borong Yuan; Xiangzhou Zhang; Weiqi Chen; Kang Liu; Mei Liu Journal: Int J Med Inform Date: 2020-09-11 Impact factor: 4.046
Authors: Lijuan Wu; Yong Hu; Xiaoxiao Liu; Xiangzhou Zhang; Weiqi Chen; Alan S L Yu; John A Kellum; Lemuel R Waitman; Mei Liu Journal: Sci Rep Date: 2018-11-23 Impact factor: 4.379
Authors: Sharon E Davis; Thomas A Lasko; Guanhua Chen; Edward D Siew; Michael E Matheny Journal: J Am Med Inform Assoc Date: 2017-11-01 Impact factor: 4.497
Authors: Lijuan Wu; Yong Hu; Xiangzhou Zhang; Weiqi Chen; Alan S L Yu; John A Kellum; Lemuel R Waitman; Mei Liu Journal: BMC Nephrol Date: 2020-08-02 Impact factor: 2.388