Approaches to predicting outcomes in patients with acute kidney injury

Danielle Saly, Alina Yang, Corey Triebwasser, Janice Oh, Qisi Sun, Jeffrey Testani, Chirag Parikh, Joshua Bia, Aditya Biswas, Chess Stetson, Kris Chaisanguanthum, F. Perry Wilson

Research output: Contribution to journalArticle

Abstract

Despite recognition that Acute Kidney Injury (AKI) leads to substantial increases in morbidity, mortality, and length of stay, accurate prognostication of these clinical events remains difficult. It remains unclear which approaches to variable selection and model building are most robust. We used data from a randomized trial of AKI alerting to develop time-updated prognostic models using stepwise regression compared to more advanced variable selection techniques. We randomly split data into training and validation cohorts. Outcomes of interest were death within 7 days, dialysis within 7 days, and length of stay. Data elements eligible for model-building included lab values, medications and dosages, procedures, and demographics. We assessed model discrimination using the area under the receiver operator characteristic curve and r-squared values. 2241 individuals were available for analysis. Both modeling techniques created viable models with very good discrimination ability, with AUCs exceeding 0.85 for dialysis and 0.8 for death prediction. Model performance was similar across model building strategies, though the strategy employing more advanced variable selection was more parsimonious. Very good to excellent prediction of outcome events is feasible in patients with AKI. More advanced techniques may lead to more parsimonious models, which may facilitate adoption in other settings.

Original languageEnglish (US)
Article numbere0169305
JournalPLoS One
Volume12
Issue number1
DOIs
StatePublished - Jan 1 2017

Fingerprint

Acute Kidney Injury
kidneys
Dialysis
Length of Stay
Area Under Curve
Demography
dialysis
Morbidity
Mortality
death
prediction
drug therapy
morbidity
demographic statistics
methodology
dosage

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Agricultural and Biological Sciences(all)

Cite this

Saly, D., Yang, A., Triebwasser, C., Oh, J., Sun, Q., Testani, J., ... Wilson, F. P. (2017). Approaches to predicting outcomes in patients with acute kidney injury. PLoS One, 12(1), [e0169305]. https://doi.org/10.1371/journal.pone.0169305

Approaches to predicting outcomes in patients with acute kidney injury. / Saly, Danielle; Yang, Alina; Triebwasser, Corey; Oh, Janice; Sun, Qisi; Testani, Jeffrey; Parikh, Chirag; Bia, Joshua; Biswas, Aditya; Stetson, Chess; Chaisanguanthum, Kris; Wilson, F. Perry.

In: PLoS One, Vol. 12, No. 1, e0169305, 01.01.2017.

Research output: Contribution to journalArticle

Saly, D, Yang, A, Triebwasser, C, Oh, J, Sun, Q, Testani, J, Parikh, C, Bia, J, Biswas, A, Stetson, C, Chaisanguanthum, K & Wilson, FP 2017, 'Approaches to predicting outcomes in patients with acute kidney injury', PLoS One, vol. 12, no. 1, e0169305. https://doi.org/10.1371/journal.pone.0169305
Saly D, Yang A, Triebwasser C, Oh J, Sun Q, Testani J et al. Approaches to predicting outcomes in patients with acute kidney injury. PLoS One. 2017 Jan 1;12(1). e0169305. https://doi.org/10.1371/journal.pone.0169305
Saly, Danielle ; Yang, Alina ; Triebwasser, Corey ; Oh, Janice ; Sun, Qisi ; Testani, Jeffrey ; Parikh, Chirag ; Bia, Joshua ; Biswas, Aditya ; Stetson, Chess ; Chaisanguanthum, Kris ; Wilson, F. Perry. / Approaches to predicting outcomes in patients with acute kidney injury. In: PLoS One. 2017 ; Vol. 12, No. 1.
@article{5e2ff9f818e84628bbb6391d9938a318,
title = "Approaches to predicting outcomes in patients with acute kidney injury",
abstract = "Despite recognition that Acute Kidney Injury (AKI) leads to substantial increases in morbidity, mortality, and length of stay, accurate prognostication of these clinical events remains difficult. It remains unclear which approaches to variable selection and model building are most robust. We used data from a randomized trial of AKI alerting to develop time-updated prognostic models using stepwise regression compared to more advanced variable selection techniques. We randomly split data into training and validation cohorts. Outcomes of interest were death within 7 days, dialysis within 7 days, and length of stay. Data elements eligible for model-building included lab values, medications and dosages, procedures, and demographics. We assessed model discrimination using the area under the receiver operator characteristic curve and r-squared values. 2241 individuals were available for analysis. Both modeling techniques created viable models with very good discrimination ability, with AUCs exceeding 0.85 for dialysis and 0.8 for death prediction. Model performance was similar across model building strategies, though the strategy employing more advanced variable selection was more parsimonious. Very good to excellent prediction of outcome events is feasible in patients with AKI. More advanced techniques may lead to more parsimonious models, which may facilitate adoption in other settings.",
author = "Danielle Saly and Alina Yang and Corey Triebwasser and Janice Oh and Qisi Sun and Jeffrey Testani and Chirag Parikh and Joshua Bia and Aditya Biswas and Chess Stetson and Kris Chaisanguanthum and Wilson, {F. Perry}",
year = "2017",
month = "1",
day = "1",
doi = "10.1371/journal.pone.0169305",
language = "English (US)",
volume = "12",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "1",

}

TY - JOUR

T1 - Approaches to predicting outcomes in patients with acute kidney injury

AU - Saly, Danielle

AU - Yang, Alina

AU - Triebwasser, Corey

AU - Oh, Janice

AU - Sun, Qisi

AU - Testani, Jeffrey

AU - Parikh, Chirag

AU - Bia, Joshua

AU - Biswas, Aditya

AU - Stetson, Chess

AU - Chaisanguanthum, Kris

AU - Wilson, F. Perry

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Despite recognition that Acute Kidney Injury (AKI) leads to substantial increases in morbidity, mortality, and length of stay, accurate prognostication of these clinical events remains difficult. It remains unclear which approaches to variable selection and model building are most robust. We used data from a randomized trial of AKI alerting to develop time-updated prognostic models using stepwise regression compared to more advanced variable selection techniques. We randomly split data into training and validation cohorts. Outcomes of interest were death within 7 days, dialysis within 7 days, and length of stay. Data elements eligible for model-building included lab values, medications and dosages, procedures, and demographics. We assessed model discrimination using the area under the receiver operator characteristic curve and r-squared values. 2241 individuals were available for analysis. Both modeling techniques created viable models with very good discrimination ability, with AUCs exceeding 0.85 for dialysis and 0.8 for death prediction. Model performance was similar across model building strategies, though the strategy employing more advanced variable selection was more parsimonious. Very good to excellent prediction of outcome events is feasible in patients with AKI. More advanced techniques may lead to more parsimonious models, which may facilitate adoption in other settings.

AB - Despite recognition that Acute Kidney Injury (AKI) leads to substantial increases in morbidity, mortality, and length of stay, accurate prognostication of these clinical events remains difficult. It remains unclear which approaches to variable selection and model building are most robust. We used data from a randomized trial of AKI alerting to develop time-updated prognostic models using stepwise regression compared to more advanced variable selection techniques. We randomly split data into training and validation cohorts. Outcomes of interest were death within 7 days, dialysis within 7 days, and length of stay. Data elements eligible for model-building included lab values, medications and dosages, procedures, and demographics. We assessed model discrimination using the area under the receiver operator characteristic curve and r-squared values. 2241 individuals were available for analysis. Both modeling techniques created viable models with very good discrimination ability, with AUCs exceeding 0.85 for dialysis and 0.8 for death prediction. Model performance was similar across model building strategies, though the strategy employing more advanced variable selection was more parsimonious. Very good to excellent prediction of outcome events is feasible in patients with AKI. More advanced techniques may lead to more parsimonious models, which may facilitate adoption in other settings.

UR - http://www.scopus.com/inward/record.url?scp=85010953626&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85010953626&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0169305

DO - 10.1371/journal.pone.0169305

M3 - Article

C2 - 28122032

AN - SCOPUS:85010953626

VL - 12

JO - PLoS One

JF - PLoS One

SN - 1932-6203

IS - 1

M1 - e0169305

ER -