Can machine learning algorithms accurately predict discharge to nonhome facility and early unplanned readmissions following spinal fusion? Analysis of a national surgical registry

Anshit Goyal, Che Ngufor, Panagiotis Kerezoudis, Brandon McCutcheon, Curtis Storlie, Mohamad Bydon

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

OBJECTIVE Nonhome discharge and unplanned readmissions represent important cost drivers following spinal fusion. The authors sought to utilize different machine learning algorithms to predict discharge to rehabilitation and unplanned readmissions in patients receiving spinal fusion. METHODS The authors queried the 2012–2013 American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) for patients undergoing cervical or lumbar spinal fusion. Outcomes assessed included discharge to nonhome facility and unplanned readmissions within 30 days after surgery. A total of 7 machine learning algorithms were evaluated. Predictive hierarchical clustering of procedure codes was used to increase model performance. Model performance was evaluated using overall accuracy and area under the receiver operating characteristic curve (AUC), as well as sensitivity, specificity, and positive and negative predictive values. These performance metrics were computed for both the imputed and unimputed (missing values dropped) datasets. RESULTS A total of 59,145 spinal fusion cases were analyzed. The incidence rates of discharge to nonhome facility and 30-day unplanned readmission were 12.6% and 4.5%, respectively. All classification algorithms showed excellent discrimination (AUC > 0.80, range 0.85–0.87) for predicting nonhome discharge. The generalized linear model showed comparable performance to other machine learning algorithms. By comparison, all models showed poorer predictive performance for unplanned readmission, with AUC ranging between 0.63 and 0.66. Better predictive performance was noted with models using imputed data. CONCLUSIONS In an analysis of patients undergoing spinal fusion, multiple machine learning algorithms were found to reliably predict nonhome discharge with modest performance noted for unplanned readmissions. These results provide early evidence regarding the feasibility of modern machine learning classifiers in predicting these outcomes and serve as possible clinical decision support tools to facilitate shared decision making.

Original languageEnglish (US)
Pages (from-to)568-578
Number of pages11
JournalJournal of Neurosurgery: Spine
Volume31
Issue number4
DOIs
StatePublished - Oct 2019

Fingerprint

Spinal Fusion
Registries
Area Under Curve
Clinical Decision Support Systems
Patient Readmission
Quality Improvement
Ambulatory Surgical Procedures
ROC Curve
Cluster Analysis
Linear Models
Decision Making
Rehabilitation
Machine Learning
Costs and Cost Analysis
Sensitivity and Specificity
Incidence

Keywords

  • Bayes theorem
  • Cervical
  • Discharge
  • Elastic net
  • Generalized linear model
  • Gradient boosting machines
  • Logistic regression
  • Lumbar
  • Machine learning
  • Neural networks
  • NSQIP
  • Outcomes
  • Penalized discriminant analysis
  • Predictive modeling
  • Random forest
  • Rehabilitation
  • Skilled nursing facility
  • Spinal fusion
  • Spine surgery
  • Unplanned readmission

ASJC Scopus subject areas

  • Surgery
  • Neurology
  • Clinical Neurology

Cite this

Can machine learning algorithms accurately predict discharge to nonhome facility and early unplanned readmissions following spinal fusion? Analysis of a national surgical registry. / Goyal, Anshit; Ngufor, Che; Kerezoudis, Panagiotis; McCutcheon, Brandon; Storlie, Curtis; Bydon, Mohamad.

In: Journal of Neurosurgery: Spine, Vol. 31, No. 4, 10.2019, p. 568-578.

Research output: Contribution to journalArticle

@article{858628be2fef4ac4b8fe3ba60721da83,
title = "Can machine learning algorithms accurately predict discharge to nonhome facility and early unplanned readmissions following spinal fusion? Analysis of a national surgical registry",
abstract = "OBJECTIVE Nonhome discharge and unplanned readmissions represent important cost drivers following spinal fusion. The authors sought to utilize different machine learning algorithms to predict discharge to rehabilitation and unplanned readmissions in patients receiving spinal fusion. METHODS The authors queried the 2012–2013 American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) for patients undergoing cervical or lumbar spinal fusion. Outcomes assessed included discharge to nonhome facility and unplanned readmissions within 30 days after surgery. A total of 7 machine learning algorithms were evaluated. Predictive hierarchical clustering of procedure codes was used to increase model performance. Model performance was evaluated using overall accuracy and area under the receiver operating characteristic curve (AUC), as well as sensitivity, specificity, and positive and negative predictive values. These performance metrics were computed for both the imputed and unimputed (missing values dropped) datasets. RESULTS A total of 59,145 spinal fusion cases were analyzed. The incidence rates of discharge to nonhome facility and 30-day unplanned readmission were 12.6{\%} and 4.5{\%}, respectively. All classification algorithms showed excellent discrimination (AUC > 0.80, range 0.85–0.87) for predicting nonhome discharge. The generalized linear model showed comparable performance to other machine learning algorithms. By comparison, all models showed poorer predictive performance for unplanned readmission, with AUC ranging between 0.63 and 0.66. Better predictive performance was noted with models using imputed data. CONCLUSIONS In an analysis of patients undergoing spinal fusion, multiple machine learning algorithms were found to reliably predict nonhome discharge with modest performance noted for unplanned readmissions. These results provide early evidence regarding the feasibility of modern machine learning classifiers in predicting these outcomes and serve as possible clinical decision support tools to facilitate shared decision making.",
keywords = "Bayes theorem, Cervical, Discharge, Elastic net, Generalized linear model, Gradient boosting machines, Logistic regression, Lumbar, Machine learning, Neural networks, NSQIP, Outcomes, Penalized discriminant analysis, Predictive modeling, Random forest, Rehabilitation, Skilled nursing facility, Spinal fusion, Spine surgery, Unplanned readmission",
author = "Anshit Goyal and Che Ngufor and Panagiotis Kerezoudis and Brandon McCutcheon and Curtis Storlie and Mohamad Bydon",
year = "2019",
month = "10",
doi = "10.3171/2019.3.SPINE181367",
language = "English (US)",
volume = "31",
pages = "568--578",
journal = "Journal of Neurosurgery: Spine",
issn = "1547-5654",
publisher = "American Association of Neurological Surgeons",
number = "4",

}

TY - JOUR

T1 - Can machine learning algorithms accurately predict discharge to nonhome facility and early unplanned readmissions following spinal fusion? Analysis of a national surgical registry

AU - Goyal, Anshit

AU - Ngufor, Che

AU - Kerezoudis, Panagiotis

AU - McCutcheon, Brandon

AU - Storlie, Curtis

AU - Bydon, Mohamad

PY - 2019/10

Y1 - 2019/10

N2 - OBJECTIVE Nonhome discharge and unplanned readmissions represent important cost drivers following spinal fusion. The authors sought to utilize different machine learning algorithms to predict discharge to rehabilitation and unplanned readmissions in patients receiving spinal fusion. METHODS The authors queried the 2012–2013 American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) for patients undergoing cervical or lumbar spinal fusion. Outcomes assessed included discharge to nonhome facility and unplanned readmissions within 30 days after surgery. A total of 7 machine learning algorithms were evaluated. Predictive hierarchical clustering of procedure codes was used to increase model performance. Model performance was evaluated using overall accuracy and area under the receiver operating characteristic curve (AUC), as well as sensitivity, specificity, and positive and negative predictive values. These performance metrics were computed for both the imputed and unimputed (missing values dropped) datasets. RESULTS A total of 59,145 spinal fusion cases were analyzed. The incidence rates of discharge to nonhome facility and 30-day unplanned readmission were 12.6% and 4.5%, respectively. All classification algorithms showed excellent discrimination (AUC > 0.80, range 0.85–0.87) for predicting nonhome discharge. The generalized linear model showed comparable performance to other machine learning algorithms. By comparison, all models showed poorer predictive performance for unplanned readmission, with AUC ranging between 0.63 and 0.66. Better predictive performance was noted with models using imputed data. CONCLUSIONS In an analysis of patients undergoing spinal fusion, multiple machine learning algorithms were found to reliably predict nonhome discharge with modest performance noted for unplanned readmissions. These results provide early evidence regarding the feasibility of modern machine learning classifiers in predicting these outcomes and serve as possible clinical decision support tools to facilitate shared decision making.

AB - OBJECTIVE Nonhome discharge and unplanned readmissions represent important cost drivers following spinal fusion. The authors sought to utilize different machine learning algorithms to predict discharge to rehabilitation and unplanned readmissions in patients receiving spinal fusion. METHODS The authors queried the 2012–2013 American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) for patients undergoing cervical or lumbar spinal fusion. Outcomes assessed included discharge to nonhome facility and unplanned readmissions within 30 days after surgery. A total of 7 machine learning algorithms were evaluated. Predictive hierarchical clustering of procedure codes was used to increase model performance. Model performance was evaluated using overall accuracy and area under the receiver operating characteristic curve (AUC), as well as sensitivity, specificity, and positive and negative predictive values. These performance metrics were computed for both the imputed and unimputed (missing values dropped) datasets. RESULTS A total of 59,145 spinal fusion cases were analyzed. The incidence rates of discharge to nonhome facility and 30-day unplanned readmission were 12.6% and 4.5%, respectively. All classification algorithms showed excellent discrimination (AUC > 0.80, range 0.85–0.87) for predicting nonhome discharge. The generalized linear model showed comparable performance to other machine learning algorithms. By comparison, all models showed poorer predictive performance for unplanned readmission, with AUC ranging between 0.63 and 0.66. Better predictive performance was noted with models using imputed data. CONCLUSIONS In an analysis of patients undergoing spinal fusion, multiple machine learning algorithms were found to reliably predict nonhome discharge with modest performance noted for unplanned readmissions. These results provide early evidence regarding the feasibility of modern machine learning classifiers in predicting these outcomes and serve as possible clinical decision support tools to facilitate shared decision making.

KW - Bayes theorem

KW - Cervical

KW - Discharge

KW - Elastic net

KW - Generalized linear model

KW - Gradient boosting machines

KW - Logistic regression

KW - Lumbar

KW - Machine learning

KW - Neural networks

KW - NSQIP

KW - Outcomes

KW - Penalized discriminant analysis

KW - Predictive modeling

KW - Random forest

KW - Rehabilitation

KW - Skilled nursing facility

KW - Spinal fusion

KW - Spine surgery

KW - Unplanned readmission

UR - http://www.scopus.com/inward/record.url?scp=85072760959&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072760959&partnerID=8YFLogxK

U2 - 10.3171/2019.3.SPINE181367

DO - 10.3171/2019.3.SPINE181367

M3 - Article

C2 - 31174185

AN - SCOPUS:85072760959

VL - 31

SP - 568

EP - 578

JO - Journal of Neurosurgery: Spine

JF - Journal of Neurosurgery: Spine

SN - 1547-5654

IS - 4

ER -