Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke

Kenneth Ottenbacher, Pam M. Smith, Sandra B. Illig, Richard T. Linn, Roger C. Fiedler, Carl V. Granger

Research output: Contribution to journalArticle

60 Citations (Scopus)

Abstract

Context: Rehospitalization following inpatient medical rehabilitation has important health and economic implications for patients who have experienced a stroke. Objective: Compare logistic regression and neural networks in predicting rehospitalization at 3-6-month follow-up for patients with stroke discharged from medical rehabilitation. Design: The study was retrospective using information from a national database representative of medical rehabilitation patients across the US. Setting: Information submitted to the Uniform Data System for Medical Rehabilitation from 1997 and 1998 by 167 hospital and rehabilitation facilities from 40 states was examined. Participants: 9584 patient records were included in the sample. The mean age was 70.74 years (SD = 12.87). The sample included 51.6% females and was 77.6% non-Hispanic White with an average length of stay of 21.47 days (SD = 15.47). Main Outcome Measures: Hospital readmission from 80 to 180 days following discharge. Results: Statistically significant variables (P < .05) in the logistic model included sphincter control, self-care ability, age, marital status, ethnicity and length of stay. Area under the ROC curves were 0.68 and 0.74 for logistic regression and neural network analysis, respectively. The Hosmer-Lemeshow goodness-of-fit chi-square was 11.32 (df = 8, P = 0.22) for neural network analysis and 16.33 (df = 8, P = 0.11) for logistic regression. Calibration curves indicated a slightly better fit for the neural network model. Conclusion: There was no statistically significant or practical advantage in predicting hospital readmission using neural network analysis in comparison to logistic regression for persons who experienced a stroke and received medical rehabilitation during the period of the study.

Original languageEnglish
Pages (from-to)1159-1165
Number of pages7
JournalJournal of Clinical Epidemiology
Volume54
Issue number11
DOIs
StatePublished - 2001

Fingerprint

Rehabilitation
Logistic Models
Stroke
Patient Readmission
Length of Stay
Aptitude
Neural Networks (Computer)
Marital Status
Self Care
Information Systems
ROC Curve
Calibration
Area Under Curve
Inpatients
Retrospective Studies
Economics
Outcome Assessment (Health Care)
Databases
Health

Keywords

  • Logistic regression
  • Neural networks
  • Stroke

ASJC Scopus subject areas

  • Medicine(all)
  • Public Health, Environmental and Occupational Health
  • Epidemiology

Cite this

Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke. / Ottenbacher, Kenneth; Smith, Pam M.; Illig, Sandra B.; Linn, Richard T.; Fiedler, Roger C.; Granger, Carl V.

In: Journal of Clinical Epidemiology, Vol. 54, No. 11, 2001, p. 1159-1165.

Research output: Contribution to journalArticle

Ottenbacher, Kenneth ; Smith, Pam M. ; Illig, Sandra B. ; Linn, Richard T. ; Fiedler, Roger C. ; Granger, Carl V. / Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke. In: Journal of Clinical Epidemiology. 2001 ; Vol. 54, No. 11. pp. 1159-1165.
@article{432be1fba84b4e64a81d95a02a8cff77,
title = "Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke",
abstract = "Context: Rehospitalization following inpatient medical rehabilitation has important health and economic implications for patients who have experienced a stroke. Objective: Compare logistic regression and neural networks in predicting rehospitalization at 3-6-month follow-up for patients with stroke discharged from medical rehabilitation. Design: The study was retrospective using information from a national database representative of medical rehabilitation patients across the US. Setting: Information submitted to the Uniform Data System for Medical Rehabilitation from 1997 and 1998 by 167 hospital and rehabilitation facilities from 40 states was examined. Participants: 9584 patient records were included in the sample. The mean age was 70.74 years (SD = 12.87). The sample included 51.6{\%} females and was 77.6{\%} non-Hispanic White with an average length of stay of 21.47 days (SD = 15.47). Main Outcome Measures: Hospital readmission from 80 to 180 days following discharge. Results: Statistically significant variables (P < .05) in the logistic model included sphincter control, self-care ability, age, marital status, ethnicity and length of stay. Area under the ROC curves were 0.68 and 0.74 for logistic regression and neural network analysis, respectively. The Hosmer-Lemeshow goodness-of-fit chi-square was 11.32 (df = 8, P = 0.22) for neural network analysis and 16.33 (df = 8, P = 0.11) for logistic regression. Calibration curves indicated a slightly better fit for the neural network model. Conclusion: There was no statistically significant or practical advantage in predicting hospital readmission using neural network analysis in comparison to logistic regression for persons who experienced a stroke and received medical rehabilitation during the period of the study.",
keywords = "Logistic regression, Neural networks, Stroke",
author = "Kenneth Ottenbacher and Smith, {Pam M.} and Illig, {Sandra B.} and Linn, {Richard T.} and Fiedler, {Roger C.} and Granger, {Carl V.}",
year = "2001",
doi = "10.1016/S0895-4356(01)00395-X",
language = "English",
volume = "54",
pages = "1159--1165",
journal = "Journal of Clinical Epidemiology",
issn = "0895-4356",
publisher = "Elsevier USA",
number = "11",

}

TY - JOUR

T1 - Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke

AU - Ottenbacher, Kenneth

AU - Smith, Pam M.

AU - Illig, Sandra B.

AU - Linn, Richard T.

AU - Fiedler, Roger C.

AU - Granger, Carl V.

PY - 2001

Y1 - 2001

N2 - Context: Rehospitalization following inpatient medical rehabilitation has important health and economic implications for patients who have experienced a stroke. Objective: Compare logistic regression and neural networks in predicting rehospitalization at 3-6-month follow-up for patients with stroke discharged from medical rehabilitation. Design: The study was retrospective using information from a national database representative of medical rehabilitation patients across the US. Setting: Information submitted to the Uniform Data System for Medical Rehabilitation from 1997 and 1998 by 167 hospital and rehabilitation facilities from 40 states was examined. Participants: 9584 patient records were included in the sample. The mean age was 70.74 years (SD = 12.87). The sample included 51.6% females and was 77.6% non-Hispanic White with an average length of stay of 21.47 days (SD = 15.47). Main Outcome Measures: Hospital readmission from 80 to 180 days following discharge. Results: Statistically significant variables (P < .05) in the logistic model included sphincter control, self-care ability, age, marital status, ethnicity and length of stay. Area under the ROC curves were 0.68 and 0.74 for logistic regression and neural network analysis, respectively. The Hosmer-Lemeshow goodness-of-fit chi-square was 11.32 (df = 8, P = 0.22) for neural network analysis and 16.33 (df = 8, P = 0.11) for logistic regression. Calibration curves indicated a slightly better fit for the neural network model. Conclusion: There was no statistically significant or practical advantage in predicting hospital readmission using neural network analysis in comparison to logistic regression for persons who experienced a stroke and received medical rehabilitation during the period of the study.

AB - Context: Rehospitalization following inpatient medical rehabilitation has important health and economic implications for patients who have experienced a stroke. Objective: Compare logistic regression and neural networks in predicting rehospitalization at 3-6-month follow-up for patients with stroke discharged from medical rehabilitation. Design: The study was retrospective using information from a national database representative of medical rehabilitation patients across the US. Setting: Information submitted to the Uniform Data System for Medical Rehabilitation from 1997 and 1998 by 167 hospital and rehabilitation facilities from 40 states was examined. Participants: 9584 patient records were included in the sample. The mean age was 70.74 years (SD = 12.87). The sample included 51.6% females and was 77.6% non-Hispanic White with an average length of stay of 21.47 days (SD = 15.47). Main Outcome Measures: Hospital readmission from 80 to 180 days following discharge. Results: Statistically significant variables (P < .05) in the logistic model included sphincter control, self-care ability, age, marital status, ethnicity and length of stay. Area under the ROC curves were 0.68 and 0.74 for logistic regression and neural network analysis, respectively. The Hosmer-Lemeshow goodness-of-fit chi-square was 11.32 (df = 8, P = 0.22) for neural network analysis and 16.33 (df = 8, P = 0.11) for logistic regression. Calibration curves indicated a slightly better fit for the neural network model. Conclusion: There was no statistically significant or practical advantage in predicting hospital readmission using neural network analysis in comparison to logistic regression for persons who experienced a stroke and received medical rehabilitation during the period of the study.

KW - Logistic regression

KW - Neural networks

KW - Stroke

UR - http://www.scopus.com/inward/record.url?scp=0034765730&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034765730&partnerID=8YFLogxK

U2 - 10.1016/S0895-4356(01)00395-X

DO - 10.1016/S0895-4356(01)00395-X

M3 - Article

C2 - 11675168

AN - SCOPUS:0034765730

VL - 54

SP - 1159

EP - 1165

JO - Journal of Clinical Epidemiology

JF - Journal of Clinical Epidemiology

SN - 0895-4356

IS - 11

ER -